Oleate activates SREBP-1 signaling activity in SCD1-deficient hepatocytes

Abstract

Stearoyl-CoA desaturase-1 (SCD1) is a key player in lipid metabolism. SCD1 catalyzes the synthesis of monounsaturated fatty acids (MUFA). MUFA are then incorporated into triacylglycerols and phospholipids. Previous studies have shown that Scd1 deficiency in mice induces metabolic changes in the liver characterized by a decrease in de novo lipogenesis and an increase in β-oxidation. Interestingly, Scd1-deficient mice show a decrease in the expression and maturation of the principal lipogenic transcription factor sterol receptor element binding protein-1 (SREBP-1). The mechanisms mediating this effect on de novo lipogenesis and β-oxidation have not been fully elucidated. We evaluated the role of SCD1 on de novo lipogenesis and β-oxidation in HepG2 cells. We also used Scd1-deficient mice and two strains of transgenic mice that produce either oleate (GLS5) or palmitoleate (GLS3) in a liver-specific manner. We demonstrate that the expression of β-oxidation markers increases in SCD1-deficient hepatocytes and suggest that this is due to an increase in cellular polyunsaturated fatty acid content. We also show that the changes in the level of SREBP-1 expression, for both the precursor and the mature forms, are mainly due to the lack of oleate in SCD1-deficient hepatocytes. Indeed, oleate treatment of cultured HepG2 cells or hepatic oleate production in chow-fed GLS5 mice can restore SREBP-1 expression and increase hepatic de novo lipogenesis. Finally, we show that oleate specifically increases SREBP-1 nuclear accumulation, suggesting a central role for oleate in SREBP-1 signaling activity.

de novo lipogenesis (DNL) is the metabolic pathway that synthesizes fatty acids from excess carbohydrates. These fatty acids can then be incorporated into triacylglycerols (TAG) for energy storage. In normal conditions, DNL mainly takes place in liver and adipose tissues (37). Palmitic acid (palmitate, 16:0), the predominant fatty acid generated by fatty acid synthase (FAS), is desaturated by stearoyl-CoA desaturase-1 (SCD1) to produce palmitoleic acid (palmitoleate, 16:1n-7) or it can be elongated to yield stearic acid (stearate, 18:0). SCD1 catalyzes the conversion of stearoyl-CoA to oleoyl-CoA, which is a major substrate for TAG synthesis. Oleic acid (oleate, 18:1n-9) is formed as a result of desaturation of stearic acid, and it is thought to be the end product of de novo fatty acid synthesis (7). DNL is principally regulated by the membrane-bound transcription factor sterol receptor element binding protein-1 (SREBP-1), which transactivates several genes involved in lipid synthesis, such as acetyl-CoA carboxylase (ACC), FAS, and SCD1 (37).

SREBP-1 is initially synthesized as an inactive precursor that forms a complex with SCAP (SREBP cleavage-activating protein) in the endoplasmic reticulum (ER) membrane. Retention of the SCAP-SREBP-1 complex in the ER membrane is mediated by the binding of SCAP to the Insig protein. SCAP binds sterols when their local concentration reaches high levels, inducing conformational changes that allow the formation of a complex between SCAP and the COPII vesicle-formation proteins Sar1, Sec23, and Sec24 in the ER (30, 31). When SCAP binds to the COPII proteins, the SCAP-SREBP-1 complex travels to the Golgi where the SREBP-1 precursor undergoes a sequential two-step proteolytic processing mediated by the site 1 and site 2 proteases (19). This cleavage process leads to the release of the transcriptionally active NH₂-terminal domain of the protein (mature form, ~68 kDa) from the Golgi apparatus. Once cleaved, the mature SREBP-1 is translocated into the nucleus where it can bind to the sterol regulatory element sequence present in promoters/enhancers of lipogenic genes (including itself), activating their transcription (30, 31). Currently, it is still unclear if SCAP or other proteins involved in SREBP-1 processing play a role as cellular sensors for fatty acids in similar manner as with sterols.

SREBP-1 activity alone is not sufficient for the maximal stimulation of lipogenic gene expression. Other transcription factors are also involved in the regulation of DNL such as the liver X receptor, carbohydrate-responsive element binding protein, and the farnesoid X receptor (37). AMP-activated protein kinase (AMPK) is a key player in the regulation of DNL as it directly phosphorylates ACC, inducing a decrease in its activity and a subsequent reduction of DNL. ACC inhibitory phosphorylation also activates β-oxidation as a reduction in the concentration of its enzymatic product malonyl-CoA leads to carnitine palmitoyltransferase-1 (CPT1) activation. AMPK also activates fatty acid oxidation through direct phosphorylation and activation of CPT1 and CPT2 (18).

