Acetate is produced at a level comparable with ketone bodies in energy stresses.

(A) Enrichment of glucose, 3-HB, AcAc and acetate in clinical serum samples from healthy volunteers and patients with diabetes mellitus (Health, n=8; Diabetes, n=17).

(B) Enrichment of glucose, 3-HB, AcAc and acetate in the serum of STZ-induced diabetic mice (C57BL/6, n=5). (C) The levels of acetate, 3-HB, AcAc and glucose in the serum of C57BL/6 mice (n=5) starved for indicated time course. Abbreviations: 3-HB, 3-hydroxybutyrate; AcAc, acetoacetate; NT, untreated control; STZ, streptozotocin.

Values are expressed as mean±SD and analyzed statistically by two-tailed unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Acetate is derived from FFAs in mammalian cells.

(A) The amount of U-13C-acetate secreted by indicated cells cultured in U-13C-palmitate-containing HBSS for 20 h (n=3). (B-D) The amount of U-13C-acetate secreted by MPH (B), LO2 (C) and AML12 (D) cells cultured in HBSS supplemented with increasing doses of U-13C-palmitate for 20 h (n=3). (E) Enrichment of acetate in the serum of untreated or STZ-induced diabetic C57BL/6 mice (n=10) fed with or without high fat diet (HFD). Abbreviations: MPH, mouse primary hepatocytes; UD, undetectable; STZ, streptozotocin.

Values are expressed as mean±SD and analyzed statistically by two-tailed unpaired Student’s t test (A, E) or one-way ANOVA (B-D), individually (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

ACOT12 and ACOT8 are involved in acetate production in mammalian cells.

(A) Heatmap showing hepatic differentially expressed genes between fed group and fasted group, RNAseq analysis data from Goldstein et al. (2017). (B) The secretion of acetate (upper panel) by HEK-239T cell lines overexpressing various ACOTs and the protein levels of expressed ACOTs (lower panel). (C, D) HEK-293T (C) and Huh7 (D) cell lines overexpressing control vector, wildtype (WT) ACOT12 and ACOT8 or their enzyme activity-dead mutants (Mut) were cultured in HBSS containing U-13C-palmitate for 20 h, followed by detection of U-13C-acetate. (E, F) U-13C-acetate secreted by ACOT12-or ACOT8-knockdown MPH after incubation in U-13C-palmitate-containing HBSS for 20 h. Abbreviations: shACOT12, short hairpin RNA targeting mouse ACOT12 gene; shACOT8, short hairpin RNA targeting mouse ACOT8 gene. UD, undetectable; ACOT8 Mut, ACOT8 H78A mutant; ACOT12 Mut, ACOT12 R312E mutant.

Values are expressed as mean±SD (n=3) of three independent experiments and analyzed using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference)

ACOT12 and ACOT8 are responsible for acetate production in energy stresses.

(A, C) ACOT12 in mice (C57BL/6) liver was knocked down by adenovirus-based shRNA, followed by detection of ACOT12 protein with Western Blot (A) and evaluation of knockdown efficiency by calculating ACOT12 level relative to β-actin (C). (B, D) The knockdown efficiency of ACOT8 was determined as that of ACOT12. (E) Enrichment of serum acetate in normal and 16 h fasting mice (C57BL/6) with adenovirus-mediated knockdown of ACOT12 or ACOT8 in liver. (F) Enrichment of serum acetate in STZ-induced diabetic mice (C57BL/6) with adenovirus-mediated knockdown of ACOT12 or ACOT8 in liver. (G, H) ACOT12 (G) or ACOT8 (H) in mice (C57BL/6) liver was conditionally deleted by Cre-Loxp in liver, followed by detection of ACOT12 and ACOT8 protein with Western Blot. (I, J) Enrichment of serum acetate in normal and 16 h fasting mice (C57BL/6) with Cre-Loxp-mediated conditional deletion of ACOT12 (I) or ACOT8 (J) in liver.

Results are expressed as mean±SD of three independent experiments in (C, D), n=10 mice per group in (E, F) and n=6 mice per group in (I, J), and analyzed by using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Acetate production is dependent on FFAs oxidation in both mitochondrion and peroxisome.

