Learn more: PMC Disclaimer | PMC Copyright Notice
Presecretory oxidation, aggregation, and autophagic destruction of apoprotein-B: A pathway for late-stage quality control
Associated Data
Abstract
Hepatic secretion of apolipoprotein-B (apoB), the major protein of atherogenic lipoproteins, is regulated through posttranslational degradation. We reported a degradation pathway, post-ER pre secretory proteolysis (PERPP), that is increased by reactive oxygen species (ROS) generated within hepatocytes from dietary polyunsaturated fatty acids (PUFA). We now report the molecular processes by which PUFA-derived ROS regulate PERPP of apoB. ApoB exits the ER; undergoes limited oxidant-dependent aggregation; and then, upon exit from the Golgi, becomes extensively oxidized and converted into large aggregates. The aggregates slowly degrade by an autophagic process. None of the oxidized, aggregated material leaves cells, thereby preventing export of apoB-lipoproteins containing potentially toxic lipid peroxides. In summary, apoB secretory control via PERPP/autophagosomes is likely a key component of normal and pathologic regulation of plasma apoB levels, as well as a means for remarkably late-stage quality control of a secreted protein.
Apolipoprotein-B100 (apoB), a 550-kDa glycoprotein, is an essential component of atherogenic plasma lipoproteins (1). Retention, or trapping, of apoB-lipoproteins within the arterial wall is the key initiating event in the pathogenesis of atherosclerosis, the major cause of death world-wide (2–4). Thus, apoB-lipoprotein assembly and secretion, which are often dysregulated (5, 6), attract considerable interest.
ApoB is constitutively synthesized in the liver, and its secretion is controlled mainly through posttranslational destruction. Three processes are known (1, 7, 8): (i) endoplasmic reticulum association degradation (ERAD) mediated by the proteasome and stimulated by severe scarcity of lipids for transfer to apoB during its translocation across the ER (1); (ii) post-ER presecretory proteolysis (PERPP) (8), a nonproteasomal pathway provoked by diverse metabolic factors in vitro and in vivo, including intracellular lipid peroxidation (9–14); and (iii) receptor-mediated degradation, also known as reuptake, which occurs via interactions of nascent apoB-particles with LDL receptors (8, 10, 15–18) and heparan sulfate proteoglycans (17, 19).
We recently reported that dietary polyunsaturated fatty acids (PUFAs), clinically important lipid-lowering agents, increase presecretory apoB degradation and inhibit apoB secretion from hepatocytes in vitro and in vivo (9). The effects of PUFAs require their intracellular conversion into lipid peroxides that then trigger the PERPP pathway. Both conventional Ω-6 PUFAs and marine (fish oil) Ω-3 PUFAs, such as docosahexaenoic acid (DHA), have this effect. Antioxidants inhibit apoB degradation in PUFA-treated cells, and thereby enhance the secretion of atherogenic apoB-lipoproteins in vitro and in vivo (9, 10). In the current study, we sought to identify the novel molecular processes by which PUFA-derived lipid peroxides trigger PERPP of newly synthesized apoB.
Results
PUFA-Treatment of Hepatic Cells Provokes Newly Synthesized apoB to Aggregate Before Its Destruction.
Incubation of hepatocytes with DHA or other PUFAs substantially lowers the total (intracellular + secreted) recovery of labeled monomeric apoB (MW ≈550 kDa) owing to stimulation of PERPP (7–9). To characterize the degradative process, we tested approximately two dozen inhibitors against members of the major classes of known intracellular proteases. Surprisingly, none had any significant effect on PUFA-induced loss of newly synthesized labeled monomeric apoB (data not shown).
