Fatty Acid-binding Proteins Interact with Comparative Gene Identification-58 Linking Lipolysis with Lipid Ligand Shuttling

Abstract

The coordinated breakdown of intracellular triglyceride (TG) stores requires the exquisitely regulated interaction of lipolytic enzymes with regulatory, accessory, and scaffolding proteins. Together they form a dynamic multiprotein network designated as the "lipolysome." Adipose triglyceride lipase (Atgl) catalyzes the initiating step of TG hydrolysis and requires comparative gene identification-58 (Cgi-58) as a potent activator of enzyme activity. Here, we identify adipocyte-type fatty acid-binding protein (A-Fabp) and other members of the fatty acid-binding protein (Fabp) family as interaction partners of Cgi-58. Co-immunoprecipitation, microscale thermophoresis, and solid phase assays proved direct protein/protein interaction between A-Fabp and Cgi-58. Using nuclear magnetic resonance titration experiments and site-directed mutagenesis, we located a potential contact region on A-Fabp. In functional terms, A-Fabp stimulates Atgl-catalyzed TG hydrolysis in a Cgi-58-dependent manner. Additionally, transcriptional transactivation assays with a luciferase reporter system revealed that Fabps enhance the ability of Atgl/Cgi-58-mediated lipolysis to induce the activity of peroxisome proliferator-activated receptors. Our studies identify Fabps as crucial structural and functional components of the lipolysome.

Keywords: adipose triglyceride lipase (Atgl); fatty acid-binding protein; lipid signaling; lipolysis; peroxisome proliferator-activated receptor (Ppar).

FIGURE 1.

A-Fabp interacts with Cgi-58. a, in solid phase assays, polystyrene plates were coated with purified A-Fabp and incubated with lysates of COS-7 cells transfected with plasmids encoding His-tagged murine Atgl, murine or human Cgi-58, murine Hsl, and as control β-galactosidase (LacZ). Bound proteins were detected using anti-His primary and Hrp-conjugated secondary antibody. Plates were developed using tetramethylbenzidine as substrate, and the absorbance was measured at 450/620 nm. Absorbance values were normalized to the relative expression levels of prey proteins determined by densitometry. b, Western blot analysis of respective cell lysates using anti-His primary and Cy3-conjugated secondary antibody. Fluorescence was quantified in a typhoon instrument. c, Cgi-58 interacts with various Fabp isoforms. Polystyrene plates were coated with purified A-, H-, L-, I-, or E-Fabp and incubated with COS-7 cell lysates (10, 30, and 60 μg total protein) containing His-tagged murine Cgi-58, human Cgi-58, and as control LacZ, expressed at comparable levels. Binding of proteins was detected using anti-His primary and Hrp-conjugated secondary antibody. d, A-Fabp/Cgi-58 interaction demonstrated by microscale thermophoresis. Ten μmol of purified and fluorescently labeled Cgi-58 was titrated against increasing amounts of unlabeled A-Fabp and fluorescence distribution inside the capillary determined using the Monolith NT.115. F_norm, normalized fluorescence. e, co-immunoprecipitation of endogenous A-Fabp and Cgi-58. White adipose tissue from overnight fasted mice was lysed and incubated with antibodies directed against A-Fabp, Cgi-58, or Gfp as negative control. Antibody/antigen complexes were precipitated using protein A/G-Sepharose and analyzed together with the lysate (= input) for the presence of antigens by Western blotting. Data are shown as mean ± S.D. (n = 4) and are representative of three independent experiments. Statistical difference was determined as compared with LacZ control (*, p < 0.05; **, p < 0.01; ***, p < 0.001) and of 60 and 30 μg versus 10 μg of each lysate (§, p < 0.05; §§, p < 0.01; §§§, p < 0.001).

FIGURE 2.