(A) Co-immunostaining of Flag-ACOT8 with peroxisome marker catalase and Flag-ACOT12 with cytosol marker GAPDH in LO2 cells. Nuclei were stained with DAPI. Scale bars represent 10 μm. (B) The protein levels of ACOT12 and ACOT8 in the subcellular fractions of MPH cells. Abbreviations: Lyso, lysosome; ER, endoplasmic reticulum; Mito, mitochondria; Perox, peroxisome. (C, D) U-13C-acetate production (C) and the relative β-oxidation rate (D) in carnitine palmitoyltransferase 1A (CPT1A)-knockdown LO2 cells cultured in HBSS containing U-13C-palmitate for 20 h. (E) U-13C-acetate production (left) and the relative β-oxidation rate (right) of LO2 cells cultured in U-13C-palmitate-containing HBSS w/wo CPT1 inhibitor etomoxir (20 μM) for 20 h. (F) U-13C-acetate production in ATP citrate lyase (ACLY)-knockdown LO2 cells cultured in HBSS supplemented with U-13C-palmitate for 20 h. (G) U-13C-acetate production in ATP binding cassette subfamily D member 1 (ABCD1)-knockdown LO2 cells cultured in HBSS containing U-13C-palmitate for 20 h. (H) A schematic diagram depicting the mitochondrion and peroxisome pathways of acetate production from FFAs oxidation in hepatocytes. Very long- and long-chain fatty acids (VL/LCFAs) is transported through ABCD1 into peroxisome where it is further degraded into medium-chain fatty acids (MCFAs) via fatty acid oxidation (FAO) process, accompanied by production of acetyl-CoA which is further converted to acetate by peroxisome-localized ACOT8. MCFAs generated in peroxisome are exported into cytosol and absorbed directly by mitochondria. Cytosolic acyl-CoA derived from medium- and long-chain fatty acids (M/LCFAs) is transferred into mitochondria through CPT1A. All fatty acids and acyl-CoA in mitochondria undergo FAO to be degraded to acetyl-CoA. Then acetyl-CoA together with oxaloacetate is synthesized to citrate in TCA cycle, and citrate is exported into cytosol where it is lysed to acetyl-CoA by ACLY. Acetyl-CoA is finally converted to acetate by cytosol-localized ACOT12.

Values in (C-G) are expressed as mean±SD (n=3) of three independent measurements. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 by two-tailed unpaired Student’s t test.

ACOT12 and ACOT8 serve to maintain CoA pool for sustained FAO.

(A, B) MPHs knocked down for ACOT12 (A) or ACOT8 (B) were cultured in glucose free reaction buffer containing 0.8 μCi/mL [9,10-3H(N)]-oleic acid for 20 h, followed by determination of the relative β-oxidation rate. (C) Relative abundance of reduced CoA in MPHs knocked down for ACOT12 or ACOT8. (D) Relative abundance of acetyl-CoA in MPHs knocked down for ACOT12 or ACOT8. (E) The ratio of reduced CoA to acetyl-CoA in MPHs knocked down for ACOT12 or ACOT8. (F) Relative abundance of reduced CoA and various oxidized CoA in MPHs. Abbreviations: Ac-CoA, acetyl-CoA; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA.

Values are expressed as mean±SD (n=3) of three independent experiments and analyzed using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

ACOT12 and ACOT8 are required for ketone bodies’ production in STZ-induced diabetes.

(A) Relative abundance of HMG-CoA in MPHs knocked down for ACOT12 or ACOT8 (n=3). (B, C) Serum levels of AcAc (B) and 3-HB (C) in STZ-induced diabetic C57BL/6 mice with adenovirus-mediated knockdown of ACOT12 or ACOT8 in liver. (D, E) The protein levels of HMGCS2 in MPHs knocked down for ACOT12 (D) and ACOT8 (E). (F, G) ACOT12 (F) and ACOT8 (G) in mice (C57BL/6) liver were knocked down by adenovirus-based shRNA, followed by detection of HMGCS2 protein with Western Blot. (H) Western Blot (upper panel) and evaluation of the relative acetylation (Ac-Lys) level by calculating HMGCS2 acetylation relative to HMGCS2 (lower panel). Abbreviations: HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; HMGCS2, 3-hydroxy-3-methylglutaryl-CoA synthase 2.

Results are expressed as mean±SD of three independent experiments in (A) and n=10 mice per group in (B, C), and analyzed by using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Brain exhibits increased acetate consumption during energy stresses.

(A-G) Relative abundance of 13C-acetyl-CoA (A), 13C-citrate (B), 13C-aconitate (C), 13C-isocitrate (D), 13C-succinate (E), 13C-fumarate (F) and 13C-malate (G) in the brain of starved or diabetic mice (C57BL/6) was determined 1 h after intraperitoneal injection of 2-13C-acetate (310 mg/kg). (H) The abundance of acetate and 3-HB in the serum of fasting mice (C57BL/6) after intraperitoneal injection (acetate 300 mg/kg, 3-HB 520 mg/kg). (I) A working model describing the biological significance of ACOT12- and ACOT8-catalized conversion of acetyl-CoA to acetate and CoA. In the status of energy stress such as diabetes mellitus and prolonged starvation, body takes at least two advantages by converting acetyl-CoA to acetate and CoA: 1) CoA is required for sustained FAO and ketone bodies production in liver; 2) acetate serves as a novel ketone body to fuel extrahepatic tissues, particularly brain.

