The authors have declared that no competing interests exist.
Conceived and designed the experiments: TZ MP JF. Performed the experiments: TZ MP. Analyzed the data: TZ MP RE JF. Contributed reagents/materials/analysis tools: MP JF. Wrote the paper: TZ MP JF.
Long chain acyl-CoA synthetases are essential enzymes of lipid metabolism, and have also been implicated in the cellular uptake of fatty acids. It is controversial if some or all of these enzymes have an additional function as fatty acid transporters at the plasma membrane. The most abundant acyl-CoA synthetases in adipocytes are FATP1, ACSVL4/FATP4 and ACSL1. Previous studies have suggested that they increase fatty acid uptake by direct transport across the plasma membrane. Here, we used a gain-of-function approach and established FATP1, ACSVL4/FATP4 and ACSL1 stably expressing 3T3-L1 adipocytes by retroviral transduction. All overexpressing cell lines showed increased acyl-CoA synthetase activity and fatty acid uptake. FATP1 and ACSVL4/FATP4 localized to the endoplasmic reticulum by confocal microscopy and subcellular fractionation whereas ACSL1 was found on mitochondria. Insulin increased fatty acid uptake but without changing the localization of FATP1 or ACSVL4/FATP4. We conclude that overexpressed acyl-CoA synthetases are able to facilitate fatty acid uptake in 3T3-L1 adipocytes. The intracellular localization of FATP1, ACSVL4/FATP4 and ACSL1 indicates that this is an indirect effect. We suggest that metabolic trapping is the mechanism behind the influence of acyl-CoA synthetases on cellular fatty acid uptake.
The lipid metabolism of adipose tissue plays an important role in health and is involved in the pathogenesis of several diseases
One family of proteins involved in fatty acid uptake is the acyl-CoA synthetase family. They are highly conserved enzymes which catalyze the ATP-dependent esterification of long chain fatty acids (LCFAs) with coenzyme A, transforming them into activated intermediates for either beta-oxidation or the biosynthesis of lipids
Insulin increases the fatty acid uptake of adipocytes, and FATP1 was proposed to mediate this effect
ACSVL4/FATP4 was initially reported to be the major intestinal fatty acid transporter
ACSL1 is quantitatively the most abundant ASCL in adipocytes and its expression is highly increased during differentiation
In this study, we investigated the subcellular localization of FATP1, ACSVL4/FATP4 and ACSL1 in 3T3-L1 adipocytes by stable overexpression. We found that FATP1 and ACSVL4/FATP4 share a distinct intracellular localization which corresponds to the ER, while ACSL1 was localized primarily on mitochondria. The intracellular localization of all three proteins was sufficient to enhance fatty acid uptake. Insulin increased the uptake of fluorescent fatty acids in FATP1 and ACSVL4/FATP4 overexpressing adipocytes without changing the intracellular localization of both proteins. Thus, we could demonstrate that acyl-CoA-synthetases located intracellularly are sufficient to drive basal and insulin-stimulated fatty acid in 3T3-L1 adipocytes.
Antibodies used were obtained from the following sources: rabbit anti-FATP4 was generated as described earlier
3T3-L1 fibroblasts from ATCC (CL-173) were kindly provided by Susanne Mandrup (University of Southern Denmark, Denmark) and Christoph Thiele (University of Bonn, Germany) and cultured in Dulbecco’s modified Eagle’s medium with 4.5 g/L glucose (DMEM; Invitrogen, Karlsruhe, Germany), 10% fetal calf serum (Biochrom, Germany), 8 mg/L pantothenic acid, 8 mg/L D-biotin, 100 U/ml penicillin/streptomycin and 1% GlutaMax (Invitrogen, Karlsruhe, Germany). Phoenix-gp cells were grown in Dulbecco’s modified Eagle’s medium containing 4.5 g/l glucose, 10% fetal calf serum, 1% GlutaMax and 100 U/ml penicillin/streptomycin under standard tissue culture conditions.
3T3-L1 fibroblasts were differentiated as described previously
The retroviral vector pRJ is derived from the moloney murine leukemia retrovirus based plasmid pQCXIP (BD Biosciences, Heidelberg, Germany) and contains a modified multiple cloning site
The generation of infectious pseudotyped retroviral particles was essentially done as described
Total RNA was extracted with High Pure RNA Isolation Kit (Roche, Mannheim, Germany) and reverse transcription was performed with Transcriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany) using hexanucleotides for random priming.
