Adipose triglyceride lipase (ATGL) is the key triacylglycerol hydrolase in adipocytes. The precise mechanisms by which ATGL action is regulated by lipid droplet (LD) coat proteins and responds to hormonal stimulation are incompletely defined. By combining usage of loss- and gain-of-function approaches, we sought to determine the respective roles of perilipin 1 and fat-specific protein 27 (FSP27) in the control of ATGL-mediated lipolysis in adipocytes. Knockdown of endogenous perilipin 1 expression resulted in elevated basal lipolysis that was less responsive to β-adrenergic agonist isoproterenol. In comparison, depletion of FSP27 protein increased both basal and stimulated lipolysis with no significant impact on the overall response of cells to isoproterenol. In vitro assays showed that perilipin but not FSP27 was able to inhibit the triacylglycerol hydrolase activity of ATGL. Perilipin 1 also attenuated dose-dependent activation of ATGL by its Coactivator Comparative Gene identification-58. Accordingly, depletion of perilipin 1 and CGI-58 in adipocytes inversely affected basal lipolysis specifically mediated by overexpressed ATGL. Moreover, although depletion of perilipin 1 abolished the LD translocation of ATGL stimulated by isoproterenol, absence of FSP27 resulted in multilocularization of LDs along with increased LD presence of ATGL under both basal and stimulated conditions. Interestingly, knockdown of ATGL expression increased LD size and decreased LD number in FSP27-depeleted cells. Together, our results demonstrate that although FSP27 acts to constitutively limit the LD presence of ATGL, perilipin 1 plays an essential role in mediating the response of ATGL action to β-adrenergic hormones.
Adipose tissue functions as the primary fuel depot in vertebrates. The ability of adipose tissue to store energy as triacylglycerols (TAGs) in intracellular lipid droplets (LDs) and, on demand, to mobilize these stores via lipolysis is a highly conserved process essential for survival (1–4). Adipocytes in white adipose tissue (WAT) are characterized by large unilocular LDs that are suitable for TAG storage. During extended fasting or exercise, a majority of whole-body energy expenditure is reliant on fatty acids (FAs) derived from WAT lipolysis stimulated by catecholamines (5). In comparison, adipocytes in brown adipose tissue contain numerous smaller LDs and a large number of mitochondria along with high oxidative activity. Lipolysis in brown adipocytes feeds FAs into mitochondria locally for β-oxidation and subsequent heat production (6, 7).
Lipolysis in adipocytes is mediated by a complex proteome comprised of specific acylglycerol lipases and their regulators (1, 8). The adipose lipases include adipose triglyceride lipase (ATGL), hormone-sensitive lipase (HSL), and monoacylglycerol lipase. ATGL favors TAG substrates and catalyzes the rate-limiting first step of lipolysis. HSL, on the other hand, is believed to function as a diacylglycerol lipase in vivo (9–14). Although HSL is mainly regulated by protein kinase A (PKA) phosphorylation (15–18), ATGL's activity is coactivated by the protein comparative gene identification-58 (CGI-58) and inhibited by the protein G0/G1 switch gene 2 (19, 20).
It is well established that only mature adipocytes possess the complete LDs that support hormone-stimulated lipolysis. During adipogenic differentiation of preadipocytes, LDs emerge and enlarge through a sequence of biochemical and morphological changes (21). Notably, a series of LD surface proteins of the perilipin/adipophilin/ 47-kDa tail interacting protein (TIP47) family is seen. First, perilipin 3 (TIP47), perilipin 4 (S3-12), and perilipin 2 (adipophilin) on small nascent droplets and then perilipin 1 on large mature droplets. Research conducted in the past decade has revealed that phosphorylation of perilipin 1 by PKA plays a critical role in controlling the LD localization of HSL (15–17).
Although ATGL is not a direct substrate for PKA (9), β-adrenergic signals are known to up-regulate the enzymatic action of ATGL in adipocytes (1). An earlier study has shown that phosphorylation of Ser517 in perilipin 1 is an essential step for ATGL activation (22). Other reports have indicated a mechanism for ATGL activation in adipocytes that involves the dissociation of CGI-58 from phosphorylated perilipin 1 (23–26). Experiments reconstituting interactions using ectopically expressed proteins have demonstrated that in unstimulated cells, CGI-58 is in complex with perilipin 1 at the surface of LDs and therefore is separated from ATGL. Upon stimulation, perilipin is phosphorylated on Ser492 or Ser517 by PKA, thereby releasing CGI-58 to act on ATGL (25). Recently, others and we have demonstrated that ATGL also undergoes HSL-like LD translocation in adipocytes upon β-adrenergic stimulation (14, 20, 27). However, it remains unclear whether perilipin 1 or its phosphorylation plays an essential role in LD recruitment of ATGL.