Values in (A-H) are expressed as mean±SD (n=5 mice per group in (A-G) and n=7 mice per group in (H)) and analyzed statistically by employing unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Increased level of acetate in diabetes mellitus.

(A) Typical 2D 1H-13C HSQC spectrum of clinical serum sample. (B) The levels of serum glucose, 3-HB, AcAc and acetate in STZ-induced diabetic mice (BALB/c, n=5 per group). (C) The levels of serum glucose, 3-HB and acetate in db/db mice (C57BL/6, n=6 per group). Abbreviations: 3-HB, 3-hydroxybutyrate; AcAc, acetoacetate; STZ, streptozotocin.

Values in (B, C) are expressed as mean±SD and analyzed statistically by two-tailed unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Increased level of acetate in fasting mice.

The levels of acetate, 3-HB, AcAc and glucose in the serum of BALB/c mice starved for indicated time course.

Values are presented as mean±SD (n=5). Statistics were performed employing two-tailed unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Acetate is increased independently of gut microbiota upon energy stress.

(A) The levels of serum glucose and acetate of C57BL/6 mice pretreated with antibiotics for 3-weeks and then starved for additional 24 h (n=5). (B) The levels of serum glucose and acetate of BALB/c mice pretreated with antibiotics for 3-weeks and then starved for another 12 h (n=5). (C) Serum glucose and acetate levels of STZ-induced diabetic mice pretreated with antibiotics for 3-weeks (C57BL/6, n=5).

Values are expressed as mean±SD. Statistical analyses were carried out by using two-tailed unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Acetate is secreted by in vitro cultured mammalian cells.

(A) Acetate secreted by various cells cultured in fresh DMEM medium for 20 h was detected by employing NMR. (B) A typical NMR 2D 1H-13C HSQC spectrum of culture medium in which HCT116 cells were culture for 20 h. (C, D) The same cells as in (A) were detected for the production of acetate, propanoate and butyrate with GC-MS. Propanoate and butyrate were used as an internal control.

Values in (A, C) are expressed as mean±SD (n=3) of three independent experiments.

Acetate is derived from other nutrients besides glucose.

(A) The amount of U-13C-acetate secreted by indicated cells cultured in U-13C-glucose-containing medium for 20 h. Values are expressed as mean±SD (n=3) of three independent experiments. UD, undetectable. (B) 1D 1H NMR CPMG spectra of aqueous extracts from medium of LO2 cells cultured with U-13C-glucose for 20 h.

ES-Acetate is mainly derived from FFAs.

(A) A representative 1D 1H NMR CPMG spectra of aqueous extracts from the DMEM medium with 4×AAs in which LO2 cells were cultured for 20 h. (B, C) The acetate production of primary hepatocyte (MPH) (B) and LO2 (C) cells cultured in Hanks’ balanced salt solution (HBSS, free for glucose, fatty acids and amino acids) supplemented with or without indicated doses of amino acids (AAs) for 20 h. (D, E) The acetate production of MPH (D) and LO2 (E) cells cultured in HBSS with or without 500 μM of indicated free fatty acids (FFAs) (C14:0, Myristate; C16:0, Palmitate; C18:0, Stearate) for 20 h.

Values are expressed as mean±SD (n=3) of three independent measurements for (B-E). Statistics in (B-E) were analyzed by using two-tailed unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

ACOT12 and ACOT8 are responsible for acetate production during energy stress.

(A) Volcano plot of RNAseq analysis data from Goldstein et al. (2017) (Goldstein et al., 2017). Taxa with fold change >2 and p-value< 0.05 are labeled in red and taxa with fold change < -2 and p-value < 0.05 are labeled in green. (B, C) ACOT12 and ACOT8 in livers of C57BL/6 mice with STZ-induced type I diabetes or 48 h starvation were detected by Western Blot (B) and their protein levels relative to β-actin are analyzed (C). ACOT, acyl-CoA thioesterase. (D, E) HEK-293T (D) and Huh7 (E) cell lines overexpressing ACOT12 or ACOT8 were cultured in medium containing indicated FFAs for 20 h, followed by detection of acetate secretion (n=3). (F) The enrichment of U- 13C-acetate in LO2 cells knocked down and further rescued for ACOT12 expression and then cultured in medium supplemented with U-13C- palmitate for 20 h (n=3). (G) The enrichment of U-13C-acetate in LO2 cell lines knocked down for ACOT8 and cultured in medium containing U-13C-palmitate for 20 h (n=3).

