ATGL is a biosynthetic enzyme for fatty acid esters of hydroxy fatty acids

Abstract

Branched fatty acid (FA) esters of hydroxy FAs (HFAs; FAHFAs) are recently discovered lipids that are conserved from yeast to mammals^1,2. A subfamily, palmitic acid esters of hydroxy stearic acids (PAHSAs), are anti-inflammatory and anti-diabetic^1,3. Humans and mice with insulin resistance have lower PAHSA levels in subcutaneous adipose tissue and serum¹. PAHSA administration improves glucose tolerance and insulin sensitivity and reduces inflammation in obesity, diabetes and immune-mediated diseases^1,4-7. The enzyme(s) responsible for FAHFA biosynthesis in vivo remains unknown. Here we identified adipose triglyceride lipase (ATGL, also known as patatin-like phospholipase domain containing 2 (PNPLA2)) as a candidate biosynthetic enzyme for FAHFAs using chemical biology and proteomics. We discovered that recombinant ATGL uses a transacylation reaction that esterifies an HFA with a FA from triglyceride (TG) or diglyceride to produce FAHFAs. Overexpression of wild-type, but not catalytically dead, ATGL increases FAHFA biosynthesis. Chemical inhibition of ATGL or genetic deletion of Atgl inhibits FAHFA biosynthesis and reduces the levels of FAHFA and FAHFA-TG. Levels of endogenous and nascent FAHFAs and FAHFA-TGs are 80-90 per cent lower in adipose tissue of mice in which Atgl is knocked out specifically in the adipose tissue. Increasing TG levels by upregulating diacylglycerol acyltransferase (DGAT) activity promotes FAHFA biosynthesis, and decreasing DGAT activity inhibits it, reinforcing TGs as FAHFA precursors. ATGL biosynthetic transacylase activity is present in human adipose tissue underscoring its potential clinical relevance. In summary, we discovered the first, to our knowledge, biosynthetic enzyme that catalyses the formation of the FAHFA ester bond in mammals. Whereas ATGL lipase activity is well known, our data establish a paradigm shift demonstrating that ATGL transacylase activity is biologically important.

Fig. 1. FAHFA biosynthesis is increased in AG4OX SVF adipocytes and is sensitive to fluorophosphonate inhibitors.

a–d, Biosynthesis of 9-PAHSA (a), 9-POHSA (b), 9-OAHSA (c) and 9-C17:1HSA (d) in WT and AG4OX SVF adipocytes incubated with 0.1% DMSO (vehicle (Veh)), increasing concentrations of 9-HSA (a–c) or both C17:1 and 9-HSA (25 µM, each) (d). n = 6 wells per condition with cells pooled from 9 mice per genotype; mean ± s.e.m. *P < 0.003 versus WT, same treatment (t-test correcting for Holm–Sidak multiple comparisons). #P < 0.05, ###P < 0.0001 versus vehicle, same genotype, @P < 0.05, @@P < 0.0001 versus WT, same treatment (two-way analysis of variance (ANOVA)). ND, not detected. e, 9-C17:1HSA biosynthesis in 3T3-L1 adipocytes pre-incubated with the indicated concentrations of MAFP or 0.1% DMSO (Veh) for 1 h, and then co-incubated with C17:1 and 9-HSA (10 µM each) or 0.1% DMSO (Veh) for 2 h. n = 4 wells except for MAFP 6.75 µM group n = 3; mean ± s.e.m. *P < 0.05, **P < 0.0001 versus C17:1 and 9-HSA alone (one-way ANOVA). f, Biosynthesis of 9-PAHSA (left) and 9-OAHSA (right) in WT and AG4OX SVF adipocytes pre-incubated with MAFP or 0.1% DMSO for 1 h, and then co-incubated with 9-HSA (10 µM) or 0.1% DMSO (Veh) for 2 h. n = 6 wells per condition; mean ± s.e.m. *P < 0.003 versus WT, same treatment (t-test correcting for Holm–Sidak multiple comparisons). #P < 0.05, ##P < 0.0001 versus Veh, @P < 0.0001 versus 9-HSA alone, same genotype (two-way ANOVA). g, Structure of FP-alkyne (top) and schematic diagram of activity-based proteomics (bottom). Created with BioRender.com. h, ATGL protein levels in AG4OX SQ WAT, normalized to total protein between 37–25 kDa on western blot. n = 8 mice, mean ± s.e.m. *P < 0.0012 versus WT (t-test, two-tailed). Similar results were obtained in at least two independent experiments. See also Extended Data Fig. 1.