Fat-specific protein 27 (FSP27) is one of the additional proteins discovered to be colocalized with LDs (28–30). FSP27 is a member of the cell death-inducing DNA fragmentation factor 45-like effector (CIDE) protein family, members of which share a conserved N-terminal CIDE-N domain and a C-terminal CIDE-C domain. When expressed in nonadipocyte cells, FSP27 enhanced the accumulation of TAGs and the size of LDs (31–33). Similar to what was observed in perilipin 1 null mice (34, 35), FSP27-deficient animals showed increased energy expenditure and resistance to diet-induced obesity (36, 37). Depletion of perilipin 1 or FSP27 in adipocytes both resulted in elevated basal lipolysis. However, adipocytes depleted of both perilipin 1 and FSP27 showed further increase in lipolysis compared with cells depleted of FSP27 alone (37). Thus, in spite of their functional resemblance, different mechanisms of action may be employed by perilipin 1 and FSP27 in the regulation of lipolysis and LD metabolism in adipocytes. Importantly, ablation of FSP27 but not perilipin 1 led to the dramatic appearance of smaller multilocular LDs as well as brown adipose-like properties in white adipocytes (36, 37), indicating a unique role of FSP27 in maintaining the LD integrity and adipocyte identity. New studies also showed that FSP27-mediated LD growth by promoting lipid exchange from smaller to larger LDs (38, 39).
In this study, we performed small interfering RNA (siRNA)-mediated knockdown and adenovirus-mediated overexpression experiments in 3T3-L1 adipocytes to assess the roles of perilipin 1 and FSP27 in ATGL-mediated lipolysis. The differentiated 3T3-L1 adipocytes have been used extensively as a model system for investigation of hormone-regulated lipolysis. Our results provide novel information on the molecular mechanisms governing the enzymatic action of ATGL in adipocytes.
Materials and Methods
Antibodies and reagents
The rabbit polyclonal antibodies against ATGL (catalog no. 2138) and HSL (catalog no. 4107) were from Cell Signaling (Beverly, MA). Monoclonal β-actin antibody (catalog no. A1978) was obtained from Sigma-Aldrich (St. Louis, MO). Polyclonal anti-CGI-58 antibody (catalog no. 12201-1-AP) was purchased from Proteintech Group, Inc. (Chicago, IL). The goat polyclonal antibody against perilipin 1 (catalog no. ab616682) was purchased from Abcam, Inc. (Cambridge, MA). The rabbit polyclonal antibody against FSP27 was generated as described previously (32). Horseradish peroxidase-linked secondary antibodies were from Pierce Chemical Co. (Rockford, IL). The protease inhibitor minitablets (catalog no. 11 836 170 001) were obtained from Roche Diagnostics (Indianapolis, IN). Lipolysis assay kit (catalog no. LIP-1-NC-L1) was purchased from Zenbio (Research Triangle Park, NC). 3H-labeled triolein was from PerkinElmer Co. (Waltham, MA). Reagents for tissue culture were obtained from Invitrogen (Carlsbad, CA). Isoproterenol, insulin, dexamethasone, and 3-isobutyl-1-methylxanthine were purchased from Sigma-Aldrich.
RNA extraction, PCR cloning of cDNA, and site-directed mutagenesis
Total RNA was prepared from mouse 3T3-L1 adipocytes using the RNeasy Mini kit (QIAGEN, Valencia, CA) according to the manufacturer's instruction. cDNA was prepared from mRNA using SuperScript Reverse Transcriptase protocol (Invitrogen). The sequences containing the complete open reading frame of mouse ATGL, CGI-58, perilipin 1, and FSP27 were amplified by PCR using ultra Pfu DNA Polymerase Mix (Stratagene, La Jolla, CA). The PCR products containing CGI-58, perilipin, and FSP27 cDNA were cleaved by BamHI/XhoI, and the PCR product containing ATGL was digested by BglII/XhoI. The digested products were purified and ligated to BamHI (5′) and XhoI (3′) sites of the eukaryotic expression vector pRK7.
Cell culture
HeLa cells were obtained from American Type Culture Collection (Manassas, VA) and maintained in high-glucose DMEM supplemented with 10% fetal bovine serum (FBS). Mouse 3T3-L1 preadipocytes were maintained in high-glucose DMEM supplemented with 10% newborn calf serum, 100 U/ml penicillin G sodium, and 100 μg/ml streptomycin sulfate. Differentiation to adipocytes was induced by treatment of postconfluent cells with 10% FBS, 1 μg/ml insulin, 1 μm dexamethasone, and 0.5 mm 3-isobutyl-1-methylxanthine. The differentiation medium was withdrawn 3 d later and replaced with medium supplemented with 10% FBS and 1 μg/ml insulin. After 2 d in insulin-containing medium, the cells were then cultured in DMEM containing 10% FBS.