Scd1 global knockout mice (GKO) are protected from high-fat or high-carbohydrate diet-induced obesity (27). These mice show an increase in the hepatic expression of genes implicated in β-oxidation (Cpt1 and Pparα), as well as a decrease in the hepatic expression of several genes implicated in DNL such as Fas, Acc, and Srebp-1 (25, 27). In addition, an increase in hepatic AMPK activity is observed in GKO mice fed a standard chow diet. This increase in AMPK activity is associated with increased inhibitory phosphorylation of ACC (9).

Recently, transgenic GKO mice were developed with the ability to restore hepatic oleate (GLS5) or palmitoleate (GLS3) monounsaturated fatty acid (MUFA) synthesis by taking advantage of substrate preferences for the human SCD5 (stearic acid substrate) and the mouse Scd3 (palmitic acid substrate) isoforms, respectively (4). GLS5, but not GLS3, mice exhibited increased hepatic lipid accumulation and adiposity compared with GKO mice when fed a lipogenic diet. A decrease in DNL and increase in β-oxidation in white adipose tissue were also specifically observed in the GLS5 mice (4).

As previously mentioned, SCD1 deficiency in hepatocytes decreases DNL, notably through modulation of SREBP-1 expression (both precursor and mature forms). However, the molecular mechanisms inducing these changes are still unknown. To investigate the molecular mechanism mediating changes in SREBP-1 expression in response to SCD1 deficiency, we transfected HepG2 cells with a small interfering (si) RNA targeting SCD1 or used a specific SCD1 inhibitor. We also used GKO, GLS3, and GLS5 mice liver tissue to confirm in vitro results. Our results suggest that oleate is essential for SREBP-1 activity, facilitating its accumulation into the nucleus where it can trigger lipogenic gene expression, including its own expression. This effect is specific to oleate as palmitoleate has no measurable effect. Our results also suggest that an elevated intracellular content of polyunsaturated fatty acid (PUFA) can activate the β-oxidation as observed in SCD1-deficient hepatocytes.

MATERIALS AND METHODS

Mice.

Mice were generated by the laboratory of J. M. Ntambi. All animal procedures were approved by the Institutional Animal Care and Use Committee at the University of Wisconsin-Madison. The SCD5 and Scd3 transgenes (conferring hepatocyte-specific expression) were crossed into the Scd1 GKO C57BL/6 background to produce GKO liver-specific SCD5 (GLS5) and Scd3 (GLS3) transgenic mice, which produce hepatic oleate and palmitoleate, respectively (4, 20). We used five male mice of 11–13 wk of age from each group: wild-type, GKO, GLS5, and GLS3. Mice were fed a chow diet since weaning and fasted for 4 h before tissue collection.

Cell culture and treatments.

HepG2 cells (ATCC) were cultured in Eagle’s minimum essential medium supplemented with 10% fetal bovine serum (Wisent), 100 U/ml penicillin, and 100 μg/ml streptomycin (Life Technologies). At confluence, cells were starved for 24 h in serum-free medium before we proceeded with the various treatments. To inhibit SCD1 expression, cells were transfected for 24 h using DharmaFECT4 reagent (Life Technologies) with Silencer Select Negative Control #1 or SCD-targeting siRNA (s12503; Life Technologies) according to the manufacturer’s instructions. To mimic SCD1 activity, cells were incubated for 24 h with 100 μM oleate (2 mol/mol BSA; Sigma-Aldrich), 100 μM palmitoleate (2 mol/mol BSA; Cayman), or 50 μM fatty acid-free BSA (control condition). To inhibit SCD1 activity, cells were treated for 24 h with 1 μM of SCD1 inhibitor (A939572; Biofine), as previously described (40). Cells were then washed six times with ice cold 1× PBS before subsequent analyses were performed.

Lipid droplet imaging.

After siRNA transfection and 24-h incubation with 100 μM oleate, HepG2 cells were fixed with 4% paraformaldehyde for 30 min. Lipid droplets (LD) were stained using Bodipy 493/502 (1 μg/ml; Life Technologies) and nuclei using 4,6 diamidino-2-phenylindole (DAPI; 1 μg/ml; Life Technologies). Fluorescence was visualized with a Nikon A1 confocal microscope. The number and average size of LD per cell were quantified with ImageJ software using the threshold function.

Quantitative real-time PCR.

Total RNA from cells and mice livers were extracted using the TRIzol reagent (Life Technologies), according to the manufacturer’s instructions. Total RNA (500 ng) was reverse transcribed (SuperScript II reverse transcriptase; Life Technologies) and quantitative real-time PCR was performed using SYBR Green I Master Mix on a LightCycler 480 Real-Time PCR System (Roche Applied Science). Sequences of the primers used are listed in Table 1. Results are represented as arbitrary units indicating relative expressions based on the comparative Ct (ΔΔCt) method. Data were normalized using the housekeeping gene (HPRT1/Hprt) and expressed as fold-changes relative to control samples.

Table 1.

