Abstract
In mature adipocytes, triglyceride is stored within lipid droplets, which are coated with the protein perilipin, which functions to regulate lipolysis by controlling lipase access to the droplet in a hormone-regulatable fashion. Adipocyte differentiation-related protein (ADRP) is a widely expressed lipid droplet binding protein that is coexpressed with perilipin in differentiating fat cells but is minimally present in fully differentiated cultured adipocytes. We find that fibroblasts ectopically expressing C/EBPα (NIH-C/EBPα cells) differentiate into mature adipocytes that simultaneously express perilipin and ADRP. In response to isoproterenol, perilipin is hyperphosphorylated, lipolysis is enhanced, and subsequently, ADRP expression increases coincident with it surrounding intracellular lipid droplets. In the absence of lipolytic stimulation, inhibition of proteasomal activity with MG-132 increased ADRP levels to those of cells treated with 10 μm isoproterenol, but ADRP does not surround the lipid droplet in the absence of lipolytic stimulation. We overexpressed a perilipin A construct in NIH-C/EBPα cells where the six serine residues known to be phosphorylated by protein kinase A were changed to alanine (Peri A Δ1–6). These cells show no increase in ADRP expression in response to isoproterenol. We propose that ADRP can replace perilipin on existing lipid droplets or those newly formed as a result of fatty acid reesterification, under dynamic conditions of hormonally stimulated lipolysis, thus preserving lipid droplet morphology/structure.
TRIACYLGLYCEROL-CONTAINING lipid droplets in adipocytes serve as the principal long-term energy storage depot of animals. There is increasing recognition that these lipid droplets are a type of metabolic organelle that is surrounded by a phospholipid monolayer to which a number of proteins bind and participate in the regulation of lipid fuel metabolism (1). In adipocytes, the most prominent of these droplet proteins is perilipin whose tissue distribution is restricted to fat and steroidogenic cells (1–4). Perilipin is a member of the so-called PAT family of proteins (5) consisting of perilipin, adipocyte differentiation-related protein (ADRP), also called adipophilin (6, 7), and TIP47 (5, 8) as well as the more recently described S3–12 protein (9, 10). In contrast to perilipin and S3–12, ADRP and TIP47 have a broad tissue distribution. ADRP was originally identified as an mRNA transcript whose levels increased 100-fold early in the differentiation program of an adipocyte cell line, suggesting possible adipocyte-specific expression (11). Subsequent studies revealed ADRP to be expressed widely in a variety of cells, where it constitutes the major lipid droplet binding protein (7, 12), except for adipocytes, where perilipin is predominant. In addition to decorating lipid droplets, TIP47 has been postulated to have a role in intracellular membrane trafficking (13) and S3–12, the least homologous PAT family member, associates with adipocyte lipid droplets early in their formation in adipocytes (14).
The major role of perilipin is to regulate lipolysis by virtue of its ability to protect the intracellular lipid droplets from neutral lipases within the cell under basal conditions (15–17). In response to lipolytic stimuli, perilipin becomes phosphorylated on six serine residues and recruits hormone-sensitive lipase (HSL), and probably other lipases, to the surface of the lipid droplet so that the stored lipids will undergo breakdown and release (2, 17–19). The exact mechanism by which hormonally mediated perilipin phosphorylation facilitates access of lipases such as HSL to the lipid droplet remains unclear.