Transient expression of HeLa cells
Transient transfection was performed using Lipofectamine 2000 reagent according to the manufacturer's instructions. Briefly, 5 μg of each DNA construct were transfected into cells cultured at subconfluent density in 6-cm dishes. Cells were extracted 16–18 h after transfection for immunoblotting analysis and TAG hydrolase activity measurement.
Cell lysis and immunoblotting
For immunoblotting, cells were washed twice with ice-cold PBS and were lysed at 4 C with a buffer containing 50 mm Tris-HCl (pH 8.0), 135 mm NaCl, 10 mm NaF, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate (SDS), 0.5% sodium deoxycholate, 1.0 mm EDTA, 5% glycerol, and protease inhibitors (1 minitablet per 7 ml of buffer). The lysates were clarified by centrifugation at 10,000 × g for 10 min and then mixed with an equal volume of 2× SDS sample buffer. The solubilized proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes. Individual proteins were blotted with primary antibodies at appropriate dilutions. Peroxide-conjugated secondary antibodies were incubated with the membrane at a dilution of (1:5000). The signals were then visualized by enhanced chemiluminescence (Amersham ECL Western Blotting Detection Reagents; GE Healthcare, Princeton, NJ).
siRNA-mediated gene knockdown
For perilipin 1, FSP27, and CGI-58 knockdown in 3T3-L1 adipocytes, 4 nmol oligonucleotides were delivered into cells 7 d after differentiation via electroporation using Bio-Rad Gene Pulser II (Bio-Rad, Hercules, CA) at 950 μF and 160 V as described previously (40). After 3 d of incubation, cells were processed for designated assays.
The following double-stranded stealth siRNA oligonucleotides (Invitrogen) were used: set 1 for mouse FSP27, sense 5′-GCACAAUCGUGGAGACAGAAG-3′ and antisense 5′-UAUUCUUCUGUCUCCACGAUU-3′; set 1 for mouse perilipin 1, sense 5′-CAGGAACAGCAUCAGUGUGCCCAUU-3′ and antisense 5′-AAUGGGCACACUGAUGCUGUUCCUG-3′; set 1 for mouse CGI-58, sense 5′-CCCAAGUGGUGAGACAGCUUUCAAA-3′ and antisense 5′-UUUGAAAGCUGUCUCACCACUUGGG-3′; and set2 for mouse CGI-58, sense 5′-GGAUAGGUGGCUUGCAUCCUGACAU-3 and antisense 5′-AUGUCAGGAUGCAAGCCACCUAUCC-3′. Control oligonucleotides with comparable GC content were also from Invitrogen.
Adenoviral infection of adipocytes
Recombinant adenovirus (dE1/dE3) containing murine ATGL cDNA under the control of a cytomegalovirus promoter was custom generated by Vector Biolabs (Philadelphia, PA). A null adenovirus (Ad-null) was also obtained for use in control experiments. Infection of differentiated adipocytes was performed 1 d after electroporation with siRNA by using a previously published method (41) with minor modifications. For infection of 3T3-L1 adipocytes, Ad-G0/G1 switch gene 2, Ad-ATGL, Ad-CGI-58, or Ad-null was diluted to 3 × 107 plaque-forming unit/ml in serum-free Opti-MEM containing 5.0 μg/ml polybrene and preincubated for 45 min at room temperature. Adipocytes plated in six-well dishes were washed once with PBS, and 1.2 ml of preincubated virus mix were added to each well. Cells were centrifuged at 800 × g for 1 h at room temperature followed by incubation at 37 C. Fresh DMEM containing 10% FBS (1.0 ml/well) was added after 4 h, and cells were incubated for 48 before use.
In vitro transcription-and-translation expression
In vitro transcription/translation was carried out by using TNT SP6 High-Yield Protein Expression System (Promega, Madison, WI) according to the manufacture's instruction. Specifically, 10 μg of total plasmid DNA per reaction were used to produce ATGL and perilipin 1 for TAG hydrolase activity assays.
Assay for TAG hydrolase activity
HeLa cells in 6-cm dishes or 3T3-L1 adipocytes in six-well dishes were washed twice in ice-cold PBS and then lysed on ice by sonication in 0.5 ml of cell extraction buffer [0.25 m sucrose, 1 mm EDTA, 1 mm dithiothreitol, and protease inhibitors (1 minitablet per 7 ml of volume)]. The cell extracts were clarified by centrifugation at 1000 × g for 15 min. The TAG hydrolase activity against 3H-labeled triolein was measured as described previously (9, 19) by mixing 0.1 ml of extracts with 0.1 ml of substrate solution. Statistical analysis was determined by Student's t test or one-way ANOVA.