Mapping of the A-Fabp/mCgi-58 binding interface by NMR titrations. Two-dimensional ¹H-¹⁵N HSQC experiments result in a spectrum showing peaks (with their intensity values depicted in contour plots) for each N-H pair. Each peak represents a ¹H-¹⁵N pair, with the respective ¹H and ¹⁵N resonance frequencies. Consequently, the resulting peaks correspond mostly to backbone amide groups of A-Fabp. Assignment of the peaks to individual backbone residues was carried out upon analysis of triple resonance experiments. a, ¹H-¹⁵N HSQC spectra were recorded for 50 μm ¹⁵N-labeled A-Fabp in free form (cyan) and in complex with 100 μm m^solCgi-58 (magenta). Protein/protein interaction induces changes in the chemical environment of ¹H-¹⁵N pairs in the binding interface and results in intensity perturbations of the peaks corresponding to the involved residues. b, selected regions of the spectrum show a subset of peaks representing free A-Fabp in cyan and A-Fabp complexed with mCgi-58 (magenta). Intensity changes in peaks corresponding to residues involved in the A-Fabp/mCgi-58 interface ranged from a significant decline of intensity to full disappearance of the magenta peaks. c, ribbon and surface representations of the three-dimensional structure of A-Fabp (Protein Data Bank code 3Q6L). Residues in blue correspond to residues experiencing a significant decrease in peak intensity upon addition of m^solCgi-58. The inset shows the primary sequence of helix α1 and helix α2 with the affected residues highlighted in blue. Structure representations were prepared with PyMOL (PyMOL Molecular Graphics System, Version 1.5.0.4 Schrödinger, LLC.).

FIGURE 3.

Binding of Cgi-58 to various mutant A-Fabp variants. a, in solid phase assays, polystyrene plates were coated with 3 μg of purified wild type A-Fabp or the mutant variants A-Fabp^F58S, A-Fabp^K59E, A-Fabp^D77G, and A-Fabp^D78G and incubated with COS-7 lysates (60 μg of total protein) containing murine His-Cgi-58. Bound proteins were detected using anti-His primary and Hrp-conjugated secondary antibody. Plates were developed using tetramethylbenzidine as substrate, and the absorbance was measured at 450/620 nm. b, solid phase assays detecting the binding of wild type A-Fabp or the mutant variants A-Fabp^D18K, A-Fabp^K22E, and A-Fabp^R31E as well as the double mutant A-Fabp^K22E,R31E to murine His-Cgi-58 expressed in COS-7 cells. Each mutant variant features an exchange in one or two amino acids involved in binding of Hsl to A-Fabp. The mutant variant A-Fabp^R127Q, which is not capable of binding FAs, was also tested for binding to Cgi-58. Data are shown as mean ± S.D. (n = 4) and are representative of three independent experiments. Statistical difference was determined as compared with A-Fabp control (n.s., nonsignificant; *, p < 0.05; **, p < 0.01; ***, p < 0.001).

FIGURE 4.

A-Fabp does not compete with Atgl for Cgi-58 binding. In solid phase assays, polystyrene plates were coated with 1 μg of purified Atgl and incubated with bacterial lysates (60 μg of total protein) containing LacZ, A-Fabp, or mCgi-58 together with increasing amounts of purified Strep-A-Fabp (0.5-, 1-, or 5-fold molar excess compared with Atgl). Bound proteins were visualized by anti-Strep primary and Hrp-conjugated secondary antibody. Plates were developed using tetramethylbenzidine as substrate, and the absorbance was measured at 450/620 nm. Data are shown as mean ± S.D. (n = 4) and are representative of three independent experiments. Statistical difference was determined as compared with LacZ control (***, p < 0.001).

FIGURE 5.

Identification of Cgi-58 regions involved in binding to A-Fabp. a, in solid phase assays, polystyrene plates were coated with purified A-Fabp and increasing amounts of purified Gst, Gst-Cgi-58, or N-terminally truncated Cgi-58 variants. Binding of proteins was detected using anti-Gst primary and Hrp-conjugated secondary antibody. Plates were developed using tetramethylbenzidine as substrate, and the absorbance was measured at 450/620 nm. b, SDS-PAGE of purified Gst, Gst-Cgi-58, and N-terminally truncated Cgi-58 variants after Coomassie Brilliant Blue staining. c, solid phase assay detecting the interaction between purified A-Fabp and His-tagged human Cgi-58, the naturally occurring mutant variants of human Cgi-58, and LacZ (control) contained in lysates of transfected COS-7 cells. d, Western blot analysis of respective cell lysates using anti-His primary and Cy3-conjugated secondary antibody. Statistical difference was determined as compared with Gst/LacZ control (*, p < 0.05; **, p < 0.01; ***, p < 0.001) and of 0.6/60 and 0.3/30 μg versus 0.1/10 μg of each lysate (§, p < 0.05; §§, p < 0.01; §§§, p < 0.001).