During these experiments, however, we noticed that PUFAs consistently induced the appearance of material that was precipitated from cellular homogenates by anti-apoB antiserum but exhibited retarded migration on SDS/PAGE relative to monomeric apoB. As shown in Fig. 1, in either rat primary hepatocytes or rat hepatoma McArdle RH-7777 (McA) cells, DHA (complexed to BSA) caused intracellular accumulation of immunoreactive material of high molecular weight in two regions of the lanes: a compact collection within the very top of the stacking gel [large aggregates (LAs)] and a dispersed population between the monomeric apoB band and the interface of the stacking and running gels [intermediate aggregates (IAs)]. ApoB aggregates were not detected in the conditioned media (Fig. 1). Cells treated with either BSA or BSA complexed with oleic acid (OA) contained no or nearly no LAs and much lower amounts of IAs than in DHA-treated cells (Fig. 1).
DHA induces intracellular aggregation of apoB and blocks its secretion, whereas BSA and OA allow release of monomeric apoB into the medium. Primary rat hepatocytes (A) or McA cells (B) were pretreated for 1 h in medium supplemented with BSA without fatty acids or with DHA or OA (0.6 mM, complexed to BSA) as indicated, and then metabolically labeled to steady-state. Cell lysates (Cell) and medium (Medium) were subjected to anti-apoB immunoprecipitation and SDS/PAGE. Displayed are fluorograms of labeled immunoreactive material. LA, large aggregates; IA, intermediate aggregates; apoB100, monomeric apoB. The line to the left of each fluorogram indicates the interface between the stacking and running gels. Each LA band penetrated into the top of the stacking gel.
Homogenates of DHA-treated cells are enriched in lipid peroxides (9) that might artifactually alter apoB after cell lysis. Thus, we prepared homogenates of cells that had been treated with BSA and [35S]methionine (met)/cysteine (cys) in the absence of added fatty acids, i.e., in which little aggregated material should be present. These labeled homogenates were mixed 1:1 either with buffer or with unlabeled homogenates from BSA- or DHA-treated cells, and then subjected to immunoprecipitation and SDS/PAGE. No LAs and no increase in IAs were detected in these mixed homogenates (data not shown). Thus, the aggregates form in intact cells, not after homogenization.
Intracellular abundance of radio-labeled LAs or IAs was directly related to the DHA concentration and a corresponding loss of apoB from the medium (Fig. 2). Quantification of intracellular plus extracellular band intensities demonstrated a remarkably consistent total signal across the DHA concentration range, indicating that the accumulated immunoreactive material inside the cells accounts for the loss of secreted apoB (Fig. 2, line graph).
Dose-response to DHA. (Top and Middle) McA cells were pretreated for 1 h with the indicated concentrations of DHA and then metabolically labeled to steady-state. Cell lysates and media were subjected to anti-apoB immunoprecipitation and SDS/PAGE. Displayed are resulting fluorograms. Abbreviations and other indicators are as in Fig. 1. (Bottom) Intensities of the bands of LAs, IAs, and cellular and secreted monomeric apoB at each DHA concentration from the above fluorograms were determined by densitometry. Sums (cells plus media) were calculated taking into account the difference in exposure time of the two fluorograms (24 h and 12 h, respectively). Data are displayed in arbitrary units (AU).
Two key non-marine PUFAs, arachidonic (ARA) and linolenic (LNA) acids, also inhibit apoB secretion through ROS-dependent PERPP (9). Incubation with either of these fatty acids provoked robust intracellular formation of apoB aggregates [supporting information (SI) Fig. 7]. LNA is unusually potent at inhibiting apoB secretion in vitro via PERPP (9), and dietary intake of this fatty acid is linked to low plasma triglyceride levels and decreased cardiovascular risk (20). Of interest, it provoked the conversion of essentially all monomeric apoB into LAs and IAs by 120 min of chase (SI Fig. 7).
We next performed pulse–chase time courses on DHA-treated McA cells (Fig. 3). During longer chase incubations (Fig. 3B), labeled LAs and IAs eventually disappear from the intracellular plus secreted pool. As in Fig. 1, apoB-aggregates do not appear in the medium during 3-h or longer (data not shown) incubations. Taken together, these data indicate that conversion of newly synthesized monomeric apoB into IAs and LAs is required for PUFA-stimulated degradation of the protein.