Interestingly, in 3T3-L1 cultured murine adipocytes, ADRP mRNA expression continues to rise as differentiation proceeds, whereas protein levels decrease and are hardly detectable by d 8 of differentiation (6, 7). This is in contrast to perilipin, in which both the mRNA and protein levels increase throughout the differentiation process (7). As perilipin protein levels increase, ADRP levels decrease and the protein composition surrounding the lipid droplets, as detected by immunofluorescence, switches from ADRP, originally observed as a punctuate pattern, to perilipin. This switch occurs around d 3 of differentiation, at which point both ADRP and perilipin surround lipid droplets but by d 5 all of the lipid droplets are surrounded by perilipin consistent with the lack of ADRP expression (7). Perilipin has been shown to be stabilized post translationally because, upon incubation of adrenal cortical cells with fatty acids, the perilipin protein levels increase 6-fold, whereas the mRNA levels remain the same (20). In addition, the loading of perilipin mRNA onto free polysomes occurs irrespective of the fatty acid loading, demonstrating that this increase in protein is a posttranslational stabilization effect (20). There is evidence suggesting ADRP may also be stabilized posttranslationally by increased intracellular fatty acid/triacylglycerol levels in a manner similar to perilipin because its expression is stabilized by exogenous fatty acid treatment of cells or by factors that increase the intracellular lipid content such as etomoxir, an inhibitor of carnitine palmitoyl-transferase 1 (16, 21, 22).
There appears to be some form of competition for lipid droplet association that exists between perilipin and ADRP as supported by various lines of evidence. The N-terminal 100 amino acids of ADRP are involved in lipid droplet binding (23), and this region displays 65% homology with the same region of perilipin (3), suggesting a common mode of binding. As adipocytes differentiate, perilipin protein levels increase above those of ADRP, thus preventing its droplet binding and facilitating its degradation (7). Also, perilipin null mice have ADRP coating the lipid droplets of their adipocytes (24). Finally, adenoviral-mediated ectopic expression of perilipin in lipid-loaded NIH-3T3 cells or Chinese hamster ovary cells decreased the protein expression of ADRP dramatically (16, 17). Thus, as perilipin protein levels increase above those of ADRP, it may preclude ADRP from binding to the lipid droplets thus resulting in its destabilization.
To get a better understanding of the relationship between ADRP and perilipin, we used an NIH-3T3-derived adipocyte cell line expressing the adipogenic transcription factor C/EBPα (CCAAT/enhancer binding protein) as a result of retroviral-mediated infection (25, 26). Unlike the situation in 3T3-L1 adipocytes, both ADRP and perilipin protein levels increase upon the induction of differentiation in these cells (Fig. 1), thus allowing more facile detection of the former. We have used this engineered cell line to dissect and understand aspects of insulin-stimulated glucose transport, and here we use them to study the dynamics of protein association with lipid droplets.
NIH-C/EBPα Cells Express HSL, Perilipin, and ADRP and Exhibit Hormone-Stimulated Lipolysis A, On the indicated days after induction of differentiation (see Materials and Methods), whole-cell extracts (150 μg) were prepared and resolved by SDS-PAGE (10% polyacrylamide). Western blotting was performed with the indicated antibodies, and detection was achieved with horseradish peroxidase-conjugated secondary antibodies and chemiluminescence. B, Isoproterenol stimulates lipolysis in NIH-C/EBPα adipocytes. Differentiated C/EBPα-adipocytes (d 10) were incubated with 10 μm isoproterenol for the times indicated. Aliquots were removed at these times and glycerol release was measured as described in Materials and Methods and normalized to total protein content. Fold-stimulation is the ratio of the values in response to 10 μm isoproterenol vs. those without agonist measured at the indicated times. A, Representative of three such experiments; B, mean ± se of three such measurements.
RESULTS
NIH-C/EBPα Adipocytes Exhibit Hormonally Stimulated Lipolysis
We have previously shown that the NIH-3T3 fibroblast cells ectopically expressing the adipogenic transcription factor C/EBPα behave similarly to 3T3-L1 adipocytes (27) by becoming laden with lipid droplets and exhibiting insulin-dependent glucose transport as they differentiate (25, 26, 28). Figure 1A shows that these cells express hormone-sensitive lipase, perilipin, and ADRP in a differentiation-dependent fashion. This could be expected for the components of hormonally regulated lipolysis, i.e. HSL and perilipin, but it was a surprise for ADRP whose protein levels had been shown to decrease to negligible levels upon differentiation of 3T3-L1 cells (7). In addition, isoproterenol treatment results in increased glycerol release in NIH-C/EBPα adipocytes in a time-dependent (Fig. 1B) and dose-dependent (data not shown) manner. These data confirm that the proteins and signal transduction mechanisms, which are required for hormonally regulated lipolysis, are present in NIH-C/EBPα adipocytes.