FIGURE 6.

A-Fabp promotes lipolysis solely in the presence of Cgi-58. a, TG hydrolase activities of cell lysates derived from COS-7 cells expressing His-tagged LacZ (control), Atgl, Cgi-58, and/or A-Fabp were determined by incubating lysates with phospholipid-emulsified triolein substrate containing radiolabeled triolein as tracer. After extraction of free FAs, radioactivity was determined by liquid scintillation counting. b, Western blot analysis of respective COS-7 lysates using anti-His primary and Hrp-conjugated secondary antibody. c, TG hydrolase activities of lysates prepared from differentiated 3T3-L1 adipocytes. Lysates were incubated with artificial substrate as in a in presence of increasing concentrations of the A-Fabp-inhibitor BMS309403 (Fi). d, TG hydrolase activities of lysates prepared from murine gonadal WAT. Lysates were incubated with artificial substrate as in a in presence of BMS309403 (Fi; 100 μm), the Atgl-inhibitor Atglistatin (Ai; 20 μm), and/or the Hsl inhibitor 76-0079 (Hi; 10 μm). Data are shown as mean ± S.D. (n = 3) and are representative of three independent experiments. Statistical difference was determined as compared with control (***, p < 0.001).

FIGURE 7.

L-Fabp promotes Pparα signaling. a, Western blot analysis of COS-7 cells that were transfected with a reporter plasmid coding for firefly luciferase under transcriptional control of multiple Ppar-responsive elements (PPRE), a plasmid coding for Renilla luciferase (transfection control), and a single plasmid or a mixture of plasmids coding for Pparα, Atgl, L-Fabp, Cgi-58, or the catalytically inactive Atgl^S47A (as indicated). b, for firefly luciferase gene transactivation assays, relative luminescence units of firefly and Renilla luciferase were determined in a 24-well plate luminometer and calculated relative to Renilla luciferase activities (control values were set to 100%). c, TG hydrolase activity (as in Fig. 6) assay and data are shown as mean ± S.D. (n = 4) and are representative of three independent experiments. Statistical difference was determined as compared with control (*, p < 0.05; **, p < 0.01; ***, p < 0.001).

FIGURE 8.

A-Fabp promotes Pparγ signaling. a and b, for firefly luciferase gene transactivation assays, COS-7 cells were transfected with the reporter plasmid coding for firefly luciferase under transcriptional control of multiple PPRE, a plasmid coding for Renilla luciferase (transfection control), and a plasmid or a mixture of plasmids coding for Pparγ, Atgl, A-Fabp, Cgi-58, the non-FA binding mutant variant A-Fabp^R127Q, or the mutant variant A-Fabp^F58A incapable of ligand-induced nuclear translocation (as indicated). Relative luminescence units of firefly and Renilla luciferase were determined in a 24-well plate luminometer and calculated relative to Renilla luciferase activities (control values were set to 100%). c, Western blot analysis of cytoplasmic and nuclear fractions obtained from differentiated 3T3-L1 adipocytes under nonstimulated (= basal, −F) and forskolin-stimulated (+F) conditions. A-Fabp was detected using an antibody specific for A-Fabp and Hrp-conjugated secondary antibody. Purity of the fractions was assessed by Western blotting using anti-histone H3 and anti-Gapdh antibodies, respectively. Data are shown as mean ± S.D. (n = 4) and are representative of three independent experiments. Statistical difference was determined as compared with control (*, p < 0.05; **, p < 0.01; ***, p < 0.001).