Hyperacetylated histone H4 is a source of carbon contributing to lipid synthesis

Abstract

Histone modifications commonly integrate environmental cues with cellular metabolic outputs by affecting gene expression. However, chromatin modifications such as acetylation do not always correlate with transcription, pointing towards an alternative role of histone modifications in cellular metabolism. Using an approach that integrates mass spectrometry-based histone modification mapping and metabolomics with stable isotope tracers, we demonstrate that elevated lipids in acetyltransferase-depleted hepatocytes result from carbon atoms derived from deacetylation of hyperacetylated histone H4 flowing towards fatty acids. Consistently, enhanced lipid synthesis in acetyltransferase-depleted hepatocytes is dependent on histone deacetylases and acetyl-CoA synthetase ACSS2, but not on the substrate specificity of the acetyltransferases. Furthermore, we show that during diet-induced lipid synthesis the levels of hyperacetylated histone H4 decrease in hepatocytes and in mouse liver. In addition, overexpression of acetyltransferases can reverse diet-induced lipogenesis by blocking lipid droplet accumulation and maintaining the levels of hyperacetylated histone H4. Overall, these findings highlight hyperacetylated histones as a metabolite reservoir that can directly contribute carbon to lipid synthesis, constituting a novel function of chromatin in cellular metabolism.

Keywords: Acetylation; Epigenetics; Histone Reservoirs; Lipid Metabolism.

Figure 1. Acetyltransferase depletion increases lipids and fatty acids in AML12 hepatocytes.

(A) PCA scores plot (PC1, 71.7% against PC2, 17%) showing separation in intact lipid species acids between the scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD sample groups. PC1 is positively correlated with intact lipid content. n = 3 biological replicates/group. (B) PCA scores plot (PC1, 54.4% against PC2, 38.4%) showing separation in lipid-bound fatty acids between the scramble, MYST1-KD, GCN5-KD, NAA40-KD and NAA10-KD sample groups. PC1 is positively correlated with lipid-bound fatty acid content. n = 3 biological replicates/group. (C) Sum of total lipids, triglycerides, and lipid-bound fatty acids measured by MS in scramble, MYST1-KD, GCN5-KD, NAA40-KD and NAA10-KD cells after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; **P  ≤  0.01, ****P ≤ 0.0001. (D) Representative image of lipid droplets by Nile red (red), Perilipin-1 (green), and nuclei by DAPI (blue) in sample groups 48 h after siRNA treatment; Scale bar = 25 μm. n = 8–10 biological replicates/group. (E) Quantification of relative lipid droplets in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells. n = 8–10 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ****P ≤ 0.0001, ns non-significant. (F) Quantification of Perilipin-1 using ImageJ in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells. n = 8–10 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ***P ≤  0.001. (G) RT-qPCR analysis of expression of NAA40 (left) and Gcn5 (right) mRNA levels, after 48 h in represented treatment groups. n = 8–10 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ****P ≤ 0.0001. (H) Triglyceride levels after 48 h in represented treatment groups. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ****P ≤ 0.0001, ns non-significant. Data information: (C, G, H) are presented as mean ± SEM; (E, F) are presented as violin plots. Source data are available online for this figure.

Figure 2. Lipid synthesis upon acetyltransferase depletion is not associated with glucose-derived acetyl-CoA in AML12 hepatocytes.

(A) Schematic representation of experimental design (left) and the synthesis of cytosolic acetyl-CoA from exogenous glucose (right). (B) Heatmap of normalised levels (%label in KD/%label in Scramble) of ¹³C-labelled species in mono-, di- and tri-acetylated histone peptides in scramble, MYST1-KD, GCN5-KD, NAA40-KD and NAA10-KD cells. n = 3 biological replicates/group. (C) Normalised levels (%label in KD/%label in Scramble) of the tri-acetylated H4K5K8K12K16 peptide carrying one, two or three labelled acetyl-CoA moieties in scramble, MYST1-KD, GCN5-KD, NAA40-KD and NAA10-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, ns non-significant. (D) Normalised levels (%label in KD/%label in Scramble) of indicated methyl fatty acids in scramble, MYST1-KD, GCN5-KD, NAA40-KD and NAA10-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ns non-significant. Data information: (C, D) are presented as mean ± SEM. Source data are available online for this figure.

Figure 3. Hyperacetylated H4 is a carbon source for lipid synthesis upon acetyltransferase depletion.