Stimulation of Lipolysis in NIH/C/EBPα Adipocytes Results in a Further Increase in ADRP Protein Expression
Cells were exposed to isoproterenol, and as expected, perilipin is seen to be phosphorylated in response to the agonist as shown by its decreased mobility in gel electrophoresis (Fig. 2). Within the first 15 min of stimulation, perilipin appears to be maximally phosphorylated, and this effect is protein kinase A (PKA) dependent because incubation of the cells with H89, a PKA inhibitor, prevents this phosphorylation from occurring (Fig. 2). This phosphorylation is maintained for the 2-h duration of isoproterenol treatment, as is the level of perilipin protein expression. Interestingly and perhaps surprisingly, the level of ADRP is increased by isoproterenol at 60 and 120 min after agonist treatment. Evidently, there is some tonic level of adenylate cyclase activity in these cells as H89 slightly suppresses basal ADRP expression (see also Fig. 4A). Stimulation of adipocytes with either a β3-adrenergic receptor-selective agonist or dibutyryl cAMP (a cAMP donor) also results in phosphorylation of perilipin and increased expression of ADRP, further confirming that this effect is a result of PKA stimulation (data not shown). The time course for the increase in ADRP expression does not correlate with perilipin phosphorylation because the latter is maximal at 15 min of stimulation, whereas 1 h is required to significantly alter ADRP amounts.
Isoproterenol Treatment Enhances the Expression of ADRP in NIH-C/EBPα Cells On d 10 of differentiation, the NIH-C/EBPα-adipocytes were treated with (+) or without (−) 100 μm H89 for 30 min. Cells were then stimulated with 10 μm isoproterenol (ISO) for the indicated times in the presence (+) or absence (−) of H89. Whole cell extracts were collected as described in Materials and Methods and equal amounts of protein were resolved by SDS-PAGE and Western blotting analysis. Immunoblots were performed with the indicated antibodies and proteins were detected as in Fig. 1A. The blot shown is representative of three independent experiments.
Increased ADRP Protein Expression Is Coincident with ADRP Binding to Lipid Droplets
As previously noted, ADRP binds to lipid droplets of adipocytes and nonadipocyte cells. We therefore determined ADRP localization in lipolytically stimulated NIH-C/EBPα cells by indirect immunofluorescence. As shown in Fig. 3A, 2 h of isoproterenol treatment resulted in markedly increased staining of ADRP surrounding the intracellular lipid droplets, whereas untreated cells or cells exposed to both isoproterenol and the PKA inhibitor H89 showed no significant staining of lipid droplets. In unstimulated cells (Fig. 3A-a), ADRP is presumably diffusely dispersed throughout the cell and does not produce a signal associated with any discernable cellular structure.
Isoproterenol Exposure Enhances ADRP Binding to Lipid Storage Droplets in NIH-C/EBPα Adipocytes A, On d 10 of differentiation, NIH-C/EBPα cells were treated with (c) or without (a, b, d) 100 μm H89 for 30 min. Cells were then stimulated with 10 μm isoproterenol (b and c) for 2 h in the presence or absence of H89. Cells were fixed for 15 min in 3.7% formaldehyde and treated as described in Materials and Methods. Panels a–c represent cells incubated with anti-ADRP antibody (1:100) for 2 h followed by incubation with rabbit Cy3-conjugated secondary antibody. Arrows indicate ADRP-decorated lipid droplet. Cells stained only with secondary antibody (d) were used to establish the fluorescence exposure settings and all images were analyzed at the same voltage and aperture settings as used for controls. The fluorescent staining profiles were examined using a Zeiss 510 confocal laser-scanning microscope. Images were processed using LSM 5 Image software. B, NIH-C/EBPα cells were plated on 35-mm glass bottom microwell dishes (MatTek) and grown and differentiated as above. On d 10, they were fixed and processed as described in Materials and Methods. The top three panels represent cells under basal condition whereas the bottom panels represent cells stimulated with 10 μm isoproterenol. The arrows indicate some of the lipid droplets with colocalized perilipin and ADRP. The fixed cells were then incubated with anti-ADRP antibody and/or antiperilipin antibody as indicated followed by antirabbit Alexa647 or antiguinea Alexa488 secondary antibodies, respectively. The results are representative of three or more independent experiments.