Assay for lipolysis
Lipolysis was measured as the rate of glycerol and free fatty acid (FFA) release by using 3T3-L1 Lipolysis Assay kit from Zenbio. Ten days after differentiation induction, adipocytes in six-well dishes were washed twice with serum- and phenol red-free DMEM containing 2% FA-free BSA and then incubated in 2.5 ml of the same medium in the presence or absence of 10 μm isoproterenol. Aliquots of the culture medium were collected after 1 h. The amounts of glycerol and FFA released were determined according to the manufacturer's instructions. Lysates were then prepared from the remaining cells, and protein concentrations in the lysates were determined and used to normalize the lipolytic signals. Statistical analysis was determined by Student's t test or one-way ANOVA.
LD fractionation
LDs were isolated from 3T3-L1 adipocytes as describe previously (42). Briefly, cells were collected in ice-cold PBS followed by centrifugation for 10 min at 1000 × g, 4 C. Cell pellets were resuspended in hypotonic lysis medium (HLM) buffer [20 mm Tris-Cl (pH 7.4), 1 mm, 0.5 mm EDTA, 10 mm sodium fluoride, and protease inhibitors] and homogenized by 10 gentle strokes with a hand-driven pestle. The homogenates were centrifuged for 10 min at 1000 × g, 4 C. The supernatant was collected, and the floating lipid layer was transferred into a separate ultracentrifugation tube; 1/3 volume of ice-cold HLM containing 60% sucrose (final 20% sucrose) was added, and LDs were mixed thoroughly by gentle pipetting. The mixture was then overloaded sequentially with HLM containing 5% sucrose and HLM containing no sucrose. After centrifugation for 30 min at 28,000 × g, 4 C, the floating LD fraction was collected by using a Beckman tube slicer (Beckman Coulter, Inc., Brea, CA). Equal volume of 10% SDS was added, and proteins associated with LDs were solubilized by incubation for 2 h at 37 C in a sonicating water bath.
Live cell microscopy
Electroporated 3T3-L1 adipocytes were plated at 50–70% density in 3-cm glass bottom dishes precoated with collagen. After incubation in DMEM containing 10% FBS for 72 h, cells were labeled with 0.5 μg/ml Nile red dye in the same medium for 15 min. After washing twice with phenol red-free DMEM, images of LDs in live cells were captured on a Zeiss LSM 510 inverted laser fluorescence confocal microscope (Zeiss, Oberkochen, Germany).
Results
Effects of perilipin 1 and FSP27 knockdown on adipocyte lipolysis
To directly assess the roles of perilipin 1 and FSP27 in lipolysis, we introduced siRNA via electroporation to knockdown each protein in mature 3T3-L1 adipocytes. Compared with the endogenous protein levels in cells transfected with a GC-matched control siRNA, less than 15% of perilipin 1 and less than 5% of FSP27 were present in cells transfected with a pool of siRNA directed against respective proteins (Fig. 1A). In relation to control knockdown, reduction of perilipin 1 expression caused a 3.5-fold increase in FFA release and a 2.3-fold increase in glycerol release under basal conditions (Fig. 1B). In comparison, depletion of FSP27 resulted in approximately 80 and 77% increase in the basal release of FFAs and glycerol, respectively (Fig. 1B). In response to stimulation by a β-agonist isoproterenol, both FFA and glycerol releases were similar in between perilipin 1 knockdown cells and control knockdown cells. However, FSP27 knockdown cells showed a 47% increase in FFA release and a 27% increase in glycerol release over control cells (Fig. 1C). Although isoproterenol treatment produced a near 9-fold stimulation in FFA release from control and FSP27 knockdown cells, the combination of increased basal lipolysis and unchanged stimulated lipolysis resulted in only 2-fold corresponding increase with perilipin 1 knockdown (Fig. 1D). Similar effects were obtained when the glycerol release was measured. Taken together, these results indicate that FSP27 functions to suppress both basal and stimulated lipolysis without affecting the sensitivity of adipocytes to isoproterenol. Perilipin 1, on the other hand, controls the β-adrenergic response via inhibiting basal lipolysis.
Effects of perilipin 1 and FSP27 knockdown on protein expression and lipolysis in 3T3-L1 adipocytes. 3T3-L1 adipocytes were electroporated with nonspecific control siRNA (Ctrl KD), perilipin 1-specific siRNA (Peri KD), or FSP27-specific siRNA (FSP KD) oligos. A, Cells were extracted, and expression of perilipin 1, FSP27, and β-actin proteins was analyzed by immunoblotting with specific antibodies. B and C, Cells were pretreated for 0.5 h in serum-free DMEM/2%BSA and then were switched to fresh medium without (B) or with 10 μm isoproterenol (C) for another 1-h incubation. The FFA and glycerol concentrations in the medium were measured as described in Materials and Methods. The data are presented as the means ± se of three independent experiments. *, P < 0.05; **, P < 0.01. D, Degrees of stimulation were shown as fold changes in average FFA and glycerol release from untreated and isoproterenol-treated cells.