Primer sequences for mouse and human genes used in qPCR

Species/Gene*	Forward Primer (5′–3′)	Reverse Primer (5′–3′)
Mouse
Srebp-1 (Srebf1)	CCTGCCTTCAGGCTTCTCAGG	GAGGCCAAGCTTTGGACCTGG
Fas (Fasn)	ATTGCATCAAGCAAGTGCAG	GAGCCGTCAAACAGGAAGAG
Acc (Acaca)	TGAAGGGCTACCTCTAATG	TCACAACCCAAGAACCAC
Hprt	TCAGTCAACGGGGGACATAAA	GGGGCTGTACTGCTTAACCAG
Human
SREBP-1 (SREBF1)	ACAGTGACTTCCCTGGCCTAT	GCATGGACGGGTACATCTTCAA
HPRT1	CCTGGCGTCGTGATTAGTGAT	AGACGTTCAGTCCTGTCCATAA

qPCR, quantitative reverse-transcription-polymerase chain reaction; SREBP-1, sterol receptor element binding protein-1; FAS, fatty acid synthase; ACC, acetyl-CoA carboxylase; HPRT1, hypoxanthine phosphoribosyltransferase 1.

^*Official gene names are provided in parentheses when different from the common name used in this study.

Immunoblot analysis.

HepG2 cells and mice liver cells were lysed in RIPA buffer containing 0.1 mM PMSF and 1% of cOmplete protease inhibitor (Sigma-Aldrich). Proteins were denatured at 95°C for 5 min, 30 μg protein/lane were loaded onto SDS-PAGE, and immunoblot analyses were carried out using primary antibodies diluted at 1:1,000 in PBS-5% milk. Antibodies used were targeted against SREBP-1 (H-160; sc-8984; Santa Cruz Biotechnology); FAS (sc-55580; Santa Cruz Biotechnology); ACC, phospho-ACC, AMPKα, and phospho-AMPKα (9957S; Cell Signaling); CD36 (ab78054; Abcam); peroxisome proliferator-activated receptor-γ (PPARγ; 2435; Cell Signaling); PPARα (sc-9000; Santa Cruz Biotechnology); CPT1 (ab128568; Abcam); and GAPDH (5174), α-tubulin (2144), and HDAC1 (5356; Cell Signaling). After incubation, the membranes were washed in 1× PBS + 0.1% Tween and incubated with anti-mouse or anti-rabbit horseradish peroxidase-conjugated secondary antibodies (1:5,000; 7076 and 7074; Cell Signaling). The immunoreactive bands were revealed by chemiluminescence (Millipore) using a chemi-luminometer (FusionFX). ImageJ software was used to quantify band intensity.

Immunofluorescence.

HepG2 cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. Cells were washed with PBS, blocked with 3% BSA in Tris-buffered saline containing 1% Triton X-100, and then incubated with anti-SCD1 (ab19862; Abcam; mouse primary 1:200) and anti-SREBP-1 (sc-8984; Santa Cruz Biotechnology; rabbit primary 1:50). Cells were washed with 1× PBS and incubated with anti-rabbit Alexa Fluor 488 and anti-mouse Alexa Fluor 647 (Abcam; goat secondary 1:1,000). The cells were washed with 1× PBS, and images were acquired using a Nikon A1 confocal microscope fitted with a ×100 oil-immersion lens. The percentage of nuclear SREBP-1 expression was then analyzed with the Intensity Ratio Nuclei Cytoplasm macro from ImageJ software, using DAPI staining to define the nucleus area.

Lipid extraction and quantification using gas chromatography.

Cells transfected with siRNA (as well as corresponding controls) were trypsinized for 5 min, collected, and spun at 1,200 rpm for 5 min to pellet hepatocytes. HepG2 cell pellets were washed in 1× PBS. Lipid extractions were conducted using previously described protocols (2, 29). Samples were analyzed using a 7890B gas chromatography system (Agilent Technologies) with a flame ionization detector, and samples were separated on a J&W DBFFAP fused-silica capillary column (15 m, 0.1-μm film thickness, and 0.1-mm inner diameter; Agilent Technologies). Fatty acid peaks were identified by comparison with retention times of fatty acid methyl ester standards. To estimate SCD1 activity, we calculated the product-to-precursor fatty acid ratio (i.e., 18:1n-9/18:0 and 16:1n-7/16:0), as previously reported (1, 11). Fatty acid data were normalized to protein concentrations for each treatment condition and reported as micrograms of fatty acid per microgram per microliter of protein.

De novo lipogenesis.

DNL was evaluated by measuring the incorporation of [1,2-¹⁴C]acetic acid (acetate) into lipids, as described previously (2, 3, 16). Briefly, posttransfection cells were incubated with 18.4 nmol (1 μCi) of [¹⁴C]acetate for 4 h. Cells were then washed six times with ice-cold 1× PBS and suspended in 1× PBS, and total cellular lipids were extracted in chloroform/methanol (2:1, vol/vol). The lipid extract was dried under nitrogen and reconstituted in 100 μl hexane. Radiolabeled lipids were separated by thin-layer chromatography on silica-coated plates using a hexane/diethyl-ether/acetic acid solution (80:20:1, vol/vol) as a developing solvent (2, 3, 16). Lipids were visualized by exposure to iodine vapors, and the bands corresponding to lipid standards [triacylglycerols (TAGs), diacylglycerols (DAGs), free fatty acids, cholesterol esters, and phospholipids] were scraped into separate vials. Radioactivity was measured, and data are presented as number of counts per minute of [¹⁴C]lipids per microgram of total protein.