Fig. 2. ATGL regulates biosynthesis of FAHFA and TG-esterified FAHFA in intact cells.

Cells were incubated with both C17:1 FA and D₂₀-9-HSA (20 µM (a–d, i–l) or 25 µM (e–h) each) or 0.05% DMSO (vehicle) for 4 h. a, b, Biosynthesis of 9-PA-D₂₀HSA, 9-PO-D₂₀HSA and 9-OA-D₂₀HSA (a) and 9-C17:1-D₂₀HSA (b) was measured in WT SVF adipocytes pre-incubated with the indicated enzyme inhibitors or 0.1% DMSO (No inh). LPCATi, lysophosphatidylcholine acyltransferase inhibitor; LYPLA2i, lysophospholipase 2 inhibitor; LYPLA1i, lysophospholipase 1 inhibitor; CPT1i, carnitine palmitoyltransferase 1 inhibitor. n = 4 wells; mean ± s.e.m. *P <  0.001, **P < 0.0001 versus No inh group (one-way ANOVA). c, d, Biosynthesis of 9-PA-D₂₀HSA, 9-PO-D₂₀HSA and 9-OA-D₂₀HSA (c) and 9-C17:1-D₂₀HSA (d) in WT SVF adipocytes co-incubated with atglistatin (ATGL inhibitor (ATGLi), 10 µM) or 0.1% DMSO (No inh) for 0.5–4 h. n = 4; mean ± s.e.m. *P < 0.05, **P < 0.0001 versus 0.5 h, same treatment, #P < 0.05, ##P < 0.001, ###P < 0.0001 versus No inh, same time point (two-way ANOVA). e, Biosynthesis of 9-PA-D₂₀HSA, 9-OA-D₂₀HSA and 9-C17:1-D₂₀HSA in HEK293T cells transfected with GFP, WT ATGL and ATGL(S47A) mutant enzymes. n = 4, mean ± s.e.m. *P < 0.01, **P < 0.001 versus GFP-transfected cells (one-way ANOVA). f–h, Biosynthesis of 9-PAHSA, 9-POHSA, 9-OAHSA and 9-C17:1HSA (f), biosynthesis of TG-esterified 9-FAHFAs (g) and levels of FFA and 9-HSA (h) in 3T3-L1 adipocytes incubated with C17:1 or 9-HSA alone or with or without atglistatin (10 µM). n = 3; mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus vehicle, #P < 0.05, ##P < 0.01, ####P < 0.0001 versus C17:1 + 9-HSA group (one-way ANOVA). i–m, Biosynthesis of 9-PA-D₂₀HSA, 9-PO-D₂₀HSA and 9-OA-D₂₀HSA (i), biosynthesis of 9-C17:1-D₂₀HSA (j), biosynthesis of TG-esterified 9-FA-D₂₀HSA (k), levels of TG-esterified C17:1 and D₂₀HSA (l) and levels of FFA and HSA (m) in AT-Atgl-KO and Atgl^fl/fl SVF adipocytes incubated with atglistatin (10 µM) or 0.1% DMSO (No inh). TG-esterified 9-FA-D₂₀HSA levels were measured in n = 4 samples per condition. FFA and HSA levels were measured in the Veh condition; n = 4 mean ± s.e.m. *P < 0.05, **P < 0.0001 versus WT, same treatment, #P < 0.0001 versus No inh, same genotype (two-way ANOVA, i, j, t-test two-tailed, k). Levels above the limit of quantification are shown with white circles. Similar results were obtained in two independent experiments.