DHA-induced intracellular aggregates form from monomeric apoB and then gradually disappear over time. McA cells were pretreated for 1 h with DHA/BSA complexes, pulse-labeled for 15 min, and chased for the indicated times. Cell lysates and the corresponding conditioned media were combined 1:1, and apoB contents were analyzed by anti-apoB immunoprecipitation and SDS/PAGE. Fluorograms are shown from short-term (A) and long-term (B) chases. Abbreviations and other indicators are as in Fig. 1.
Large Aggregates Contain apoB That Is Oxidized, and Their Formation Requires Lipid Peroxidation.
We focused on the chemical nature of LAs because they are the most extensively aggregated form (Figs. 11–3) and apparently a proximal substrate for degradation (Fig. 3B). First, LAs were isolated and analyzed by mass spectroscopy. Peptide sequences from all rat proteins identified as statistically significant are given in SI Table 1. All these sequences corresponded to rat apoB, except for three rat apoA-I peptides. Six distinct apoB peptides appeared beyond residue 2179, which is definitive evidence for rat apoB100. Thus, LAs are composed primarily of apoB, with small amounts of apoA-I, a minor component of rat VLDL. No other rat proteins in LAs were detected by this method.
Because ROS play a crucial role in PERPP (9), we assessed oxidative modifications in aggregated apoB. LAs showed strong anti-MDA reactivity, demonstrating oxidized epitopes, whereas IAs and monomeric apoB showed no detectable anti-MDA reactivity (Fig. 4). To establish a causal connection between oxidation and the intracellular formation of apoB aggregates, McA cells were incubated for 2.5 h in [35S]met/cys in the presence of BSA alone, or with DHA, without or with lipid antioxidants or the iron chelator desferrioxamine (DFX). As we report in ref. 9, antioxidants or DFX blocked intracellular TBARS generation and increased the secretion of monomeric apoB (data not shown). Importantly, these treatments also inhibited the formation of both LAs and IAs (Fig. 4C). The effect on the amount of LAs was more pronounced, consistent with this species containing the most heavily oxidized apoB (Fig. 4B).
DHA-induced large aggregates of apoB contain oxidized epitopes and require lipid peroxidation for formation. (A) McA cells were incubated for 1 h with BSA or DHA/BSA complexes and then metabolically labeled to steady-state. Cellular lysates and media were subjected to anti-apoB immunoprecipitation and SDS/PAGE and then analyzed by fluorography. (B) Lysates from cells treated in parallel, but without the radioactive label, were analyzed by Western blot, using an antibody against MDA. Lane 1, positive control; lane 2, conditioned medium from DHA-treated cells; lanes 3 and 4, lysates of BSA-treated cells; lanes 5 and 6, lysates of DHA-treated cells. Other abbreviations and indicators are as in Fig. 1. (C) McA cells were treated with DHA alone or in combination with the indicated lipid anti-oxidants [100 μM Trolox; 150 μM vitamin E (Vit E)] or iron chelator (100 μM DFX). Cells were then metabolically labeled to steady-state. Displayed is a fluorogram after anti-apoB immunoprecipitation/SDS/PAGE analysis of cell lysates.
Intracellular Trafficking Is Required for apoB Aggregates to Form.
Because lipid peroxides trigger intracellular apoB aggregation in intact cells (Fig. 4C) but not in cellular homogenates, we hypothesized a role for intracellular trafficking in LA or IA formation. To test this idea, we blocked apoB exit from the ER, using a combination of brefeldin A and nocodazole (BFA/Noc) (21), which we showed to inhibit PERPP (10). Importantly, BFA/Noc only moderately decreased the intracellular content of lipid peroxides in DHA-treated cells (DHA, 81.4 ± 4.3; DHA plus BFA/Noc, 53.3 ± 4.4 pmol of MDA equivalents per milligram of cell protein; P < 0.01), but nearly completely blocked the formation of LAs and IAs (Fig. 5A). The newly synthesized apoB remained almost entirely monomeric.