(A) Heatmap of relative abundances (normalised to scramble) of histone acetylation marks measured by MS in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells. n = 3 biological replicates/group. (B) Schematic representation of experimental design (left) and the generation of acetyl-CoA from histone deacetylation (right). (C) Normalised levels (%label 6 h/%label 0 h) of mono-acetylated H3AK9me3K14 in scramble, GCN5-KD and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (D) Normalised levels (%label 6 h/%label 0 h) of H4K5K8K12K16-1ac carrying one labelled acetyl-CoA moiety in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (E) Normalised levels (%label 6 h/%label 0 h) of H4K5K8K12K16-2ac carrying two labelled acetyl-CoA moieties in scramble, GCN5-KD and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (F) Normalised levels (%label 6 h/%label 0 h) of H4K5K8K12K16-3ac carrying three labelled acetyl-CoA moieties in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01. *for comparisons between scramble and GCN5-KD and # for scramble and NAAA40-KD. (G) Normalised levels (%label 6 h/%label 0 h) of methyl palmitate in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; **P  ≤  0.01, ****P ≤ 0.0001. *For comparisons between scramble and GCN5-KD and # for scramble and NAAA40-KD. (H) Normalised levels (%label 6 h/%label 0 h) of methyl stearate in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; **P  ≤  0.01, ****P ≤ 0.0001. *For comparisons between scramble and GCN5-KD and # for scramble and NAAA40-KD. (I) Normalised levels (%label 6 h/%label 0 h) of methyl eicosanoate in scramble, GCN5-KD and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; **P  ≤  0.01. *For comparisons between scramble and GCN5-KD and # for scramble and NAAA40-KD. (J) Normalised levels (%label 6 h/%label 0 h) of methyl lignoceric acid in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ****P ≤ 0.0001. *For comparisons between scramble and GCN5-KD and # for scramble and NAAA40-KD. (K) Mass isotopomer distribution of methyl palmitate at 6 h post media change in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01. (L) Mass isotopomer distribution of methyl stearate at 6 h post media change in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01. (M) Mass isotopomer distribution of methyl eicosanoate at 6 h post media change in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ***P ≤  0.001. (N) Mass isotopomer distribution of methyl lignoceric acid at 6 h post media change in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05. Data information: (C–N) are presented as mean ± SEM. Source data are available online for this figure.

Figure 4. ACSS2 drives HDAC-dependent lipid droplet formation upon acetyltransferase depletion in AML12 hepatocytes.

(A) RT-qPCR analysis of Acss2 mRNA levels in scramble, MYST1-KD, GCN5-KD, NAA40-KD cells, and NAA10-KD 48 h after siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ***P ≤  0.001, ****P ≤ 0.0001, ns non-significant. (B) Representative immunoblot (top) and quantification (bottom) of ACSS2 in scramble, MYST1-KD, GCN5-KD, NAA40-KD cells, and NAA10-KD 48 h after siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ***P ≤  0.001, ****P ≤ 0.0001. (C) Quantification of relative lipid droplets in scramble and the indicated double-KD cells after 48 h of siRNA treatment. n = 6–8 biological replicates/group. Statistical analysis was performed using by one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ns non-significant. (D) Triglyceride levels in scramble and the indicated double-KD cells after 48 h of siRNA treatment. N = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ns non-significant. (E) Representative images of lipid droplets by Nile red (red) and nuclei by DAPI (blue) in scramble and the indicated double-KD cells after 48 h of siRNA treatment. Scale bar = 25 μm. n = 6–8 biological replicates/group. (F) Schematic representation of experimental design (left) and the generation of acetyl-CoA from histone deacetylation, through the activity of ACSS2 (right). (G) Average of normalised levels (%label 6 h/%label 0 h) of the detected fully labelled acetylated histone peptides in scramble, ACSS2-KD, and double-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01. (H) Normalised levels (%label 6 h/%label 0 h) of indicated histone peptides in scramble, ACSS2-KD, and double-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, ***P ≤  0.001, ****P ≤ 0.0001, ns non-significant. (I) Normalised levels (%label 6 h/%label 0 h) of indicated methyl fatty acids in scramble, ACSS2-KD, and double-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. Data information: (A, B, G–I) are presented as mean ± SEM; (C, D) are presented as violin plots. Source data are available online for this figure.

Figure 5. Overexpression of acetyltransferases reverses HFHG-induced fat accumulation in vitro.

(A) Volcano plot comparing histone acetylation marks, measured by MS, in cells with high fructose/high glucose (HFHG) media vs. control Media carrying EVs. n = 3 biological replicates/group. Statistical analysis was performed using a two-tailed T test. (B) Representative images of lipid droplets by Nile red (red) and nuclei by DAPI (blue) in control media, and high fructose/high glucose (HFHG) media, 48 h after co-overexpression of GCN5 and NAA40 (dO/E) or corresponding empty vectors (EV) in AML12 hepatocytes; Scale bar = 25 μm. n = 6 biological replicates/group. (C) Quantification of relative lipid droplets in control media, and high fructose/high glucose (HFHG) media, 48 h after co-overexpression of GCN5 and NAA40 (dO/E) or corresponding empty vectors (EV) in AML12 hepatocytes. n = 6 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ****P ≤ 0.0001, ns non-significant. (D) Triglyceride levels in control media, and high fructose/high glucose (HFHG) media, 48 h after co-overexpression of GCN5 and NAA40 or corresponding empty vectors in AML12 hepatocytes. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ns non-significant. (E) Representative immunoblots of GCN5, NAA40, ACSS2 and β-actin in control media, and high fructose/high glucose (HFHG) media, 48 h after co-overexpression of GCN5 and NAA40 (dO/E) or corresponding empty vectors (EV) in AML12 hepatocytes. n = 3 biological replicates/group. (F) Quantification of ACCS2 in control media and high fructose/high glucose (HFHG), 48 h after co-overexpression of GCN5 and NAA40 (dO/E) or corresponding empty vectors (EV) in AML12 hepatocytes. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, ns non-significant. (G) Volcano plot comparing histone acetylation marks, measured by MS, in cells with HFHG media Vs Control media carrying both GCN5 and NAA40 plasmids (dO/E) in AML12 hepatocytes. n = 3 biological replicates/group. Statistical analysis was performed using a two-tailed T test. Data information: Log2(FC) threshold for (A, G) based on variance=0.018. (C) is presented as a violin plot; (D, F) are presented as mean ± SEM. Source data are available online for this figure.