Protein fatty acid modification assay.

HepG2 cells were inoculated into six-well plates and incubated overnight at 37°C. These cells were then incubated for 24 h in serum-free medium containing 18.5 kBq/ml (8.3 nM) [9,10-³H]oleate, [9,10-³H]palmitoleate, or [³H]H₂O, or incubated in serum-free medium without radioactive compound. Cells were washed six times with ice cold 1× PBS to prevent background noise. Cells were then lysed in RIPA buffer containing 0.1 mM PMSF and 1% cOmplete protease inhibitor (Sigma-Aldrich). The lysate was immunoprecipitated with anti-SREBP-1 (H-160) antibody (sc-8984; Santa Cruz Biotechnology) and protein A-Sepharose 4B beads (Invitrogen). Immunoprecipitates were separated by SDS-PAGE. Bands corresponding to the precursor and mature SREBP-1 protein forms were cut out, and radioactivity was measured in scintillation liquid in a TRI Carb 2800TR liquid scintillation counter (Perkin Elmer).

Statistical analysis.

Student’s t-test was used to compare two groups (cell culture data analysis), and a two-way ANOVA was used when more than one factor was evaluated (transgenic mice data analysis). P < 0.05 was considered statistically significant.

RESULTS

SCD1 deficiency decreases fatty acid synthesis and storage in hepatocytes.

We used a validated siRNA to induce SCD1 deficiency in HepG2 hepatocarcinoma cells, resulting in an 82% knockdown of SCD1 protein expression (Fig. 1A). To confirm that SCD1 deficiency has an impact on lipid metabolism in HepG2 cells, we evaluated fatty acid synthesis by incorporation of radiolabeled acetate, a lipid precursor, followed by quantification of acetate incorporation in various lipids classes separated using thin-layer chromatography. Inhibition of SCD1 expression decreased total lipid synthesis by 20% (Fig. 1B, left), mainly due to decreased TAG and DAG content (35 and 60% decrease, respectively) (Fig. 1B, middle and right). Other lipid classes (free fatty acids, cholesterol esters, and phospholipids) were not significantly affected by SCD1 knockdown (data not shown). Concomitantly, the levels of key hepatic DNL proteins were lowered (Fig. 1C). SREBP-1 levels decreased by 20% for the precursor form and 35% for the mature form, whereas FAS and ACC levels decreased by 12 and 16%, respectively.

Fig. 1.

Stearoyl-CoA desaturase-1 (SCD1) knockdown decreases de novo lipogenesis in HepG2 cells. A: SCD1 protein expression was evaluated by Western blotting in HepG2 cells transfected with negative control (Ctrl) or SCD1-targeting small interfering (si) RNA (siSCD1). Protein levels measured by densitometry were normalized against GAPDH and presented relative to negative control. B: HepG2 cells transfected with negative control (Ctrl) or SCD1-targeting siRNA (siSCD1) were incubated with [¹⁴C]acetate. Incorporated radioactivity was measured by thin layer chromatography in total lipids as well as in lipid fractions, including triacylglycerols and diacylglycerols, and presented as counts per minute (CPM) of [¹⁴C]lipids per microgram of total protein. C: the expression of proteins involved in lipid metabolism [precursor and mature sterol receptor element binding protein-1 (SREBP-1), fatty acid synthase (FAS), and acetyl-CoA carboxylase (ACC)] was evaluated by Western blotting in SCD1-deficient HepG2 cells. Protein levels measured by densitometry were normalized against GAPDH and presented relative to negative control. Each histogram presents the means ± SD of 5 independent experiments. *P < 0.05, **P < 0.002, and ***P < 0.001.

To determine whether these effects are reflected in cellular lipid storage, we visualized LDs within SCD1-deficient HepG2 cells (Fig. 2A). Cells were also treated with oleate as it increases LD formation (32). Inhibition of SCD1 expression decreased LD number, in both the absence and presence of oleate (40 and 37% decrease, respectively; Fig. 2B, top). In fact, oleate treatment normalized LD number in SCD1-deficient cells. Inhibition of SCD1 expression had no effect on LD size in the absence of oleate yet partially blocked LD expansion (35% less) when cells were stimulated with oleate (Fig. 2B, bottom).

Fig. 2.

SCD1 knockdown modulates the number and size of lipid droplets (LDs) in HepG2 cells. A: HepG2 cells transfected with negative control (Ctrl) or SCD1-targeting siRNA (siSCD1) and treated (or not) with 100 μM oleate were stained with Bodipy 493/503 (green) and 4,6 diamidino-2-phenylindole (DAPI; blue) to visualize LDs and nuclei, respectively. Bars = 20 μm. B: number and size (diameter) of LD observed in transfected/treated HepG2 cells. Each histogram presents the means ± SD of 5 independent experiments. *P < 0.05 and **P < 0.002.