The mRNA levels of FATP1 (NM_011989.4) and ACSVL4/FATP4 (NM_011977.2) were determined by efficiency corrected relative quantification on an Applied Biosystems 7500 Fast Real-Time PCR System (Foster City, CA), using SYBR Green (Power SYBR Green Master Mix; Roche, Mannheim, Germany) for detection. The quantity of each mRNA was obtained by using a calibration curve derived from five subsequent dilutions of the corresponding plasmids. The quantities were normalized to the quantities of general transcription factor 2b (Gtf2b) (NM_145546.1), a housekeeping gene in adipocytes
Primers (5′ to 3′):
ACSVL4/FATP4
FATP1
Gtf2b
3T3-L1 adipocytes were starved in serum-free medium containing 1% bovine serum albumin (BSA). After 3 h, the medium was removed and the cells were washed with Krebs Ringer HEPES (KRH) buffer (120 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl2, 10 mM HEPES, 1.2 mM KH2PO4, 1.2 mM MgSO4, pH 7.4). Then, serum-free medium containing defined concentrations of [3H]-oleic acid (ART-198; Biotrend Chemikalien GmbH, Cologne, Germany) bound to fatty acid free BSA in a specific ratio was added (specific activity 0.5–1 Ci/mol). After 5 min, 60 min or 3 h, the labeling mix was removed and the uptake was stopped by washing the cells 2x with ice-cold PBS containing 0.5% BSA and then 2x with ice-cold PBS. The cells were then lysed with 1M NaOH and aliquots of each lysate were used for scintillation counting in a Beta-Counter LS 6500 (Beckman-Coulter, CA). Protein concentration was measured by Bradford assay.
Fluorescent Bodipy fatty acids are very long chain fatty acid analogues and have been frequently used to measure cellular fatty acid uptake
3T3-L1 adipocytes were starved for 3 h in serum-free medium containing 1% bovine serum albumin (BSA). The cells were then washed and, if indicated, stimulated with 1 µg/ml insulin in KRH buffer for 20 min at 37°C. Glucose transport was measured by adding [3H]-2-deoxy-D-glucose (NET-549, PerkinElmer, Waltham, MA) (final concentration: 0.1 mM 2-deoxy-D-glucose [DOG], 1 µCi/ml) for 10 min. Uptake was stopped by immediately removing the labeling mix and washing 4x with ice-cold PBS. Radioactivity and protein measurement were done as described for oleate uptake.
3T3-L1 adipocytes were transiently transfected with the following plasmids: OCT-GFP.pcDNA3 (N-terminus of ornithine carbamyl transferase
In brief, adipocytes on day 6–8 of differentiation were detached with trypsin and suspended in standard medium containing 4% glycerol. The cells were centrifuged for 8 min at 8,000 g and room temperature, and then resuspended in medium containing 4% glycerol. An aliquot of this cell suspension containing 2×106 cells was taken, centrifuged and the remaining medium was completely removed. This pellet was resuspended in 100 µl Nucleofector solution L (Amaxa, Cologne, Germany) containing 2–3 µg DNA. Transfection was performed in a Nucleofector (Amaxa, Cologne, Germany) with the default program A-033. After transfection, 500 µl of fresh medium was immediately added. The cell suspension was then transferred to a gelatin/fibronectin coated well containing 2 ml fresh medium. Adipocytes were used 2 d post transfection for experiments. Transfection efficiency ranged from 20–60% depending on the plasmid used.
3T3-L1 adipocytes grown on coverslips were fixed with 4% PFA at room temperature and permeabilized with PBS containing 0,01% saponin, 0,2% gelatin and 0,02% sodium azide. After blocking with PBS containing 1% BSA and 0.2% gelatin, cells were stained with indicated antibodies overnight at 4°C. The coverslips were mounted using Mowiol (Calbiochem, Germany). Images were acquired on a Leica TCS SP2 confocal microscope (Leica Microsystems, Wetzlar, Germany) and arranged with Adobe Photoshop (Adobe Systems, Mountain View, CA).
Fractionation of 3T3-L1 adipocytes was done as described previously
Oleoyl-CoA synthetase activity was determined from cell lysates as described
For each membrane fraction, 2.5% of the original amount of proteins obtained from the subcellular fractionation was loaded and separated on a 8% SDS–polyacrylamide gel. After electrophoresis, proteins were transferred to nitrocellulose membranes. Equal loading and transfer of samples were verified by Ponceau S staining. The membranes were blocked in 5% milk powder in TBS-Tween (50 mM Tris–HCl, pH 7.4, 138 mM NaCl, 2.7 KCl, 0.1% Tween-20) for 30 min and then incubated with primary antibodies in blocking buffer containing 5% milk powder. The membranes were then washed, incubated with horseradish peroxidase-conjugated secondary antibodies and the reaction was detected with an enhanced chemiluminescence system (Amersham Life Science, Buckinghamshire, UK). Quantification of Western blots was done with ImageJ 1.37v software (Wayne Rasband, NIH).