DHA-induced apoB aggregation requires intracellular trafficking. McA cells were preincubated for 1 h with BSA, DHA, or OA as indicated and then metabolically labeled by a short pulse followed by a chase. Pulse and chase temperatures were 37°C unless otherwise specified. Displayed are fluorograms after anti-apoB immunoprecipitation/SDS/PAGE of cells lysed at the end of the indicated chase periods. (A) Effect of blocking apoB exit from the ER. The pulse was 15 min, and the chase was 120 min. To block ER exit, added were nocodazole (Noc) (4 μg/ml, administered 30 min before labeling and then present in the medium thereafter) and brefeldin A (BFA) (2 μg/ml, administered at the start of the chase). (B) Effect of cooling to block apoB exit from the Golgi. The pulse was 5 min followed by an 8-min chase to reach the peak of 35S incorporation into apoB. The chase was 75 min at 37°C or 20°C, as indicated. (C) Effect of releasing the temperature block. The pulse was 15 min at 20°C. After an initial chase of 60 min at 20°C (Chase 1), the cells in one group of wells were lysed, whereas the rest were extended for another 60 min of chase at either at 20°C or 37°C as indicated (Chase 2).
Next, we blocked apoB exit from the Golgi in DHA-treated cells. Surprisingly, the standard inhibitor, monensin (5 μM), decreased cellular TBARS by nearly 50%. IAs continued to form in the presence of monensin, but the inhibitor dramatically reduced the level of LAs (SI Fig. 8). To avoid this large effect on ROS, we used a second method: cooling the cells to 20°C, which allows trafficking into, but not out of, the Golgi (22, 23). As a control, cooling OA-treated cells to 20°C blocked all secretion of 35S-labeled apoB (see Fig. 5C). Importantly, cooling DHA-treated McA cells to 20°C did not lower the cellular content of lipid peroxides (DHA at 37°C, 60.1 ± 3.7 vs. DHA at 20°C, 80.7 ± 6.1 pmol of MDA equivalents per milligram of cell protein), although, similar to monensin, it allowed continued formation of IAs, but nearly completely inhibited the appearance of LAs (Fig. 5B). Bringing the cooled DHA-treated cells from 20°C to 37°C allowed labeled intracellular apoB to be chased into LAs, whereas warming OA-treated cells allowed the intracellular apoB to be chased into the medium in monomeric form (Fig. 5C). To verify that impaired formation of LAs at 20°C is not a nonspecific consequence of low temperature, we found that DHA-treated cells cooled to 24°C, which allows proteins to exit from the Golgi (22, 24), formed abundant IAs and LAs (data not shown). Taken together, these results indicate that both IA and LA formation requires lipid peroxidation (Fig. 4C) and trafficking out of the ER (Fig. 5A) but that LA formation further requires trafficking out of the Golgi (Fig. 5B).
Degradation of LAs by the Autophagosomal/Lysosomal System.
As shown in Fig. 3B, LAs gradually disappeared over several hours. We now revisited the efficacy of several protease inhibitors, none of which was effective at blocking the disappearance of monomeric apoB from DHA-treated cells. Lactacystin, an inhibitor of the proteasome, failed to protect LAs (Fig. 6A). In contrast, E64d, an inhibitor of lysosomal/endosomal cathepsins, significantly prolonged the decay of the LA signal (Fig. 6B).