Effects of perilipin 1 and FSP27 on ATGL activity
Because ATGL is the key TAG hydrolase in adipocytes, we next evaluated the direct effects of perilipin 1 and FSP27 on the enzymatic activity of ATGL. To this end, ATGL, perilipin 1, and FSP27 were expressed singly in HeLa cells (Fig. 2, A and C). The hydrolase activity against a triolein substrate was then measured in the combined cell extracts that contain ATGL along with either perilipin 1 or FSP27. Compared with the extracts of nonexpressing control cells, inclusion of ATGL-containing extracts resulted in a near 4-fold increase in the TAG hydrolase activity (Fig. 2, B and D). Interestingly, the addition of perilipin 1-containing extracts caused an approximately 60% decrease in the TAG hydrolase activity derived from ATGL (Fig. 2B). In contrast, no such effect was observed when FSP27 extracts were added (Fig. 2D). Therefore, the presence of perilipin 1 but not FSP27 substantially inhibits the enzymatic activity of ATGL.
Perilipin 1 inhibits the TAG hydrolase activity of ATGL. A–D, Extracts of HeLa cells singly transfected with empty vector, ATGL, perilipin 1, FSP27, or CGI-58 were mixed in various combinations. Proteins in parallel mixtures were revealed by immunoblotting using specific antibodies (A and C). The TAG hydrolase activity was determined against 3H-triolein as described in Materials and Methods (B and D). E, ATGL and perilipin 1 produced by in vitro transcription/translation were mixed and analyzed by immunoblotting with anti-ATGL and antiperilipin 1 antibodies. F, The TAG hydrolase activity in the in vitro transcription/translation mixtures was measured as in B and D. All data are shown as mean ± se and represent three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Ctrl, Control; KD, knockdown; Peri, perilipin; CGI, CGI-58; NS, nonsignificant; FSP, FSP27.
Because ATGL is activated by CGI-58, we next investigated whether perilipin 1 could affect the TAG hydrolase activity of ATGL in the presence of CGI-58. As shown in Fig. 2, B and D, the addition of CGI-58-containing extracts to ATGL-containing extracts caused an over 2-fold increase in the TAG hydrolase activity derived from ATGL. Such increase was significantly attenuated by the further inclusion of perilipin-containing extracts to the mixture (Fig. 2B). On the other hand, increased levels of CGI-58 still led to gradually enhanced ATGL activity in the presence of perilipin 1 (Fig. 3), although the latter caused a significant downward shift of the activation curve. Thus, the molar ratio between CGI-58 and perilipin 1 may be important in determining the enzymatic activity of ATGL.
Effect of perilipin 1 on dose-dependent activation of ATGL by CGI-58. HeLa cells were singly transfected with perilipin 1, CGI-58, or ATGL. Cell extracts containing ATGL ± perilipin 1 were mixed with increasing amounts of CGI-58 extracts and were subjected to TAG hydrolase activity assays. All data are shown as mean ± se and represent three independent experiments. **, P < 0.01; ***, P < 0.001. Peri, Perilipin 1; CGI, CGI-58.
Effects of perilipin 1 and CGI-58 depletion on ATGL-mediated lipolysis
To examine the respective impact of perilipin 1 and CGI-58 on ATGL action, we overexpressed ATGL in adipocytes where expression of endogenous perilipin 1 or CGI-58 was reduced. As described previously, we generated a recombinant adenovirus encoding mouse ATGL (Ad-ATGL). A null adenovirus (Ad-null) was used as negative control. As shown in Figs. 4A and 5A, infection with Ad-ATGL led to a profound increase in expression of ATGL when compared with the levels of endogenous protein in Ad-null-infected cells. In the control knockdown cells, overexpression of ATGL increased the TAG hydrolase activity in total cell extracts by approximately 80% (Fig. 4B). Knockdown of perilipin 1 further increased the activity derived from overexpressed ATGL (Fig. 4B), indicating an inhibitory effect of endogenous perilipin 1 on ATGL activity. Moreover, overexpression of ATGL was able to significantly increase both basal and isoproterenol-stimulated glycerol release in control knockdown cells (Fig. 4C). In cells with reduced expression of perilipin 1, overexpression of ATGL drastically enhanced basal glycerol release without significantly affecting the stimulated lipolysis (Fig. 4C).