To determine whether the ADRP produced after lipolysis was replacing perilipin on existing droplets or, perhaps, was binding to droplets newly formed from lipolysis and reesterification, we performed double-label immunofluorescence as shown in Fig. 3B. ADRP and perilipin colocalized on many of the droplets, and some droplets remained surrounded by perilipin without a detectable ADRP signal. These data suggested that ADRP can replace perilipin mainly or exclusively on existing lipid droplets. Similar results were seen (data not shown) in the presence of triacsin C, an inhibitor of esterification, consistent with the colocalization of perilipin and ADRP on existing droplets.
Proteasome Inhibition with MG132 Results in Increased ADRP Expression in the Absence of PKA Activation
Treatment of cells with the proteasome inhibitor MG132 for 2 h resulted in unchanged perilipin levels, and after isoproterenol treatment for 2 h, perilipin is still maximally phosphorylated as demonstrated by its decreased mobility via gel electrophoresis (Fig. 4A). On the other hand, ADRP levels increase in response to MG132 treatment with or without isoproterenol exposure. Ordinarily, under these conditions ADRP levels remain decreased, suggesting that the increased levels after lipolytic stimulation are the result of increased stabilization of ADRP.
Proteasome Inhibition by MG132 Increases the Expression of ADRP, but Its Binding to Lipid Droplets Requires Lipolysis A, On d 10 of differentiation NIH-C/EBPα adipocytes were treated with (+) or without (−) 10 μm MG132 (proteasome inhibitor) for 2 h. During the last 30 min of the MG132 pretreatment, cells were incubated in the presence (+) or absence (−) of 100 μm H89. Cells were then incubated with 10 μm isoproterenol for zero or 120 min as indicated while still in the presence of the indicated inhibitors. Whole cell extracts were prepared and equal amounts of protein were resolved by SDS-PAGE followed by Western blotting analysis as in the prior figures. B, On d 10 of differentiation cells were treated as described in Fig. 4A. They were fixed and subjected to confocal microscopy as in the previous figures with arrows indicating some of the ADRP-containing droplets. The data are representative of three or more independent experiments.
Despite these increased levels of ADRP with MG132 treatment alone, there is no detectable lipid droplet binding indicating that proteasome inhibition has the ability to stabilize the protein independent of lipolysis and lipid droplet binding. This is illustrated in the immunofluorescence experiment of Fig. 4B, where cells have been treated with isoproterenol for 2 h in the presence or absence of the PKA inhibitor H89, as well as the proteasome inhibitor MG132. Under conditions that promote lipolysis (−H89), ADRP surrounds lipid droplets as shown by the lipid ring-like staining pattern (Fig. 4B). However, under conditions where lipolysis is inhibited (+H89) and ADRP expression is increased by MG132 treatment (Fig. 4A), there is no signal surrounding the lipid droplets (Fig. 4B).
Overexpression of a Nonphosphorylatable Perilipin Construct Results in Decreased ADRP Levels
As discussed in the introductory text, perilipin phosphorylation is a requisite for hormonally stimulated lipolysis by virtue of its ability to recruit HSL to the droplet where it acts. Therefore, we decided to see whether overexpressing a mutant perilipin construct lacking all possible PKA phosphorylation sites [Peri A Δ1–6 (18)] could block the increase in ADRP expression in response to isoproterenol. The hypothesis is that the Peri A Δ1–6 will coat the lipid droplet but not allow access of HSL because it cannot be phosphorylated. Therefore, lipolysis will not occur and ADRP will be unable to increase in amount and decorate the lipid droplet. Figure 5A demonstrates that the NIH-C/EBPα adipocytes overexpressing the mutant perilipin Peri A Δ1–6 express lower ADRP levels in response to isoproterenol as compared with control cells not infected with adenovirus and cells infected with green fluorescent protein (GFP) adenovirus. This would suggest that the phosphorylation of perilipin, and thus its ability to mediate HSL binding and enhance lipolysis, is necessary to an increase in ADRP expression and binding to the droplets.