Involvement of autophagosomes/lysosomes in intracellular degradation of apoB. (A–C) McA cells were preincubated with DHA for 60 min and then pulse labeled 37°C for 15 min, followed by a 3-h chase at 37°C to give enough time for labeled LAs to degrade (see Fig. 3). At the end of chase, cells were subjected to anti-apoB immunoprecipitation, SDS/PAGE, and fluorography. Typical fluorograms are displayed at Left. (Right) Quantifications of intensities of LA bands under different conditions (means ± SEM, n = 4). (A) Effect of proteasomal inhibition. Lactacystin (Lac) (10 μM) was added 30 min before labeling and kept present in the medium thereafter. The two columns were not significantly different. (B) Effect of inhibiting the autophagosome/lysosome system. Ten millimolar 3MA or 10 μM E64d was administered at the start of chase. The columns were significantly different (P < 0.001). *, P < 0.01 versus the DHA control. (C) Effect of stimulation of autophagosomes. Rapamycin (Rap) (250 nM) was added at the start of the chase period. *, P < 0.0001. (D) McA cells were transfected with an expression plasmid for LC3-GFP. After 24 h, the cells were incubated with either OA (a–f) or DHA (g–l) for 4 h, after which they were fixed and permeabilzed, and apoB was detected by indirect immunofluorescence (primary antibody, anti-apoB; secondary, goat anti-rabbit conjugated to Texas red). Nuclei were stained blue by DAPI. Laser confocal microscopy of representative cells are shown. Note the colocalization of apoB and LC3 (yellow-orange structures) in the DHA-incubated cells in the juxtanuclear region, which is consistent with the subcellular distribution of autophagosomes and lysosomes.
The protective effect of E64d implied a lysosomal process, and so we turned our attention to autophagosomes, which transport damaged cytosolic proteins and organelles into lysosomes for degradation (25–27). Autophagosomes were recently shown by Ohsaki et al. (28) to be colocalized with a small population of apoB molecules in human Huh-7 hepatoma cells after inhibition of the proteasome (28) and in preliminary reports by Thorne–Tjomsland et al. (29) and Tran et al. (30) to be induced in hepatic cells treated with Ω-3 fatty acids, perhaps consistent with the known induction of autophagy under prooxidative conditions (25). Addition of 3-methyladenine, which inhibits a PI3-kinase required for autophagosome formation (25, 26), significantly protected LAs in metabolically labeled DHA-treated cells while increasing the IA signal (Fig. 6B). This result may explain our prior finding of a role for PI3-kinases in PERPP (8). Likewise, we found that rapamycin, which stimulates autophagosomes by inhibiting the negative regulator mTOR (25, 26), significantly facilitated the degradation of LAs (Fig. 6C). None of these inhibitors affected the cellular content of lipid peroxides (data not shown).
To strengthen the relationship between apoB degradation and autophagy, colocalization studies in McA cells were performed taking a standard approach of first expressing, by transient transfection, a GFP-tagged version of the autophagosomal protein LC-3 (31). Then, the transfected McA cells were incubated with BSA complexed to either OA or DHA for 4 h, after which the cells were fixed, permeabilized, and incubated with rabbit anti-apoB antiserum, followed by Texas red-conjugated goat anti-rabbit secondary antiserum. In DHA-incubated cells, laser confocal microscopy revealed areas of colocalization of LC3-GFP and apoB, as illustrated by the yellow-orange structures in representative cells incubated with DHA (Fig. 6D i and l).
Other evidence (presented in SI Figs. 9–12) in support of a role for autophagy in PUFA-stimulated apoB aggregation and degradation was as follows: (i) immunofluorescent microscopy consistent with movement of apoB species into autophagosomes in DHA-treated McA cells: the intracellular apoB signal became concentrated in a juxtanuclear localization (as also seen in Fig. 6D), consistent with the subcellular distribution of autophagosomes and lysosomes (SI Fig. 9); (ii) the coimmunoprecipitation of LC3 and LAs, and smaller amounts of IAs and monomeric apoB, in DHA-incubated McA cells, with the degree of interaction decreased by DFX cotreatment (SI Fig. 10); (iii) increased recovery of the lipidated form of LC3 (LC3-II) in DHA-incubated McA cells (SI Fig. 11); and (iv) increased apoB recovery in DHA-treated McA cells that had been transfected with siRNA directed against the essential autophagy protein Atg7 (SI Fig. 12).