The acyl-CoA-synthetases FATP1, ACSVL4/FATP4 and ACSL1 have distinct functions in fatty acid metabolism which are possibly mediated by their specific subcellular localization
We decided to use a gain of function approach to characterize the function of the most relevant acyl-CoA-synthetases of adipocytes. For this purpose, we used a retroviral transfection system with subsequent antibiotic selection to yield a pool of 3T3-L1 cells expressing the respective enzymes. Expression levels were evaluated by quantitative PCR and Western blotting. As seen in
(A) Quantification of FATP1 and ACSVL4/FATP4 mRNA levels in overexpressing (3T3-FATP1, dark grey and 3T3-FATP4, light grey) and control adipocytes (3T3-pRJ, white bars) by efficiency corrected quantitative real-time PCR relative to general transcription factor 2b (Gtf2b). The relative increase of FATP1 mRNA level is 3.5-fold in 3T3-FATP1 (n = 2) and the increase of ACSVL4/FATP4 mRNA level is 6.9-fold in 3T3-FATP4 (n = 1) compared to 3T3-pRJ. (B, C, D) Analysis of protein expression by Western blotting of total cell lysates from FATP1, ACSVL4/FATP4 and ACSL1-FLAG overexpressing and control adipocytes. Densitometry indicates 11-fold overexpression of FATP1 (n = 1) and 9.0-fold overexpression of ACSVL4/FATP4 (n = 2).
The exact knowledge of the subcellular localization of FATP1 and ACSVL4/FATP4 is necessary for understanding their function, especially in light of the controversies over the fundamental mechanisms of fatty acid transport. Here, we focused on immunofluorescence microscopy to analyze the localization of FATP1 and ACSVL4/FATP4.
We combined immunofluorescence staining and expression of fluorescent organelle markers to investigate the localization via confocal laser scanning microscopy. As seen in
(A–D) Comparative analysis of the intracellular localization of FATP1. FATP1 (green: A, B + D/red: C; affinity purified rabbit anti FATP1) does not overlap with the plasma membrane marker CD36 (A) (red, mouse polyclonal anti CD36) or the mitochondrial marker ornithin-carbomoyl-transferase (C) (green, OCT-GFP; C-terminal coupled with GFP, expressed by nucleofection). FATP1 (green) co-localizes with ER marker Sec61β (red; C-terminal coupled with RFP, expressed by nucleofection) in adipocytes (B) and fibroblasts (D) as indicated by the yellow color in the overlap image. In adipocytes, a distinct area is covered by FATP1 but not ER (B). Representative images from single confocal sections are shown and dimension bars are 10 µm for all images. (E–H) Comparative analysis of the intracellular localization of ACSVL4/FATP4. ACSVL4/FATP4 (green: E, F + H/red: G); affinity purified rabbit anti FATP4) does not co-localize with CD36 (E) (red) or Tom20 (G) (green). ACSVL4/FATP4 (green) overlaps with Sec61β-RFP (red) in both adipocytes (F) and fibroblasts (H).
We analyzed the subcellular localization of ACSL1, an acyl-CoA-synthase that can drive fatty acid uptake
Our data show that ACSL1-FLAG is located on organelles with a worm-shaped pattern that is typical for mitochondria. This morphology is most striking at the stage of fibroblasts, when ACSL1-FLAG completely co-localizes with the mitochondrial marker GFP-OCT (
(A–D) Comparative analysis of the intracellular localization of FLAG -tagged ACSL1. (A) ACSL1-FLAG (green); mouse monoclonal anti FLAG M2) does not co-localize with CD36. (red). (B–C) Congruency of ASCL1-FLAG (red) with the mitochondrial marker OCT-GFP is excellent for lowly differentiated (B) but less complete for highly differentiated adipocytes (C). (D) In fibroblasts, ACSL1-FLAG overlaps completely with the mitochondrial marker.