Expression of Peri A Δ1–6 Blocks the Isoproterenol-Mediated Increase in ADRP Levels On d 6 of differentiation, NIH-C/EBPα adipocytes were infected with adenoviral Peri A Δ1–6, adenoviral GFP, or control as described in Materials and Methods. On d 9 of differentiation, cells were treated with 10 μm isoproterenol or not for 120 min. At the indicated time, whole cell extracts were prepared and equal amounts of protein were resolved by SDS-PAGE followed by Western blotting analysis as in the prior figures.
In further support of this notion, we created adipocytes from embryonic fibroblasts derived from perilipin null mice by means of retroviral driven peroxisome proliferator-activated receptor (PPAR) γ expression (29–31). After their differentiation, perilipin wild-type (Peri A) and Peri A Δ1–6 constructs were expressed by means of adenoviral infection. When a GFP-containing vector was used, forskolin-stimulation of cAMP levels results in no change in ADRP levels. Expression of wild type perilipin shows that ADRP levels decrease under basal conditions as expected and increase as a result of forskolin treatment but Peri A Δ1–6 suppresses ADRP expression basally and after forskolin exposure (Fig. 6), consistent with the results from the NIH-C/EBPα cells.
Expression of Peri A Δ 1–6 Blocks the Forskolin-Mediated Increase in ADRP Levels in Embryonic Fibroblasts Derived from Perilipin Null Mice Perilipin null adipocytes were infected with adenoviral (Ad) GFP, adenoviral Peri A, or adenoviral Peri A Δ1–6 on d 3 of the differentiation process. On d 10, the cells were treated with 20 μm forskolin for 2 h. Whole cell extracts were prepared and equal amounts of protein were resolved by SDS-PAGE followed by Western blotting analysis as in the prior figures.
DISCUSSION
The process of adipocyte differentiation is regulated by the hierarchical expression of a number of transcription factors that can mediate the expression of different patterns or programs of genes that are required for the metabolic processes of a mature adipocyte (30, 32). We and others have used some of these transcription factors to engineer certain fibroblast cell lines and convert them into adipocytes, and we have found that different adipogenic transcription factors can produce fat cells with different phenotypes, for example, those lacking insulin-sensitive glucose uptake due to lack of Glut4 expression (25). In the present study, we have used the model of C/EBPα-driven conversion of NIH-3T3 fibroblasts into adipocytes. These cells have insulin-responsive glucose transport, and they accumulate intracellular lipid droplets (25, 26). We show here that they express key proteins involved in lipolysis such as perilipin and HSL, and they respond to lipolytic stimuli in a time- and concentration-dependent manner (Fig. 1 and data not shown, respectively). An interesting feature of these cells is that they express significant amounts of the lipid droplet binding protein ADRP in the differentiated state under basal conditions (Fig. 1). In contrast, this protein is not significantly expressed in the basal state in primary or 3T3-L1 adipocytes. This prompted us to investigate potential changes in protein association with lipid droplets that may occur upon stimulation of lipolysis in this cell line.