Discussion
Based on the present results, we conclude that in hepatic cells lipid peroxides derived from PUFAs provoke limited oxidative modification and aggregation of apoB after its exit from the ER but before it leaves the Golgi. Extensive oxidation and massive aggregation of apoB then occurs but requires trafficking out of the Golgi. Little or none of the oxidized aggregated material leaves the cells. Instead, it is disposed of by autophagosome-lysosomes. The presence of detectable amounts of intracellular apoB-aggregates, even without PUFA supplementation (Figs. 1, ,2,2, and and5),5), extends our earlier demonstrations of ROS-dependent control of apoB secretion in the basal state, indicating broad physiologic importance (9, 10). Thus, the PERPP pathway appears to be not only a major regulator of hepatic apoB secretion but also a means for remarkably late-stage quality control of an endogenous protein after it has left the ER.
Prior studies on quality control in the secretory pathway have focused on preventing the exit of malfolded proteins from the ER (reviewed in ref. 32). In contrast, studies on autophagosomes have emphasized their importance in degradation of the cytosol, including long-lived proteins, superfluous or damaged organelles, and overflow from ERAD (25–27). We doubt that autophagic degradation of apoB is a consequence of ERAD being overwhelmed, because analysis of LAs by mass spectroscopy did not detect ubiquitin. Rather, the autophagic process appears to represent a role in secretory regulation of apoB independent from ERAD, as suggested in refs. 30 and 33. The reason may be that PUFAs, an essential part of the diets of herbivores and omnivores, are readily converted into lipid peroxides inside cells. Although one physiologic response to PUFA-rich diets is the induction of anti-oxidant enzymes in the liver (34), our results suggest that another physiologic response is the aggregation and then destruction of a portion of newly synthesized apoB.
This selective destruction of some of the apoB late in the secretory pathway (Fig. 5) is consistent with demonstrations that full lipidation of apoB-lipoproteins to VLDL buoyancy occurs post-ER (35–37). Also, given the conspicuous physical barriers against translocating large, extensively lipidated apoB-lipoproteins out of the secretory pathway, which would be required for the proteasome to have access to them, the autophagosomal/lysosomal system may be the only plausible mechanism for late-stage quality control of these particles. Although we have implicated autophagy in apoB-PERPP and have localized the later steps of the pathway to the Golgi and beyond, still to be determined are the source (e.g., mitochondria, microsomal cytochrome P450 enzymes, and peroxisomes) of the modifying free radicals and where and how they first encounter apoB.
Besides protecting the liver from potentially toxic aggregates (38), destruction of these particles may also prevent the export of lipid peroxides on VLDL to triglyceride-consuming, but oxidant-sensitive, tissues in the periphery, such as cardiac muscle (39). Thus, lipid peroxidation is kept within the liver, where it is safe, even beneficial (40), and away from potentially vulnerable peripheral organs. Reverse transport of oxidized or oxidizable lipids from periphery to liver could serve a similar function (9, 41).
Regardless of teleologic explanations, hepatocytes have adapted an ancient eukaryotic process, autophagy, to the quality control of a comparatively recent structure, the large, lipidated, post-ER apoB-lipoprotein, especially in response to metabolic perturbation. ApoB may be particularly susceptible to modification by lipid peroxides inside hepatocytes, both because it would be assembled into close contact with them and because the unique structure of apoB facilitates its aggregation under a wide range of physiologic (e.g., Figs. 1 and and2),2), pathogenic, and even artificial stimuli that do not affect other proteins to the same extent (4, 42–47). Its remarkable susceptibility to aggregation allows a normal hypolipidemic response to dietary PUFAs and other stimuli (Fig. 1 and refs. 8–11 and 14) but unfortunately may also contribute to the aggregation, retention, and atherogenicity of apoB-lipoproteins once they reach the vessel wall (2, 4). Defects in PERPP, then, might produce significant health risks through hepatic oversecretion of highly atherogenic apoB-lipoproteins, a key feature of the metabolic syndrome.
Taken together, the data here and in previous reports suggest that under typical conditions, apoB escapes ERAD (8), but then autophagy serves as a quality control process for handling adverse modifications of apoB and VLDL that occur late in the secretory pathway as consequences of the normal processing of PUFAs and their intracellular peroxidized byproducts. Secretory control via PERPP/autophagosomes likely represents a key component of normal and pathologic regulation of plasma apoB levels, given that PUFAs are a major component of the heart-healthy lifestyle owing to their lipid-lowering properties (40) and that many diseases of apoB oversecretion have become a major health problem (6).