Values are expressed as mean±SD of three independent experiments. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 by two-tailed unpaired Student’s t test.

The expression profile of ACOTs and ketogenetic enzymes in human liver.

Heatmap from GTEx database represents the expression levels of ACOTs and ketogenetic genes in a variety of normal human tissues.

The expression profile of ACOTs and ketogenetic enzymes in mouse liver.

(A) Heatmap from GEO database represents the expression levels of ACOTs and ketogenetic genes in normal mouse tissues indicated. (B) The protein levels of ACOT12 and ACOT8 in different tissues of mice (C57BL/6).

FFA-derived acetate is diminished by supplementation of glucose.

(A, B) NMR detection of the amount of U-13C-acetate secreted by MPH (A) and LO2 (B) cell lines after incubation in U-13C-palmitate-containing HBSS supplemented with or without (w/wo) glucose (20 mM) for 20 h. UD, undetectable.

Values are expressed as mean±SD (n=3) of three independent measurements. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 by two-tailed unpaired Student’s t test.

ACOT12 and ACOT8 are involved in the catabolism of fatty acid.

(A-D) Serum levels of fasted blood glucose (A), non-fasted blood glucose (B), insulin and total FFA (D) of C57BL/6 mice with adenovirus-mediated knockdown of ACOT12 or ACOT8. Fasted, mice were fasted for 12 h; Non-fasted, mice were fed normally. (E-I) Serum free fatty acids’ levels determined by GC-MS. (J) Serum levels of triacylglycerol (TG) of C57BL/6 mice with adenovirus-mediated knockdown of ACOT12 or ACOT8. (K, L) LO2 cell lines knocked down for ACOT12 (K) or ACOT8 (L) were cultured in glucose free reaction buffer containing 0.8 μCi/mL [9,10- 3H(N)]-oleic acid for 20 h, followed by determination of the relative β- oxidation rate (n=3).

Results are expressed as mean±SD of n=10 mice per group in (A-J) and three independent experiments in (K, L), and analyzed by using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

ACOT12 and ACOT8 serve to maintain CoA pool for sustained FAO.

(A-D) Relative abundance of octanoyl-CoA (A), caproyl-CoA (B), succinyl-CoA (C) and acetoacetyl-CoA (D) in MPHs knocked down for ACOT12 or ACOT8 (n=3). (E- G) Serum levels of cholesterol (CHOL) (E), high density lipoprotein cholesterol (HDL-C) (F) and low density lipoprotein cholesterol (LDL-C) (G) of C57BL/6 mice with adenovirus-mediated knockdown of ACOT12 or ACOT8. (H) Relative abundance of reduced CoA, acetyl-CoA and other oxidized CoA (octanoyl-CoA, caproyl-CoA, succinyl-CoA, acetoacetyl-CoA and HMG-CoA) in MPHs knocked down for ACOT12 or ACOT8. HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA (n=3).

Results are expressed as mean±SD of three independent experiments in (A-D, H) and n=10 mice per group in (E-G), and analyzed by using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).

Accumulation of acetate derivatives in muscle is retarded under energy stress.

Relative abundance of 13C-acetyl-CoA (A), 13C-citrate (B), 13C-aconitate (C), 13C- isocitrate (D), 13C-succinate (E), 13C-fumarate (F) and 13C-malate (G) in the muscle of diabetic and starved mice (C57BL/6) was determined 1 h after intraperitoneal injection of 2-13C-acetate (310 mg/kg).

Values are expressed as mean±SD of (n=5 mice per group). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 and n.s. P≥ 0.05 by unpaired Student’s t test.

Behavior analyses of diabetic mice with KD of ACOT12 or ACOT8.

(A, B) Normalized forelimb strength in forelimb grip force test (A) and total running time in the rotarod test (B) were determined in STZ-induced diabetic C57BL/6 mice which were knocked down for ACOT12 or ACOT8 in liver and injected intraperitoneally w/wo acetate (300 mg/kg). (C) Total time spent in the open arms during the elevated plus maze test of the diabetic C57BL/6 mice w/wo adenovirus- mediated knockdown of ACOT12 or ACOT8 in liver. (D-F) The percentage of correct alterations (D), total distance moved (E) and total number of entries into each arm (F) in the Y-maze test in the same mice as in (C). (G, H) The novel object preference index (G) and total distance travelled (H) in NOR test were determined in the same mice as in (C).

Results in (A-H) are expressed as box-plot (box extending from the 25th to 75th percentiles with whiskers indicating the minimum and maximum, and lines in boxes indicating the median) of n=10 mice per group, and analyzed by using unpaired Student’s t test (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n.s., no significant difference).