Fatty acid β-oxidation protein marker levels and PUFA content increase in SCD1-deficient hepatocytes.

We next measured β-oxidation protein markers in SCD1-deficient HepG2 cells. SCD1 knockdown induced the expression of PPARα (29% increase), a transcription factor regulating several genes involved in β-oxidation including CPT1 (28% increase), an essential transporter for mitochondrial degradation of long-chain fatty acids (Fig. 3A). Since PPARα activity can be regulated by PUFA (8, 17), we measured the cellular concentration of several key PUFAs known to directly activate this transcription factor. SCD1 inhibition in HepG2 cells increased both total omega-3 (Fig. 3B, left) and omega-6 fatty acid (Fig. 3C, left) content. In particular, we observed a significant increase in eicosapentanoic acid (EPA; 20:5n-3) and docosahexanoic acid (DHA; 22:6n-3; Fig. 3B, right), as well as in linoleic acid (18:2n-6), eicosadienoic acid (20:2n-6), dihomo-γ-linoleic acid (20:3n-6), and arachidonic acid (AA; 20:4n-6; Fig. 3C, right).

Fig. 3.

Inhibition of SCD1 expression increases β-oxidation marker levels and increase polyunsaturated fatty acid (PUFA) content in HepG2 cells. A: the expression of proteins involved in β-oxidation [peroxisome proliferator-activated receptor-α (PPARα) and carnitine palmitoyltransferase-1 (CPT1)] was evaluated by Western blot in HepG2 cells transfected with negative control (Ctrl) or SCD1-targeting siRNA (siSCD1). Protein levels measured by densitometry were normalized against GAPDH and presented relative to negative control. Each histogram presents the means ± SD of 5 independent experiments. The cellular PUFA content of siRNA-transfected HepG2 cells was determined by gas chromatography-mass spectrometry (GC-MS). Omega-3 (B) and omega-6 (C) total PUFA content, as well as PUFA fractions, were analyzed. The results of 4 experiments are presented as means ± SE. *P < 0.05, **P < 0.002, and ***P < 0.001.

Oleate treatment rescues fatty acid content and SREBP-1 expression in SCD1-deficient hepatocytes.

To better understand how SCD1 activity modulates SREBP-1 expression, we incubated SCD1-deficient HepG2 cells with oleate (18:1n-9), one of the SCD1 enzymatic products. We used gas chromatography-mass spectrometry to verify the impact of our treatments on cellular lipids. As expected, SCD1 knockdown led to a decrease in 16:1n-7 and 18:1n-9, as well as an accumulation of saturated 16:0 and 18:0 fatty acids. Incubation of cells with oleate specifically increased the 18:1n-9 level independently of SCD1 expression (Fig. 4A, left and middle). We also observed a decrease in total fatty acid content in untreated siSCD1-transfected cells (Fig. 4A, right), echoing the decrease in fatty acid synthesis observed earlier (Fig. 1B). Oleate treatment rescued total lipid content in SCD1-deficient cells (Fig. 4A, right), suggesting an effect of this SCD1 product on DNL.

Fig. 4.

Oleate treatment reverses SREBP-1 expression and maturation defects in SCD1-deficient HepG2 cells. HepG2 cells were transfected with negative control (Ctrl) or SCD1-targeting siRNA (siSCD1) and treated (or not) with 100 μM oleate. A: fatty acids were extracted and quantified using GC/MS, including the monounsaturated fatty acids (MUFA) palmitoleate and oleate and saturated fatty acids (SFA) palmitate and stearate. The results of 5 experiments are presented as means ± SE. B: SREBP-1 mRNA expression was analyzed by quantitative reverse-transcription-polymerase chain reaction (qPCR) relative to housekeeping gene HPRT1. C: SREBP-1 protein expression, for both precursor (pSREBP-1) and mature (mSREBP-1) forms, was analyzed by Western blotting. D: protein levels evaluated by densitometry were normalized against GAPDH, used to calculate the mature/total ratio, and presented relative to untreated negative control. Each histogram presents the means ± SE of 5 independent experiments. *P < 0.05, **P < 0.002, and ***P < 0.001.

We focused next on the expression of SREBP-1, an essential transcription factor in DNL activation. The level of SREBP-1 mRNA was decreased in SCD1-deficient cells (Fig. 4B). Interestingly, addition of oleate increased SREBP-1 mRNA expression in both control (64% increase) and SCD1-deficient cells (69% increase) (Fig. 4B). Western immunoblots also showed an important increase in both the precursor and mature forms of the SREBP-1 protein upon oleate treatment of SCD1-deficient cells (42 and 75% increase, respectively; Fig. 4, C and D). In addition, the mature-to-total SREBP-1 ratio was decreased in SCD1-deficient cells and was rescued by oleate treatment (Fig. 4D, right), suggesting an impact of oleate on SREBP-1 maturation.