Experimental Procedures
General Materials.
Male Sprague–Dawley rats were purchased from The Jackson Laboratory. Unless otherwise specified, reagents were from Sigma. [35S]methionine/cysteine ([35S]met/cys) was from Perkin–Elmer. Collagenase was from Worthington Biochemical. Rabbit polyclonal antiserum against rat apoB was developed in the authors' laboratory. Complete protease inhibitor mixture with EDTA was from Roche. Autofluor was from National Diagnostics. McA cells were obtained from the American Type Culture Collection.
Cell Culture.
Rat primary hepatocytes were prepared as in ref. 8, using a protocol approved by the Institutional Animal Care Committee. Hepatocytes were plated in six-well plates precoated with type I collagen and maintained in Waymouth's medium (Waymouth's MB 752/1 containing 1% streptomycin/penicillin, 1% l-glutamine, 0.2% BSA, and 0.1 nM insulin). After 12–14 h incubation, the primary hepatocytes were subjected to study protocols.
McA cells were plated in six-well plates or 100-mm Petrie dishes and maintained in complete DMEM, which was changed every 3 days. When cells reached ≈90% confluence, they were used for the experiments.
Culture media were supplemented with fatty acids, typically at a final concentration of 0.6 mM given as a complex with BSA (fatty acid:BSA molar ratio = 5:1), as described (e.g., ref. 48). PUFAs to stimulate PERPP were DHA, ARA, and LNA (8, 9). OA complexed to BSA, or BSA (0.12 mM) without fatty acids, served as controls. Unless otherwise noted, cells were incubated at 37°C in 5% CO2. For inhibition of protein trafficking, the temperature in some experiments was lowered to either 20°C (22, 23) or 24°C.
Metabolic Labeling.
Steady-state labeling of hepatic proteins.
After the pretreatments described in the figure legends, endogenous proteins were metabolically labeled to steady-state (49) by incubating hepatocytes at 37°C for 3 h in met/cys-free DMEM-complete (met/cys-free high-glucose DMEM with 0.5% FBS, 0.5% horse serum, 1% streptomycin/penicillin, and 1% l-glutamine) supplemented with 100 μCi [35S]met/cys per ml of medium and various compounds indicated in Results.
Pulse–chase studies.
After the pretreatments described in the figure legends, cellular proteins were metabolically labeled with met/cys-free DMEM-complete (see above) supplemented with 200–400 μCi [35S]met/cys per milliliter of medium and various compounds indicated in Results. The temperature and duration of labeling and the chase periods varied according the experimental goals and are shown in the appropriate figure legends.
After the labeling period, the media were removed, and the cells were washed twice with PBS at the same temperature as the chase incubation and then incubated with excess unlabeled met (10 mM)/cys (3 mM) in chase medium (for primary hepatocytes, Waymouth's MB 752/1 containing 1% streptomycin/penicillin, 1% l-glutamine, 0.2% BSA, and 0.1 nM insulin; for McA cells, high-glucose DMEM containing 0.5% FBS, 0.5% horse serum, 1% penicillin/streptomycin, and 1% l-glutamine).
Immunoprecipitation, Electrophoresis, and Fluorography.
Plates or dishes of cells were placed on ice. Media were collected and supplemented with PMSF (1 mM). Cells were washed three times with cold PBS. Cell lysis buffer [10 mM PBS (pH 7.4), 125 mM NaCl, 36 mM lithium dodecyl sulfate, 24 mM deoxycholate, and 1% Triton X-100] freshly supplemented with protease inhibitor mixture and 1 mM PMSF was used to lyse cells. Both cellular lysates and media were spun at 10,000 × g for 5 min to remove debris. One percent of each lysate and medium sample was treated with trichloroacetic acid (TCA) at 4°C, and the precipitates were resuspended and counted to quantify total 35S-labeled proteins (8). The remaining lysate and medium samples were subjected to apoB immunoprecipitation.