Effects of overexpression ATGL on TAG hydrolase activity and lipolysis in adipocytes with reduced perilipin 1. 3T3-L1 adipocytes were electroporated with either control or perilipin 1-specific siRNA oligos. Twenty-four hours after electroporation, cells were infected with either a null adenovirus (Ad-null) or a virus containing ATGL cDNA (Ad-ATGL) followed by incubation for another 48 h. A, Cells were lysed, and then protein expression was analyzed by immunoblotting with antibodies specific for perilipin 1, ATGL, and β-actin. B, The TAG hydrolase activity in the cell extracts was measured using 3H-labeled triolein as substrate. The activity was normalized with the total protein levels. C, Cells were pretreated for 0.5 h in serum-free DMEM/0.5%BSA and then were switched to fresh medium without or with isoproterenol for another 1-h incubation. The glycerol concentrations in the medium were measured, and the data are presented as the means ± se of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. D, Degrees of stimulation were shown as fold changes in average glycerol release from untreated and isoproterenol-treated cells. Peri, Perilipin; Ctrl, control; KD, knockdown; NS, nonsignificant; iso, isoproterenol.
Effect of overexpression ATGL on lipolysis in the absence of CGI-58. 3T3-L1 adipocytes were electroporated with either control or perilipin 1-specific siRNA oligos. Twenty-four hours after electroporation, cells were infected with either a null adenovirus (Ad-null) or a virus containing ATGL cDNA (Ad-ATGL) followed by incubation for another 48 h. A, Cells were lysed, and then protein expression was analyzed by immunoblotting with antibodies specific for CGI-58 and ATGL. B, Cells were pretreated for 0.5 h in serum-free DMEM/0.5%BSA and then were switched to fresh medium without or with isoproterenol for another 1-h incubation. The glycerol concentrations in the medium were measured, and the data are presented as the means ± se of three independent experiments. **, P < 0.01; ***, P < 0.001. Ctrl, Control; KD, knockdown; CGI, CGI-58.
In contrast to perilipin 1 knockdown, depletion of CGI-58 resulted in a pronounced decrease in the basal glycerol release from cells infected with both Ad-null and Ad-ATGL (Fig. 5B). Interestingly, the impact of CGI-58 reduction was modest on stimulated lipolysis (Fig. 5B). Taken together, these data suggest that under basal conditions, perilipin 1 suppresses and CGI-58 activates the enzyme action of ATGL in adipocytes. In response to β-adrenergic stimulation, the further activation of ATGL-mediated lipolysis is dependent on the presence of perilipin 1 but less on that of CGI-58.
Control of ATGL localization by perilipin 1 and FSP27
In addition to their enzymatic activity, the action of TAG hydrolases in vivo is determined by their presence at the surface of LDs. To analyze the respective roles of perilipin 1 and FSP27 on the LD localization of ATGL, we performed LD fractionation experiments using 3T3-L1 adipocytes where each protein was depleted. In control knockdown cells, only low levels of ATGL were detected in the association with LDs at basal state (Fig. 6, A and B). As demonstrated previously, stimulation of lipolysis in these cells with isoproterenol promoted translocation of ATGL to LD where upward shift of perilipin 1 was also observed due to PKA phosphorylation (Fig. 6, A and B). Interestingly, knockdown of FSP27 expression clearly increased the basal level of ATGL associated with LDs, which could be further raised by isoproterenol stimulation (Fig. 6A). In comparison, FSP27 depletion had no noticeable effects on HSL LD localization under both conditions. When perilipin 1 was knocked down, the LD localization of ATGL was unaffected in the basal state (Fig. 6B). In response to isoproterenol, however, the ATGL translocation of LDs was nearly completely abolished by perilipin 1 knockdown (Fig. 6B). Collectively, these results suggest that perilipin 1 plays a critical role in facilitating the LD translocation of ATGL in response to β-adrenergic stimulation, whereas FSP27 functions to limit the LD localization of ATGL under all conditions.
Effects of FSP27 and perilipin 1 depletion on LD localization of ATGL in adipocytes. siRNA-mediated knockdown was performed by electroporating 3T3-L1 adipocytes with either control siRNA, FSP27-specific siRNA (A), or perilipin 1-specific siRNA (B). Three days later, cells were treated with or without 10 μm isoproterenol (Iso) for 30 min followed by LDs isolation as described in Materials and Methods. Total and LD-associated levels of ATGL, HSL, perilipin 1, FSP27, and CGI-58 were analyzed by immunoblotting with specific antibodies. The data shown are representative of three independent experiments. Ctrl, Control; KD, knockdown; Peri, perilipin; FSP, FSP27; CGI, CGI-58.