Oleate specifically rescues SREBP-1 nuclear accumulation in SCD1-deficient hepatocytes.

We next explored the effect of decreasing SCD1 expression (by siRNA transfection) and activity (by treatment with inhibitor A939572) on SREBP-1 cellular localization in HepG2 cells, as observed by immunofluorescence (Fig. 5A). Oleate and palmitoleate treatments were used to mimic SCD1 activity. Consistent with what we observed previously (Fig. 4, C and D), oleate supplementation increased the expression of SREBP-1 in all the conditions tested, including in cells treated with the SCD1 inhibitor. Inhibition of SCD1 expression or activity significantly decreased cellular SREBP-1 levels, while oleate treatment restored it to normal levels (Fig. 5B). We then determined the percentage of SREBP-1 within the DAPI-stained cell nucleus as an estimation of the proportion of transcriptionally active protein. SREBP-1 nuclear localization was significantly diminished in both SCD1-deficient cells and in SCD1 inhibitor-treated cells and was rescued by incubation with oleate (Fig. 5C). On the other hand, incubating cells with palmitoleate did not reverse the effect of SCD1 deficiency or inhibition on either SREBP-1 expression (Fig. 5D) or nuclear localization (Fig. 5E). This observation emphasizes a specific role for oleate on SREBP-1 maturation and nuclear translocation.

Fig. 5.

Oleate specifically increases SREBP-1 expression and nuclear localization. A: SCD1 (red) and SREBP-1 (green) expression was evaluated by immunofluorescence on HepG2 cells transfected with negative control (Ctrl) or SCD1-targeting siRNA (siSCD1 or siR), or incubated with 1 μM SCD1 inhibitor A939572 (inh.). Cells were treated with 100 μM oleate or palmitoleate (Palm.). DAPI (blue) was used to visualize nuclei. Bars = 20 μm. The SREBP-1 signal per cell (B and D), as well as the SREBP-1 nuclear proportion (C and E), was measured in all experimental conditions. Each histogram presents the means ± SD of 3 independent experiments. *P < 0.05 and **P < 0.002.

Oleate specifically modifies the SREBP-1 protein.

To investigate a possible mechanism through which oleate triggers the nuclear accumulation of SREBP-1, we sought to determine whether oleate could modify the SREBP-1 NH₂-terminal domain. We extracted SREBP-1 from HepG2 cells incubated with radiolabeled oleate or palmitoleate. We could detect radioactivity in the mature form of SREBP-1 when cells were treated with oleate but not palmitoleate (Fig. 6). This radioactivity was detected in immunoprecipitates and following gel migration in denaturing conditions, suggesting a covalent modification of the SREBP-1 protein.

Fig. 6.

The SREBP-1 protein is specifically modified by oleate. SREBP-1 was immunoprecipitated from protein extracts of HepG2 cells previously incubated 24 h with [³H]H₂O, [³H]oleate, or [³H]palmitoleate (palm.). Immunoprecipitates (IP) were submitted to denaturing gel electrophoresis to separate precursor and mature SREBP-1 (pSREBP-1 and mSREBP-1). Incorporated radioactivity was measured in isolated mSREBP-1 protein bands and presented relative to nonradioactive control (Ctrl). WB, Western blot. The histogram presents the means ± SD of 3 independent experiments. ***P < 0.001.

Oleate specifically increases SREBP-1 levels and de novo lipogenesis in mice livers.

To test the effect of oleate on SREBP-1 expression and activity in vivo, we used mouse transgenic lines in a Scd1 GKO background that express specific desaturases in their liver, leading to the hepatic production of either oleate (GLS5 line) or palmitoleate (GLS3 line) (4, 20). Expression of both precursor and mature (presumably nuclear) forms of SREBP-1 was decreased in GKO mice livers compared with controls. In line with our in vitro results, SREBP-1 expression was partially rescued in livers of the GLS5 but not the GLS3 group (Fig. 7A). We obtained the same result using hepatic nuclear extract (data not shown). Expression of lipogenic genes (Srebp-1, Acc, and Fas) was also partially rescued in mouse livers producing oleate but not palmitoleate (Fig. 7B), highlighting a central role of oleate in SREBP-1 signaling activity. As previously described (9), AMPK and ACC phosphorylation was increased in GKO livers compared with controls (Fig. 7C). However, restoring either oleate (GLS5) or palmitoleate (GLS3) hepatic synthesis did not modify the phosphorylation level of either protein, suggesting in this case a SREBP-1-independent mechanism.

Fig. 7.