Our findings can be summarized as follows. We show that ADRP expression increases in more than one adipocyte cell line, in response to lipolytic stimulation after 1–2 h of lipolysis, a time when perilipin protein levels are essentially unchanged from the basal state (Fig. 2). Indeed, perilipin phosphorylation is maximal by 15 min of adrenergic stimulation, which corresponds to increased rates of lipolysis and does not therefore correlate with the 1–2 h required for the increased levels of ADRP expression (Fig. 2). Only after prolonged lipolytic stimulation does ADRP decorate/surround lipid droplets (Fig. 3). As noted previously, ADRP mRNA is relatively abundant in adipocytes, suggesting that either a lack of translation or enhanced protein degradation accounts for the lack of ADRP. Indeed, we find that ADRP expression can be further increased in NIH-C/EBPα cells by incubation of cells with a proteasome inhibitor, MG132 (Fig. 4), suggesting that protein degradation normally accounts for the lack or decreased levels of ADRP. However, the significant basal expression of ADRP in these cells, where it does not bind to lipid droplets, may be sufficiently high to allow some of it to escape degradation. Importantly, even when ADRP levels are significantly elevated by inhibition of proteasomal activity, it stills does not bind to lipid droplets in the absence of lipolysis. We confirmed this observation by overexpressing a perilipin construct, Peri A Δ1–6, lacking all the PKA-mediated phosphorylation sites required for hormonally stimulated lipolysis, and we found that ADRP levels did not increase under these circumstances (Figs. 5 and 6). This can be explained by inhibition of hormonally regulated lipolysis due to the lack of phosphorylatable perilipin because phosphorylated perilipin is required to recruit HSL to the lipid droplet to enhance lipolysis (19). Thus, in our experiments, hormonally stimulated lipolysis is required for ADRP decoration of lipid droplets under all conditions, suggesting that upon lipolytic stimulation, ADRP is either replacing perilipin on existing droplets or is binding to droplets newly formed as a result of fatty reesterification to triglycerides.
We confirmed the former possibility by double-label immunofluorescence showing that ADRP and perilipin can bind to the same lipid droplet (Fig. 3B), although we observed some tendency of the former to be associated mainly with smaller droplets (Figs. 3 and 4). This suggested the possibility that these could represent new lipid droplets formed as a result of fatty acid glycerol reesterification. Thus, we employed triacsin C, an inhibitor of some of the acyl coenzyme A synthetase isoforms (33) required for reesterification and triglyceride formation (15) to test this possibility. We found that this inhibitor reduces ADRP expression in the basal state as well as after 2 h of isoproterenol stimulation, but the amount of ADRP associated with the lipid droplets after exposure of cells to isoproterenol, as determined by immunofluorescence, was not significantly different in the presence and absence of triacsin C (data not shown). These data are consistent with the possibility that ADRP acylation plays a role in the enhanced stability of this protein observed upon lipolysis, and indeed, this acylation has been observed in ADRP isolated from milk fat droplets (34). We are in the process of testing this possibility in the cell system described herein.
Recently, a study by Brasaemle et al. (10) showed by mass spectrometry and immunofluorescence that ADRP was associated with the lipid droplets after 2 h of isoproterenol treatment in 3T3-L1 cells. This was a somewhat surprising result because the protein was previously thought not to be expressed in this cell line. However, we show here that ADRP shows the same behavior in two additional adipocyte cell lines. Consequently, the phenomenon of ADRP association with lipid droplets in adipocytes seems to be a general one and underscores the need for such droplets to be protected under conditions of enhanced adipocyte lipolytic stimulation. This is likely to be an important physiological mechanism for the regulation of lipid metabolism.
MATERIALS AND METHODS
Reagents
The following were purchased from Sigma (St. Louis, MO); dexamethasone, 3-isobutyl-1-methylxanthine, insulin, sodium fluoride, sodium orthovanadate, fetal bovine serum (Australian origin), donkey serum, glycerol standard, puromycin, forskolin, H89, and isoproterenol. Aprotinin, leupeptin, and pepstatin A were obtained from American Bioanalytical (Natick, MA). DMEM was from Mediatech, Inc. (Herndon, VA), Lipofectamine 2000 transfection reagent and PLUS Reagent were both purchased from Invitrogen (Carlsbad, CA). Antirabbit Cy3 conjugated secondary antibody was from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). Rabbit antiguinea pig Alexa 488 and goat antirabbit Alexa 647 were purchased form Molecular Probes (Invitrogen). Triglyceride-GPO Reagent Set was purchased from Pointe Scientific Inc. (Brussels, Belgium), and BSA-Cohn Fraction V was purchased from Intergen Co. (Purchase, NY). H89 was obtained from LC Laboratories (Woburn, MA), MG-132 was purchased from Calbiochem (San Diego, CA), and Triacsin-C was from Biomol Research Laboratories Inc. (Plymouth Meeting, PA). Troglitazone was a kind gift from Dr. John Johnson (Parke-Davis/Warner Lambert, Ann Arbor, MI).