Fig. 3. ATGL transacylation activity synthesizes FAHFAs in vitro and in vivo.

a, Left, biosynthesis of 9-OAHSA in vitro by affinity-purified WT ATGL-288 (ATGL) and the cofactor CGI-58 from 9-HSA and acyl-donor TG(18:1) and inhibition with atglistatin (50 µM). Right, inhibition of the transacylation activity of affinity-purified full-length ATGL (ATGL_full) by the indicated doses of atglistatin. b, c, WT ATGL-288 catalysed biosynthesis of 9-C17:1HSA from 9-HSA and acyl donors TG(C17:1) and C17:1(FFA) (b) and biosynthesis of 9-PA-D₂₀HSA and 9-OA-D₂₀HSA from D₂₀-9-HSA and acyl donor TG(C16:0/18:1/18:1) (c). d, WT ATGL-288 and ATGL-288(S47A) transacylation activity catalysed 9-FAHFA biosynthesis from acyl donors: TG(18:2), TG(16:1), TG(18:1), DG(C18:1/18:1/0:0), phosphatidylcholine (PC(18:1/18:1/0:0)) and 18:1-CoA. e, Lipase activity of WT ATGL-288 (ATGL) and ATGL-288(S47A) enzymes as measured by OA release during the FAHFA biosynthesis assay with the substrate TG(18:1) in d. f, 9-OAHSA biosynthesis from TG(18:1) and D₂₀-9-HSA in human SQ WAT lysates. g, 9-FAHFA biosynthesis in human SVF adipocytes incubated with D₂₀-9-HSA or vehicle (0.025% DMSO). (n = 3, mean ± s.e.m. a–g), *P < 0.0001 versus no enzyme, #P < 0.0001 versus ATGL/CGI-58 and @P < 0.0001 versus ATGL (one-way ANOVA, a, d). $P <  0.02, $$P < 0.008 versus no enzyme (t-test, two-tailed c). AT-Atgl-KO and littermate Atgl^fl/fl female mice were used for h–v. h–k, Total endogenous PAHSA (h) and PAHSA isomer levels in SQ WAT (i), PG WAT (j), BAT (k), liver (l) and serum (m). n–q, Total TG-esterified PAHSA (n) and TG-esterified PAHSA isomer levels in SQ WAT (o), PG WAT (p) and liver (q). r, Adipose tissue and liver total triglyceride levels. AT-Atgl-KO and Atgl^fl/fl mice were injected with D₂₀-9-HSA (5 mg kg⁻¹) intraperitoneally and euthanized 4 h later (s–v). s, PG WAT, liver and serum levels of newly synthesized 9-PA-D₂₀HSA. t–v, 9-PA-D₂₀HSA enrichment (9-PA-D₂₀HSA per cent of total 9-PAHSA) (t), levels of newly synthesized TG-esterified 9-PA-D₂₀HSA (u), endogenous HSA substrate level (v, left) and enrichment of D₂₀-9-HSA (v, right) in PG WAT. n = 6 mice for all tissues in h–r; n = 5 mice for serum in h, m and s–v; mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus Atgl^fl/fl mice t-test, two-tailed (h–v). White error bars are shown within the bars in o, p. Only levels above the quantification limit are shown with white circles. Source data

Fig. 4. DGATs have an upstream role in ATGL-mediated FAHFA biosynthesis.