To confirm that overexpression of FATP1, ACSVL4/FATP4 and ACSL1-FLAG results in functional acyl-CoA-synthetases, we measured the oleoyl-CoA-synthetase activity from lysates of the corresponding cell lines. Our results show that FATP1, ACSVL4/FATP4 and ACSL1-FLAG are functional because the total enzyme activity is enhanced in all cell lines (
(A) Oleoyl-CoA synthetase activity was determined from cell lysates of 3T3-pRJ control (white bars), 3T3-FATP1 (light grey), 3T3-FATP4 (dark grey) and 3T3-ACSL1-FLAG adipocytes (punctate) for ten minutes in the presence of 20 µM [3H]-oleate (specific activity 10 Ci/mol). * p<0.05 vs 3T3-pRJ; n = 4. (B) Oleate uptake was measured by incubating adipocytes for 3 h with 200 µM [3H]-oleic acid bound to 100 µM fatty acid free BSA (specific activity 0.5 Ci/mol). * p<0.05 vs 3T3-pRJ; n = 4.
Insulin is known to increase fatty acid uptake in adipocytes and a translocation of FATP1 was considered as the key mechanism behind this effect
We first measured the uptake of the physiological substrate oleate for 5 min, but could not detect an increase in uptake rate after insulin treatment (
(A+B) 3T3-FATP1 (dark grey bars), 3T3-FATP4 (light grey bars) and 3T3-pRJ control adipocytes (white bars) were pretreated with or without 1 µg/ml insulin for 10 min, followed by co-incubation with the with 340 µM [3H]-oleic acid bound to 170 µM fatty acid free BSA (specific activity 1 Ci/mol) for 5 min (A) or 170 µM [3H]-oleic acid bound to 85 µM fatty acid free BSA (specific activity 1 Ci/mol) for 60 min (B). * p<0,05; n = 3. (C+D) 3T3-FATP1 (dark grey bars), 3T3-FATP4 (light grey bars) and 3T3-pRJ control adipocytes (white bars) were pretreated with or without 1 µg/ml insulin for 18 min, followed by co-incubation with 2 µM Bodipy fatty acids (FABy12 (C); n = 3/FABy16 (D); n = 2) bound to 2 µM BSA for 2 min. Subsequent FACS analysis shows a more pronounced increase in geographical mean fluorescence signal after insulin treatment for 3T3-FATP1 (dark grey bars) and 3T3-FATP4 (light grey bars) compared to 3T3-pRJ control adipocytes (white bars), which is however restricted to specific Bodipy fatty acid species (FABy12 for 3T3-FATP4 (C) and FABy16 for 3T3-FATP1 (D)). Representative experiments are shown.
To investigate whether insulin induces a change of localization of FATP1 or ACSVL4/FATP4, we first used immunofluorescence microscopy. As seen in
(A+B) 3T3-FATP1 and 3T3-FATP4 were treated or non treated for 20 min with 1 µg/ml insulin, stained and analyzed for changes of localization with confocal laser scanning microscopy. No co-localization of FATP1 (A) or ACSVL4/FATP4 (B) with the plasma membrane localized CD36 was observed, neither for insulin treated nor for non treated cells. (C) Subcellular fractionation of wild-type 3T3-L1 adipocytes treated or non treated for 20 min with 1 µg/ml insulin. 20 µg of each subcellular fraction were applied. Upon insulin stimulation, a significant increase in GLUT4 was observed in the plasma membrane (PM) fraction with concomitant reduction in the low (LDM) and high density membrane fraction (HDM). This shift was not observed for FATP1 and ACSVL4/FATP4. FATP1, ACSVL4/FATP4 and the ER marker calnexin share the same distribution pattern. This pattern differs from the distribution of both, voltage-dependent anion-selective channel protein 1 (VDAC) that is localized primarily on mitochondria and partly on the plasma membrane and the plasma membrane localized sodium-potassium ATPase. Representative blots are shown.
In conclusion, our results show that short term incubation with insulin increases fatty acid uptake, but without changing the intracellular localization of FATP1 and ACSVL4/FATP4.
Glucose and fatty acid metabolism are intimately connected. Therefore we investigated whether overexpression of FATP1 and ACSVL4/FATP4 has an impact on insulin-mediated glucose uptake. The basal glucose uptake rate was not significantly different between 3T3-FATP1, 3T3-FATP4 or the control adipocytes. However, we found that the increase in glucose uptake upon insulin stimulation was significantly higher in 3T3-FATP1 (5.8-fold) and 3T3-FATP4 (7.6-fold) as compared to control adipocytes (2.9-fold), as shown in
3T3-FATP1 (dark grey bars) and 3T3-FATP4 adipocytes (light grey bars) and control adipo-cytes (white bars) were incubated with 1 µg/ml insulin for 20 min followed by 10 min of co-incubation with 0.1 mM 2-deoxy-D-glucose [DOG], 1 mCi/ml. Glucose uptake is significantly higher for insulin treated cells (* p<0,05) and this effect is more pronounced in FATP1 and ACSVL4/FATP4 overexpressing adipocytes (* p<0,05), n = 3.