Primary Antibodies
The polyclonal antiperilipin antibody, specific for the C terminus of perilipin A was generated using a C-terminal peptide as described (18). The polyclonal anti-ADRP antibody was generated against a mouse sequence and the polyclonal anti-HSL antibody was raised against a peptide sequence of rat HSL as described in Souza et al. (16). All three of these antibodies were affinity purified and used for Western blotting at the following concentrations: perilipin (1:3000), ADRP (1:1000), and HSL (1:1500). Guinea pig anti-ADRP used for the double-label experiment of Fig. 3B was purchased from Research Diagnostics (Concord, MA) and used as directed.
Cell Culture
Murine NIH-3T3 fibroblasts ectopically expressing either PPARγ or C/EBPα were cultured, maintained, and differentiated as described previously (25, 26). Briefly, cells were plated and grown in DMEM containing 10% fetal bovine serum and 2.0 μg/ml puromycin. At 2 d after confluence (d 0), cells were induced by changing the medium to DMEM containing 10% fetal bovine serum, 0.5 mm 3-isobutyl-1-methylxanthine, 1 μm dexamethasone, 1.7 μm insulin, and 5 μm troglitazone. After 48 h, the induction media were removed and cells were maintained in DMEM containing 10% fetal bovine serum and 5 μm troglitazone. Mouse embryonic fibroblasts were obtained from perilipin null animals (Greenberg, A. S., unpublished data) and converted to adipocytes by retroviral-mediated PPARγ infection as previously described (29–31). These were cultured and differentiated in the same way as described above.
Preparation of Whole Cell Extracts
At the indicated times, cultured cells grown in 6-cm dishes were rinsed three times with PBS and then harvested in ice-cold buffer containing 50 mm Tris (pH 7.4), 100 mm NaCl, 1% Triton X-100, 1 μm pepstatin, 1 μm aprotinin, and 10 μm leupeptin. In addition, the phosphatase inhibitors sodium fluoride (20 mm) and sodium orthovanadate (2 mm) were added to the lysis buffer to detect phosphorylated proteins. Lysates were vortexed and stored at −20 C until ready for further analysis. When ready to be analyzed, samples were thawed, vortexed, and spun for 20 min at 16,000 × g in a microcentrifuge at 4 C. The supernatants were collected and the protein content was determined using the bicinchoninic acid (BCA) kit (Pierce, Rockford, IL).
Gel Electrophoresis and Immunoblotting
Proteins were separated by SDS-PAGE (acrylamide from National Diagnostics) as described by Laemmli (35) and electrophoretically transferred to polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA) in 25 mm Tris, 192 mm glycine. The membrane was then blocked with 10% nonfat dry milk in PBS containing 2% Tween 20 for 1 h at room temperature. The membranes were then incubated with the primary antibodies described above. Horseradish peroxidase-conjugated secondary antibodies (Sigma) and either an enhanced chemiluminescent substrate kit (PerkinElmer Life and Analytical Sciences, Inc., Boston, MA) or Super Signal West Femto Maximum Sensitivity Substrate kit (Pierce, Rockford, IL) was used for detection.
Lipolysis
Cells were grown to confluence in 12-well plates and induced to differentiate as described above. On the day of the experiment, adipocytes were washed three times with warm DMEM and then treated with 10 μm isoproterenol in the presence of DMEM containing 2% BSA (fatty acid free) in a total volume of 500 μl. Aliquots were removed at 1, 2, and 4 h of stimulation and the glycerol content of each of these was measured using the Triglyceride (GPO) Reagent Set (Pointe Scientific, Brussels, Belgium) and measured at 540 nm against a set of glycerol standards. Cells were then washed with cold PBS, lysed with a 1% Triton X-100 buffer, and the protein concentration was determined and used to normalize glycerol release.