Cells were incubated with C17:1 FA and D₂₀-9-HSA (20 µM (a–c) or 25 µM (d, e) each) for 4 h. a–c, Biosynthesis of 9-PA-D₂₀HSA (a), 9-PO-D₂₀HSA (b) and 9-C17:1-D₂₀HSA (c) in WT and AG4OX SVF adipocytes with or without ATGL inhibitor (10 µM) and DGAT1 inhibitor (DGAT1i, 20 µM) or 0.2% DMSO (vehicle). n = 4 wells per condition; mean ± s.e.m. *P < 0.05, **P < 0.001, ***P < 0.0001 versus vehicle, same genotype, #P <  0.001, ##P < 0.0001 versus WT, same treatment, @P < 0.006, @@P < 0.0001 versus atglistatin alone (two-way ANOVA). d, Biosynthesis of 9-PA-D₂₀HSA, 9-PO-D₂₀HSA and 9-C17:1-D₂₀HSA in HEK293T cells transfected with control vector (Ctrl), ATGL and DGAT1 plasmids alone or together. n = 4 wells per condition except ATGL group n = 3; mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001 versus control vector t-test, two-tailed. These data are representative of three independent experiments. e, Biosynthesis of 9-PA-D₂₀HSA, 9-PO-D₂₀HSA, 9-OA-D₂₀HSA and 9-C17:1-D₂₀HSA in WT SVF adipocytes with or without DGAT2 inhibitor (DGAT2i, 20 µM) and Dgat1-KO SVF adipocytes. n = 4 wells per condition. Mean ± s.e.m. *P < 0.01, **P < 0.001, ***P < 0.0001 versus WT (one-way ANOVA). f, Model for ATGL transacylation activity synthesizing FAHFAs. Canonical ATGL lipase activity transfers an acyl chain of TG to a water molecule (H₂O) thereby releasing a FA. In the presence of TG and HFA, ATGL transacylation activity transfers an acyl chain of TG to HFA, synthesizing FAHFA. Our data show that the lipase activity of ATGL (Fig. 3e) is much higher than its biosynthetic activity (Fig. 3d) in vitro.

Extended Data Fig. 1. (related to Figure 1): (a-e) FAHFA biosynthesis is sensitive to fluorophosphonate inhibitors; (f) ATGL protein is increased in AG4OX mice adipose tissue.

(a) 9-PAHSA biosynthesis in 3T3-L1 adipocytes pre-incubated with the indicated concentrations of methyl arachidonyl fluorophosphonate (MAFP) or 0.1% DMSO (Veh) for 1hr, and then co-incubated with C17:1 and 9-HSA or 0.1% DMSO (Veh) for 2hrs. (n = 4 wells, except for MAFP 6.75 µM group n = 3, mean±SEM), *P < 0.0001 vs. C17:1 and 9-HSA only group (One-way ANOVA) (b) 9-POHSA biosynthesis in WT and AG4OX SVF-adipocytes pre-incubated with MAFP or 0.1% DMSO for 1hr, then co-incubated with 9-HSA or 0.1% DMSO (Veh) for 2hrs. (n = 6 wells, mean±SEM) *P < 0.05, **P < 0.0001 vs. WT, same treatment, t-test correcting for Holm-Sidak multiple comparisons. #P < 0.001, ##P < 0.0001 vs. the Veh for 9-HSA, same genotype, @P < 0.05 vs. 9-HSA alone, same genotype (Two-way ANOVA). (c) 9-PAHSA and (d) 9-POHSA biosynthesis in WT and AG4OX SVF-adipocytes pre-incubated with FP-alkyne (5 µM) or 0.1% DMSO for 1hr, then co-incubated with 9-HSA (10 µM) for 2hrs. (n = 4 except for AG4OX Vehicle group = 6, mean±SEM) *P < 0.001 vs. WT, same treatment, #P < 0.01 vs. Veh, same genotype (Two-way ANOVA). (e) Selected list of candidate FAHFA biosynthetic enzymes from activity-based proteomics study in WT and AG4OX SVF adipocytes. (f) WT and AG4OX SQ WAT western blots for ATGL and GAPDH. Total protein between 37 and 25 kDa was visualized by Revert stain (N = 8/genotype). Loading control, GAPDH, and ATGL were run on the same blot. This blot is representative of 3 separate blots. ND means not detected.

Extended Data Fig. 2. (related Figure 2): Effect of ATGL inhibition on FAHFA biosynthesis and cellular glycerolipids.

(a) Western-blots confirming overexpression of catalytically dead ATGL S47A and ATGL WT in HEK293T cells. Similar results were obtained in 3 independent experiments (b) Triglyceride (c), diglyceride and (d) monoglyceride levels in 3T3-L1 adipocytes incubated with C17:1 (25 µM), 9-HSA (25 µM) with or without atglistatin (ATGLi, 10 µM) or 0.1%DMSO. n = 3 wells, mean ± SEM.