Regulation of LD morphology by ATGL in FSP27-depleted adipocytes
Ablation of FSP27 results in the disappearance of large central LDs along with the formation of small multilocular LDs. To determine whether enhanced lipolysis contributes to the LD multilocularization in the absence of FSP27, we stained LDs with Nile red dye in live 3T3-L1 adipocytes where FSP27 and ATGL were knocked down individually or in combination (Fig. 7A). The morphology of LDs was examined by fluorescence confocal microscopy. As shown in Fig. 7B, a typical control cell contained two or three large central LDs along with a few small LDs at the cell periphery. Cells with reduced expression of ATGL still manifested large LDs while losing a majority of the small peripheral droplets. By contrast, depletion of FSP27 alone resulted in the complete loss of large LDs and the concomitant appearance of numerous small droplets that were evenly distributed throughout the cytosol. Interestingly, although it did not alter the overall multilocular pattern of the LDs, knockdown of ATGL in FSP27-deficient cells resulted in a decrease in LD number along with an increase in the mean area of individual droplets (Fig. 7C). These results indicate that in the absence of FSP27, enhanced ATGL-mediated lipolysis substantially contributes to the production of numerous smaller LDs.
Effect of ATGL-mediated lipolysis on LD morphology in adipocytes. 3T3-L1 adipocytes were electroporated either singly with control, ATGL, or FSP27 siRNA or doubly with ATGL and FSP27 siRNA oligos. A, Cells were lysed, and then protein expression was analyzed by immunoblotting with antibodies specific for ATGL, FSP27, or β-actin. B, Fluorescence microscopy was performed to reveal LDs in live cells labeled with Nile red dye. C, Size distribution of LDs per cell after introduction of FSP27 siRNA alone or FSP27 siRNA + ATGL siRNA. Data are means ± sem (n = 4 per group). *, P < 0.05; ***, P < 0.001. Ctrl, Control; KD, knockdown; FSP, FSP27.
Discussion
The relative abundance of LD coat proteins is crucial for the regulation of TAG turnover. Here, we demonstrate that distinctive mechanisms are employed by perilipin 1 and FSP27, two LD proteins with seemingly different functions, to control adipocyte lipolysis specifically mediated by ATGL. Ablation of perilipin 1 and FSP27 in mice leads to enhanced basal adipose lipolysis that responds differently to β-adrenergic stimulation (34–37). Through analyzing the involvement of endogenous perilipin 1 and FSP27 in the control of ATGL action in adipocytes, the present study is an attempt directed to seek explanation for the aforementioned phenotypic findings obtained from genetically mutated animals.
The efficacy of ATGL in adipocytes is decided by its enzymatic activity and subcellular localization, both of which can be regulated by perilipin 1 as evidenced by loss-of-function studies. Data derived from our in vitro assays show that perilipin 1 was capable of suppressing the TAG hydrolase activity of ATGL. Experiments combining depletion of perilipin 1 with overexpression of ATGL in adipocytes have clearly demonstrated the specific inhibitory effects of perilipin 1 on ATGL-mediated lipolysis in the basal state. The biochemical mechanisms by which perilipin 1 regulates the activity of ATGL have yet to be defined. Because perilipin 1 is not known to interact directly with ATGL, we speculate that association of perilipin 1 with the phospholipid monolayer coating LDs may limit the access of ATGL to the underlying TAGs. In this regard, perilipin 2, another perilipin/adipophilin/TIP47 protein containing the conserved N-terminal sequences, was also shown to possess a similar ability to inhibit ATGL by reducing its association with LDs (43).
In the basal state when ATGL largely resides in the cytosol of adipocytes, depletion of CGI-58 drastically diminished the rates of lipolysis mediated by endogenous and overexpressed ATGL. This finding is in accordance with our previous result that CGI-58 is highly effective in stimulating the hydrolytic action of cytosolically localized ATGL toward TAGs stored in the LDs (44). On the other hand, we found that knockdown of perilipin 1 in adipocytes specifically up-regulated the intracellular TAG hydrolase activity and, in parallel, the basal lipolysis derived from ATGL. Given that perilipin 1 attenuated the dose-dependent activation of ATGL by CGI-58, we propose that the molar ratio of CGI-58 to perilipin 1 plays a deciding role in the hydrolysis of LD-contained TAGs by ATGL in the basal state. By showing direct effects of perilipin 1 on ATGL enzyme activity both in the absence and presence of CGI-58, our study provides an extra dimension to a previously suggested model that spatially CGI-58 is sequestered by perilipin 1 at the surface of LDs in the absence of stimulation (23–25).
In response to β-adrenergic stimulation, our data demonstrate that perilipin 1 acts to facilitate the LD redistribution of ATGL thereby enhancing ATGL-mediated lipolysis. First, overexpression of ATGL greatly elevated basal lipolysis but was unable to further enhance stimulated lipolysis in the absence of perilipin 1. Second, perilipin 1 depletion abolished the LD translocation of endogenous ATGL stimulated by isoproterenol. Earlier studies have shown that PKA phosphorylation of perilipin 1 on Ser492 and Ser517 plays critical roles in ATGL activation (22, 25). In this context, it would be interesting to determine whether these phosphorylation events are required for the LD redistribution of ATGL. Because PKA activation does not promote a direct interaction with ATGL (27), phosphorylation of perilipin 1 may be permissive but not sufficient for the LD localization of ATGL. It is possible that the lipolytic complex in adipocytes may include multiple proteins that are assembled in association with perilipin 1 upon β-adrenergic stimulation. The LD translocation of ATGL may thus depend on its binding to another protein component that associates with perilipin 1.