Confocal Laser Scanning Microscopy
Cells were grown on two-well chamber slides (Fisher Scientific, Pittsburgh, PA) or 35-mm glass bottom microwell dishes (MatTek, Ashland, MA) to confluence and were induced to differentiate 2 d after confluence as already described. On d 10 of differentiation, adipocytes were washed three times with warm DMEM and then treated with 10 μm isoproterenol for 2 h in the presence of DMEM containing 2% BSA (fatty acid free). Cells were washed with PBS and then fixed with 3.7% formaldehyde in PBS for 15 min at room temperature. They were then washed three more times with PBS and then kept overnight at 4 C. The next day, the cells were permeabilized with 0.1% Triton X-100 in PBS for 5 min and then blocked for 1 h at room temperature in a PBS solution containing 5% donkey serum (Sigma). Staining was performed with rabbit polyclonal anti-ADRP antibody at room temperature at a dilution of 1/100 in blocking solution. Four more washes with PBS were then performed and the cells were covered in aluminum foil and incubated for 1 h at room temperature with antirabbit Cy-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc.) at a dilution of 1/200 in blocking solution. The cells were washed again with PBS and then mounted with a glycerol/PBS solution (Component A of the SlowFade Antifade Kit from Molecular Probes, Eugene, OR). For the double staining experiments, the primary antibody incubation was performed overnight at 4 C using antirabbit ADRP and antiguinea pig perilipin antibodies at a concentration of 1:100. Incubation with the conjugated secondary antibodies was performed for 1 h at room temperature using an antirabbit Cy3 and an antiguinea pig Cy2, both diluted 1:200. Cells stained with only the secondary antibody were used to determine the fluorescent exposure settings, and all experiments were analyzed at the same voltage and aperture settings as used for controls. The stained cells were observed using a Zeiss 510 confocal laser-scanning microscope (Carl Zeiss, Thornwood, NY). Images were processed using LSM 5 Image software. For the double label experiment of Fig. 3B, cells were prepared as described above, incubated with guinea pig antiperilipin (16) or rabbit anti-ADRP antibodies as described above, then incubated with the appropriate Alexa-derivatized secondary antibodies. Confocal microscopy was performed using a Leica (Exton, PA) TCS SP2 instrument in the Tufts/New England Medical Center Imaging Facility.
Adenoviral Infection
Adenovirus expressing PKA site-mutated perilipin A (Peri A Δ1–6) was generated as described (18) using the AdEasy adenoviral vector system (Stratagene, La Jolla, CA). Briefly, perilipin cDNA was subcloned into the pShuttle-CMV vector that was then linearized and cotransformed with supercoiled viral DNA plasmid pAdEasy into BJ5183 bacterial cells. The adenovirus expressing the Aequoria victoria GFP was generated by cloning the GFP from pEGFPC1 (CLONTECH, Palo Alto, CA) into pLEP-hCMV, a pLEP3 derivative as described (16). The recombinant DNAs were transfected into human embryonic kidney 293 cells in which viral assembly genes were complemented. Adenoviral amplification was performed and the adenoviruses were purified and concentrated by CsCl ultracentrifugation to 0.6 × 1012 plaque-forming units/ml. Cells were infected on either d 3 or 6 (both with similar results) by incubation with Lipofectamine and Plus Reagent (Invitrogen) for 3 h at a multiplicity of infection of 1000 plaque-forming units/cell.
Acknowledgments
We thank Dr. Stephen Farmer for valuable discussions.
This work was supported by National Institutes of Health Grants DK30425 and 56935 (to P.F.P.) and DK50647 (to A.S.G.) and grants from the American Diabetes Association, U.S. Dept of Agriculture, Agriculture Research Service, contract 52-KO6-5-10.
Abbreviations
- ADRP
Adipocyte differentiation-related protein;
- C/EBP
CCAAT/enhancer binding protein;
- GFP
green fluorescent protein;
- HSL
hormone-sensitive lipase;
- PKA
protein kinase A;
- PPAR
peroxosime proliferator-activated receptor.