Extended Data Fig. 3. ATGL regulates FAHFA biosynthesis in intact cells.

(a) 12-PAHSA, 12-POHSA, 12-OAHSA, (b) 5-PAHSA, 5-POHSA and 5-OAHSA biosynthesis in Atgl^fl/fl and AT-Atgl-KO SVF-derived adipocytes incubated with 12-HSA, 5-HSA or 0.05% DMSO (Vehicle) for 4hrs. n = 5 wells, mean±SEM. *P < 0.002, **P  = 0.001, ***P < 0.0001, vs. Atgl^fl/fl, same treatment, #P < 0.01, ##P < 0.0001 vs. 25µM HSA same genotype (Two-way ANOVA a-b). ND means not detected.

Extended Data Fig. 4. (related to Figure 3 a-e) Coomassie stain of affinity-purified ATGL-288 WT and ATGL-288[S47A] mutant enzymes.

(a) Coomassie-stained SDS gel showing induction and purification of ATGL-288 WT enzyme from E. coli Arctic Express. (b) Coomassie-stained SDS gel showing affinity-purified WT ATGL-288 and mutant ATGL-288[S47A] enzymes. Depiction of ATGL-288[S47A] plasmid construct and host e-coli culture (bottom panel).

Extended Data Fig. 5. (related to Figure 3): PG WAT and serum 5-PAHSA, free and TG-esterified OAHSA and POHSA/OAHPA levels are lower in adipose tissue of AT-Atgl-KO mice.

Atgl-KO and littermates Atgl^flfl female mice (10-11 weeks old) were sacrificed in the ad-lib fed state for experiments in panels (a-n). Endogenous 5-PAHSA levels in (a) PG WAT and (b) serum. Endogenous OAHSA isomers levels in (c) SQ WAT, (d) PG WAT, (e) BAT and (f) liver. Tissue levels of (g) total OAHSA and (h) total POHSA/OAHPA. TG-esterified OAHSA isomer levels in (i) SQ WAT, (j) PG WAT and (k) liver. Tissue levels of (l) total TG-esterified OAHSA and (m) total TG-esterified POHSA/OAHPA. (n) Adipose and liver tissue weights. (n = 5 per genotype for serum samples and 6 for all other tissues, mean ± SEM), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 vs. FL/FL mice, t-test, two-tailed (a-n). Fewer than 5 samples for panel b or 6 samples for panels a and c-n indicates the other samples were below the limit of quantification for our MS method. ND means not detected. Source data

Extended Data Fig. 6. Newly synthesized FAHFAs are present in mouse liver and are synthesized by human hepatocytes.

(a) WT mice were injected with FAHFA substrate D₂₀-9-HSA (25mg/kg) IP and sacrificed 4hrs later to collect serum and tissues. WAT, liver and serum levels of newly synthesized 9-PA-D₂₀HSA. (n = 6 mice per condition, mean±SEM) (b) 9-FAHFA biosynthesis in intact HEPG2 cells (human hepatocyte) incubated with D₂₀-9-HSA, C17:1 or Vehicle (0.05% DMSO) for four hrs. (n = 4 wells per condition, mean±SEM), *P < 0.008 vs. 25µM substrate group t-test, two-tailed. ND means not detected. Source data

Extended Data Fig. 7. DGAT2 overexpression increases 12-FAHFA biosynthesis in HEK293T cells.

(a) 12-PAHSA, 12-OAHSA and 12-C17:1HSA biosynthesis in DGAT2-overexpressing HEK293T cells incubated with C17:1 and 12-HSA (25µM each) for 2hrs. n = 3 wells per condition, mean±SEM *P<0.03, **P<0.025 vs GFP group, t-test, two tailed. Representative of 3 independent experiments. Western blot confirming overexpression of DGAT1 (b) and DGAT2 (c) in HEK 293T cells.