Upon its anchorage to LDs with the TAG core in close proximity, we speculate that the enzyme action of ATGL may be less reliant on the CGI-58 coactivation. This hypothesis is supported by the result that CGI-58 knockdown only modestly decreased the ATGL-mediated lipolysis under the stimulated conditions. To this end, it is also important to note that humans with inactivating mutations in CGI-58 are not obese (45, 46), and knockdown of CGI-58 expression results in smaller but not larger fat pads in mice (47). Consequently, further studies are needed to resolve the discrepancies between the results obtained from our cell-based system and the aforementioned settings. One possibility is that additional mechanisms may exist in human and animals that compensates for the changes associated with CGI-58 in adipose tissue.
FSP27 deficiency in adipocytes leads to the appearance of numerous small LDs along with enhanced lipolysis (31, 33, 36, 37). In contrast to perilipin 1, FSP27 did not directly affect the TAG hydrolase activity of ATGL or the activation of ATGL by CGI-58. The fact that depletion of FSP27 dramatically induced association of ATGL with the multilocular small LDs indicates that FSP27 acts as a lipolytic barrier via limiting the presence of ATGL at the droplet surface. Although FSP27 depletion failed to similarly affect the LD distribution of HSL, it remains to be determined whether the increased LD association of ATGL occurs independently of the expansion of total LD surface area. A previous study has shown that FSP27 alone is not sufficient to protect against ATGL-mediated lipolysis in nonadipocyte cells (48). Thus, we speculate that other factors may cooperate with FSP27 in the unique control of ATGL localization in adipocytes.
Two new studies have shown that FSP27 mediates lipid transfer and fusion between smaller and larger LDs (38, 39). Our data add to these findings by offering decreased LD association of ATGL as an additional mechanism for the promotion of LD growth by FSP27. Considering that further depletion of ATGL increased the size and decreased the number of LDs in cells without FSP27, it would be interesting to determine whether the reduced LD presence of ATGL is prerequisite for the lipid transfer mediated by FSP27 at the interface of LDs. Furthermore, FSP27 is known to exist at low levels in brown adipocytes (36, 37). Work by Wang et al. (27) has demonstrated a localization pattern of ATGL in brown adipocytes similar to what we observed in 3T3-L1 adipocytes depleted of FSP27. In these cells, the LD fraction already contains high content of ATGL in the basal state and as a result, PKA activation does not cause significant recruitment of ATGL to LDs. Interestingly, ablation of ATGL in mice promotes the conversion of brown adipose tissue to a WAT-like tissue where cells contain larger, fewer LDs (49). Therefore, enhanced lipolysis in the absence of FSP27 may be important in maintaining the proper LD morphology and identity of brown adipocytes.
In summary, this study demonstrates that perilipin 1 and FSP27 work coordinately to control the lipolytic capacity in adipocytes. FSP27 functions to constitutively limit the LD presence of ATGL, whereas perilipin 1 plays a pivotal role in the control of the lipolytic response mediated by ATGL to β-adrenergic hormones. In the basal state, perilipin 1 deters the hydrolytic action of cytosolic ATGL toward TAG substrates in the LDs via inhibiting its enzymatic activity as well as sequestering its coactivator CGI-58. Upon stimulation, phosphorylated perilipin 1 augments the ATGL action by promoting the LD translocation of ATGL. In nonadipocyte cells expressing neither perilipin 1 nor FSP27, we predict that ATGL is constitutively associated with the LDs thereby mediating TAG turnover in a hormone-independent manner.
Acknowledgments
The work was supported by the National Institutes of Health Grant DK089178 and a Junior Faculty Award form the American Diabetes Association (J.L.).
Disclosure Summary: The authors have nothing to disclose.
Abbreviations
- Ad-null
Null adenovirus
- ATGL
adipose triglyceride lipase
- CGI-58
comparative gene identification-58
- CIDE
cell death-inducing DNA fragmentation factor 45-like effector
- FA
fatty acid
- FBS
fetal bovine serum
- FFA
free fatty acid
- FSP27
fat-specific protein 27
- HLM
hypotonic lysis medium
- HSL
hormone-sensitive lipase
- LD
lipid droplet
- PKA
protein kinase A
- SDS
sodium dodecyl sulfate
- siRNA
small interfering RNA
- TAG
triacylglycerol
- TIP47
47-kDa tail interacting protein
- WAT
white adipose tissue.