With our work, we sought to resolve some of the ongoing controversies about the mechanism of FATP-mediated fatty acid transport. Our approach was to focus on identifying the subcellular localization of FATP1, ACSVL4/FATP4 and the acyl-CoA-synthetase ACSL1. We selected adipocytes as our model system because they are highly relevant for the pathogenesis of diseases while the localization of FATP1, ACSVL4/FATP4 and ACSL1 is poorly defined in this cell line.
Unlike previous studies, we chose to overexpress FATP1 and ACSVL4/FATP4 instead of knockdown by RNAi
We localized ACSVL4/FATP4 to the ER of 3T3-L1 adipocytes. This is consistent with previous studies that identified the same location for ACSVL4/FATP4 in different cell lines
In our opinion, two fundamental problems cause the incoherent reports concerning the localization of FATP1 and ACSVL4/FATP4 in adipocytes: first, the specific morphology of this type of cells and secondly, the experimental approaches used to localize the proteins. We observed that identifying organelles by confocal laser scanning microscopy was much easier and more distinctive for 3T3-L1 fibroblasts than for differentiated adipocytes. Unlike fibroblasts that spread horizontally, adipocytes primarily grow on the vertical axis due to the high cell density required for their differentiation. This growth pattern makes the correct identification of cell structures difficult. Furthermore, adipocytes incorporate several lipid droplets that scatter laser light, which affects the focus and the resolution of confocal microscopy. A high resolution however is required as the cytoplasm is tightly compressed by lipid droplets, posing a challenge for the identification of subcellular organelles. The second problem contributing to the confusion is the experimental approach. Immunofluorescence microscopy is limited by the drawbacks mentioned above. Another common approach is subcellular fractionation. Most fractionation protocols for adipocytes are optimized for GLUT4 translocation, but lack the specificity and resolution required for a clear-cut distinction between different cellular compartments. Similar to our own results, other groups found FATP1
Our results also show that the intracellular localization of overexpressed FATP1, ACSVL4/FATP4 and ACSL1 is sufficient to significantly enhance basal fatty acid uptake. This observation was already made by us and others for ACSVL4/FATP4
Insulin has been shown before to enhance the fatty acid uptake of adipocytes and other cell types. Our results support this observation even if the magnitude measured now is at the lower end compared to previous data. Interestingly, we found a significant increase of oleate uptake after a 60 min incubation with insulin and fatty acids but not after 15 min (
There are conflicting ideas about the molecular mechanism by which insulin enhances fatty acid uptake
If FATP1 and ACSVL4/FATP4 are not insulin sensitive fatty acid transporter proteins, and the localization of CD36 at the plasma membrane of adipocytes remains undisturbed, then how is the effect of insulin mediated? At present, we are not able to answer this question satisfyingly. A very recent discovery was that the enzyme activity of ACSVL4/FATP4 stably expressed in C2C12 muscle cells was increased in a manner dependent on insulin signaling
Our data also suggest that both overexpressed FATP1 and ACSVL4/FATP4 enhance the effect of insulin on fatty acid uptake (
Overexpression of FATP1 and ACSVL4/FATP4 also increased insulin-mediated glucose uptake (
Taken together, our results demonstrate that the intracellular localization of FATP1 and ACSVL4/FATP4 on the ER and ACSL1 on mitochondria is sufficient to enhance fatty acid uptake. Short term insulin treatment leads to an increased uptake of fluorescent fatty acids, which is more pronounced in FATP1 and ACSVL4/FATP4 overexpressing adipocytes, but not accompanied by a change of localization of either protein.
Our interpretation is that the FATP acyl-CoA synthetases are metabolically trapping intracellular fatty acids, and through this mechanism contribute to the efficiency of insulin mediated fatty acid uptake. It remains for future studies to determine if the adipocyte FATP acyl-CoA synthetases themselves are activated by insulin, or if they support insulin mediated fatty acid uptake more indirectly.
(TIF)
We would like to thank Christoph Thiele (University of Bonn, Germany), Susanne Mandrup and Ronni Nielsen (both from the University of Southern Denmark, Odense, Denmark) for providing 3T3-L1 cells and helping us with the differentiation protocol. Furthermore, we thank Dieter Stefan (Institute of Immunology, Heidelberg University, Germany) for helping with the FACS analysis, Simone Staffer (Department of Gastroenterology, Heidelberg University) for excellent technical assistance, David Bernlohr (University of Minnesota, USA) for providing FATP1 antibodies, and Wolfgang Stremmel (University of Heidelberg, Germany) for continuous support.