SREBP-1 expression and de novo lipogenesis are partially restored in the liver of Scd1 knockout mice specifically producing oleate. Livers of control wild-type mice (Ctrl), Scd1 knockout mice (GKO), and mice specifically producing oleate or palmitoleate (GLS5 and GLS3, respectively) were used to evaluate the importance of SCD1 MUFA products in SREBP-1 expression and de novo lipogenesis. A: SREBP-1 expression, for both precursor (pSREBP-1) and mature (mSREBP-1) forms, in total protein extracts was analyzed by Western blotting, in triplicate. Protein levels evaluated by densitometry were normalized against α-tubulin and presented relative to wild-type control. B: expression of key genes implicated in DNL (Srebp-1, Acc, and Fas) were measured by qPCR, normalized against Hprt and presented relative to wild-type control. C: protein expression of pAMPK, AMPK, pACC, and ACC was measured by Western blotting, in triplicate. Protein levels measured by densitometry were normalized against α-tubulin, used to calculate the phosphorylated/total ratio, and presented relative to wild-type control. Each histogram presents the means ± SD of 5 tissue measurements. *P < 0.05, **P < 0.002, and ***P < 0.001.

DISCUSSION

Scd1 (GKO) mice are resistant to obesity and hepatic steatosis in response to a high-fat diet (27). Compared with wild-type mice, GKO mice also present a lower activation of hepatic DNL and an induction of hepatic β-oxidation (27). However, the molecular mechanisms underlying these effects are still unknown. To investigate these mechanisms, we used HepG2 cells where expression or activity of SCD1 has been inhibited. Our results were also validated in GKO mice and in transgenic GKO mice expressing human SCD5 (GLS5) or mouse Scd3 (GLS3) in their liver. These mice specifically synthesize either hepatic oleate (GLS5) or palmitoleate (GLS3) (4, 20). We confirm here that SCD1 deficiency decreases DNL and lipid storage (in the form of LD) while markers of β-oxidation are increased. We show that SCD1 deficiency is associated with a higher concentration of cellular PUFA, specifically DHA and EPA, which probably trigger β-oxidation via the activation of PPARα. Our study also suggests that oleate covalently modifies the SREBP-1 protein and increases its accumulation within the nucleus where it can activate transcription of lipogenic genes.

SCD1 deficiency is associated with a decrease in the number of LD (Fig. 2). This result suggests that SCD1 may play an important role in the biogenesis of these lipid-storing organelles. This hypothesis is consistent with previous studies showing that SCD1 activity is implicated in LD formation (22, 35). Interestingly, SCD1 deficiency also diminishes HepG2 LD size, but only the presence of oleate, when SCD1 activity is presumably bypassed and LD are particularly large. This latter result argues for an enzymatic activity-independent role for the SCD1 protein in the LD expansion process. The exact molecular role of SCD1 in LD formation remains to be clarified. Production of MUFA by SCD1 in the ER could influence the activity of enzymes such as diacylglycerol acyltransferase (DGAT), acyl-CoA cholesterol acyltransferase, and microsomal glycerol phosphate acyltransferase involved in TAG and cholesterol ester formation, the principal lipids incorporated in the LD core (13). This hypothesis is supported by a study showing that SCD1 colocalizes with DGAT2 in the ER (23). These enzymes, including SCD1, may exist as an ER membrane-resident complex promoting substrate channeling and fuel partitioning into various metabolic pathways.

As observed in SCD1-deficient mouse livers (27), inhibition of SCD1 expression in cultured hepatocytes modifies lipid metabolism. Upon SCD1 knockdown in HepG2 cells, we observed a decrease in DNL markers as well as an increase in β-oxidation markers (Figs. 1C and 3A). Not surprisingly, we showed that SCD1 deficiency is associated with an intracellular accumulation of the saturated fatty acid (SFA) palmitic acid (16:0) and stearic acid (18:0) (Fig. 4A). Interestingly, we also observed an increase in omega-3 and omega-6 PUFA species (Fig. 3, B and C). Elevated SFA and PUFA concentrations could underlie the β-oxidation increase observed upon SCD1 inhibition. SFA are activating ligands of PPARα, the main transcriptional activator of genes (such as CPT1) implicated in β-oxidation (5, 6). Elevated cellular PUFA content (especially EPA, DHA, and AA) has also been associated with activation of β-oxidation through an increase in the expression and activity of PPARα (10, 21, 39). Mice on a PUFA-rich diet (containing mostly EPA and DHA) show a decrease in the expression of genes implicated in lipogenesis and an increase in the expression of genes implicated in β-oxidation, notably PPARα and CPT1 (10, 39). PUFAs are able to regulate metabolic gene expression through direct binding and activation of PPARs (8, 12, 17). Recently, a study performed with mice fed a MUFA-enriched diet showed that desaturase-5 and desaturase-6 expression is decreased and that cellular EPA and DHA content is lower relative to mice fed with SFA (28). These results suggest that PUFA synthesis is regulated by MUFA, although the mechanism through which this happens is still unknown.