As previously, we focused on apoB100, because it is the form of apoB that is secreted from human liver and is the predominant form of rat hepatic apoB on large, lipid-rich lipoproteins (50), the targets of PERPP (8). In addition, the effect of PUFAs on total apoB48 recovery has been small in prior and the present studies (data not shown), owing to its under-representation in large lipid-rich particles (8, 48). As indicated in Results, depending on the experiment, either the total amount of labeled apoB100 (intracellular plus secreted) or the separate recoveries from cells or conditional media were determined by immunoprecipitation followed by SDS/PAGE/fluorography/densitometry, as described in ref. 9.
Detection of MDA-Modified apoB.
Cellular content of lipid peroxides was determined by measuring thiobarbituric acid-reactive substances (TBARS) as described in ref. 9. Results are expressed as malondialdehyde (MDA) equivalents normalized to cellular protein mass (DC protein assay kit; Bio Rad).
For immunodetection of oxidatively modified intracellular apoB, cells were incubated for 4 h in DMEM supplemented with 0.5% FBS, 0.5% horse serum, 1% l-glutamine, and 1% penicillin/streptomycin with 0.12 mM BSA or 0.6 mM DHA. After the preincubation, steady-state labeling with [35S]met/cys was done as above but in the presence of the corresponding BSA and DHA treatments.
At the end of the labeling period, treatment media were collected, cells were washed twice with PBS, and lysed. 35S-labeled apoB was then immunoprecipitated from cell lysates.
Aliquots of immunoprecipitates (normalized according to total TCA-precipitable counts) were resolved on two separate gels. One gel was subjected to fluorography, and the other was used to transfer proteins to Polyscreen PVDF membrane (Perkin–Elmer). Rabbit anti-MDA (Alpha Diagnostic) primary antibody (1:1,000) was followed by anti-rabbit IgG-peroxidase (1:10,000; Sigma). Detection was by Western Lightning Chemiluminescence reagent plus kit (Perkin–Elmer). MDA-Ovalbumin (Alpha Diagnostic) was the positive control.
Colocalization Studies.
McA cells were transfected with an expression plasmid for LC3-GFP [a generous gift from T. Yoshimori (National Institute of Genetics, Mishima, Japan)] using the TransIT reagent (Mirus Bio), following the manufacturer's protocol. After 24 h, cells were then replated on collagen-coated chambered slides. After 2 h to allow attachment, cells were then incubated for 4 h at 37°C with BSA complexed to either OA (0.6 mM) or DHA (0.6 mM). Cells were fixed with 3.3% paraformaldehyde and then permeabilized with 0.1% Triton X-100. Cells were incubated with rabbit anti-rat apoB serum (1:250). After washing, cells were next incubated with Texas red-conjugated goat anti-rabbit secondary antibody (1:250; Jackson ImmunoResearch). Nuclei were stained with DAPI (Vector Laboratories). Cells were imaged with a Zeiss 510 Meta laser scanning confocal microscope with a 63× Plan-Apochromat NA 1.4 objective and the digitized files processed by Adobe Photoshop 5.5.
Statistical Analyses.
Unless otherwise noted, quantitative results are displayed as mean ± SEM, n = 4. For most comparisons, two-tailed t tests were used. For comparisons involving several groups, ANOVA was initially used, and, when indicated, pairwise comparisons of each experimental group versus the control group were performed using the Dunnett q′x2032 statistic (51).
Acknowledgments.
We thank Ms. Cristina Villagra for technical assistance, Dr. Chris Cardozo (Mount Sinai School of Medicine, New York, NY) for advice on protease inhibitors, and Dr. Ana Maria Cuervo (Albert Einstein College of Medicine, Bronx, NY) for advice and protocols in the area of autophagy. This work was supported by National Institutes of Health Grants HL58541 (to E.A.F.), HL73898 and HL56984 (K.J.W.), and DK27083 (F.R.M.).
Footnotes
The authors declare no conflict of interest.
This article contains supporting information online at www.pnas.org/cgi/content/full/0707460104/DC1.