Our study also revealed that oleate treatment of HepG2 cells increases SREBP-1 gene and SREBP-1 protein expression as well as its nuclear accumulation, although the latter was specific to cells in which SCD1 activity or expression was diminished (Figs. 4, B–D, and 5). These observations suggest that SCD1 could play an important role in activating DNL via an oleate-dependent activation of SREBP-1. Previous in vitro studies in HEK-293 cells have shown that oleate and palmitoleate suppress the proteolytic processing of the SREBP-1 protein and reduce expression of the SREBP-1 gene (14). This discrepancy with our result could be explained by differences in SREBP-1 isoform ratios in various cell types. SREBP-1c is the predominant isoform in the liver but in other tissues the SREBP-1a isoform is the predominant one (36). Perhaps SREBP-1a is more prevalent than SREBP-1c in HEK-293 cells. Nevertheless, our result is consistent with a previous study showing that SREBP-1 is overexpressed in the livers of SCD1-deficient mice raised on a fructose- and oleate-enriched diet (25). While long-term oleate feeding restored SREBP-1 expression in these mice, lipogenic gene expression was only partially rescued. The authors suggested that endogenously synthetized oleate is a more readily accessible regulator of lipogenic gene expression than dietary oleate and hypothesized the existence of a SREBP-1-independent pathway regulating DNL (25). We also showed that restoring hepatic oleate synthesis in GKO animals partially restores SREBP-1 expression, which could explain the lipogenic gene expression increase specifically observed in GLS5 mice (Fig. 7, A and B). SCD1 inhibition increases SFA cellular level. Elevated intracellular SFA has been shown to promote ACC polymerization and inactivation, and the subsequent inhibition of DNL (15). In addition, an increase in intracellular SFA levels has been associated with elevated AMPK activity (34). Subsequently, AMPK may phosphorylate ACC, decreasing its activity and resulting in the inhibition of DNL (18). In agreement with this, we observed an increase in the phosphorylation of AMPK and ACC upon SCD1 inhibition, which is not abrogated by either restoration of hepatic oleate (GLS5) or palmitoleate (GLS3) (Fig. 7C). This confirms the presence of a SREBP-1-independent regulation mechanism. Taken together, this evidence may explain why only a partial restoration of lipogenesis was observed, as oleate supplementation (in both HepG2 cells and GLS5 mice) increases MUFA levels while SFA levels remain elevated.

To investigate the mechanism by which oleate increases SREBP-1 nuclear accumulation and expression, we tested a potential fatty acid modification of the SREBP-1 protein. Radiolabeled oleate is closely associated to SREBP-1 (Fig. 6). The presence of a radiolabel in SREBP-1 immunoprecipitates separated in denaturing gel conditions suggests the implication of a covalent bond. Although we cannot exclude the possibility that oleate metabolites are somehow specifically involved in this protein modification, our results suggest the intriguing possibility that SREBP-1 is acylated by an oleate moiety. Mass spectrometry analyses have yet to support such a possibility, perhaps because an oleate moiety would render peptide fragments particularly hydrophobic, which could hinder analyses. More work is needed to determine whether SREBP-1 can indeed be acylated. Nevertheless, we believe that oleate association might facilitate SREBP-1 precursor cleavage and/or translocation of the mature form to the nucleus. In agreement with our observations, many studies have shown that protein acylation facilitates the trafficking of proteins through cellular membranes, such as in the case of Ghrelin and Wnt modifications (26, 33). Wnt3a is acylated by palmitoleate (16:1n-7) on the Ser209 residue through the direct action of the Porcupine enzyme (38). Porcupine is an ER-resident protein that belongs to the membrane-bound O-acyltransferase protein family. This family contains 11 distinct members, each with a distinct substrate preference (24). A specific Porcupine inhibitor did not affect the oleate modification of SREBP-1 (data not shown). This result is in line with the fact that oleate, and not palmitoleate, is involved in SREBP-1 modification and nuclear accumulation.

Taken together, our data demonstrate that inhibition of SCD1 expression is associated with a decrease in DNL, associated with lower SREBP-1 expression and nuclear translocation. This effect is specifically rescued by oleate treatment. Our data also suggest that oleate can modify the SREBP-1 protein, increasing its nuclear translocation and leading to increased expression of lipogenic genes.

GRANTS

The Discovery Grants Program of the National Science and Engineering Research Council of Canada (NSERC) funded this research. M. A. Lounis was supported by a Fond de Recherche du Québec-Nature et Technologie (FRQNT) fellowship. J. M. Ntambi was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant R01-DK-062388, American Diabetes Association Grant 7-13-BS-118, and US Department of Agriculture Grant Hatch W2005. M. S. Burhans was supported by NIDDK Grant T32-DK-007665.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.A.L. and C.M. conceived and designed research; M.A.L. and M.S.B. performed experiments; M.A.L., K.-F.B., and C.M. analyzed data; M.A.L., K.-F.B., and C.M. interpreted results of experiments; M.A.L. and K.-F.B. prepared figures; M.A.L. and K.-F.B. drafted manuscript; M.A.L., K.-F.B., J.M.N., and C.M. edited and revised manuscript; M.A.L., K.-F.B., M.S.B., J.M.N., and C.M. approved final version of manuscript.