Figure 6. Overexpression of acetyltransferases reverses HFHG-induced fat accumulation in vivo.

(A) Schematic representation of in vivo experimental design. (B) Body weight (grams) of mice supplied with tap water or HFHG water 48 h after IP injections with both GCNA and NAA40 (dO/E) or their corresponding empty vectors (EV). n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ns non-significant. (C) Volcano plot comparing histone acetylation marks, measured by MS, in livers from mice supplied with high fructose/high glucose (HFHG) water vs. tap water carrying EVs. n = 3 biological replicates/group. Statistical analysis was performed by a two-tailed T test. (D) Representative images of lipid droplets by Nile red (red) and nuclei by DAPI (blue) of mouse liver sections from each treatment group. Scale bar = 25 μm. n = 3 biological replicates/group (9 technical replicates). (E) Quantification of relative lipid droplets of mouse liver sections from each treatment group. n = 3 biological replicates/group (9 technical replicates). Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ns non-significant. (F) Triglyceride quantification of mouse liver from each treatment group. n = 3 biological replicates/group. Statistical analysis was performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; **P  ≤  0.01, ns non-significant. (G) Representative immunoblots (left) and quantification (right) of GCN5, NAA40, ACSS2 and β-actin in tap water and high fructose/high glucose (HFHG) conditions, 48 h after co-overexpression of GCN5 and NAA40 (dO/E) or corresponding empty vectors (EV). n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, ****P ≤ 0.0001, ns non-significant. (H) Volcano plot comparing histone acetylation marks in livers of mice supplied with HFHG water vs. tap water carrying both GCN5 and NAA40 plasmids (dO/E). n = 3 biological replicates/group. Statistical analysis was performed using a two-tailed T test. Data information: Log2(FC) threshold for (C, H) based on variance = 0.02. (B, F–G) Are presented as mean ± SEM; (E) is presented as a violin plot. Source data are available online for this figure.

Figure 7. Model of hyperacetylated H4 reservoir contributing carbon to lipids.

Based on the overall activity between acetyltransferases and HDACs, carbons derived from deacetylation of hyperacetylated histone H4 can contribute to lipid synthesis in an ACSS2-dependent manner.

Figure EV1. Acetyltransferase depletion attenuates insulin signalling and does not affect the transcription of genes involved in lipid synthesis.

(A) RT-qPCR analysis of expression of Myst-1, Gcn5, Naa40, and Naa10 mRNA levels in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells, respectively, after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ****P ≤ 0.0001. (B) Relative intensity (normalised to scramble) of indicated metabolites in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells measured by MS after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test. Only P values which are *P  ≤  0.05 are indicated. All comparisons between KD and Scramble that are non-significant are not indicated. (C) Relative intensity (normalised to scramble) of free fatty acids and cholesterol in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells measured by MS. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ns non-significant. (D) Representative immunoblots of phopshoAKT ser473, totalAKT and β-actin in scramble, MYST1-KD, GCN5-KD, NAA40-KD cells, and NAA10-KD respectively, after 48 h of siRNA treatment. n = 3 biological replicates/group. (E) Mean fluorescent uptake quantification by FACS in scramble, MYST1-KD, GCN5-KD, NAA40-KD cells, and NAA10-KD, respectively, after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; **P  ≤  0.01, ns non-significant. (F) RT-qPCR analysis of expression of Acly, Fasn, Sbrebf, mRNA levels in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells, respectively, after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using a two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (G) Volcano plot comparing mRNA levels between NAA40-KD and SCR control cells as determined by RNA-seq analysis after 12 h of siRNA treatment. Upregulated genes upon loss of NAA40 are shown in red (adjusted P  <  0.05 and logFC > 0.5) and downregulated genes in blue (adjusted P  <  0.05 and logFC < −0.5). n = 3 biological replicates/group. Statistical analysis was performed by a paired T test and corrected with the False Discovery Rate (FDR). (H) Volcano plot comparing mRNA levels between NAA40-KD and SCR control cells as determined by RNA-seq analysis after 48 h of siRNA treatment. Upregulated genes upon loss of NAA40 are shown in red (adjusted P  <  0.05 and logFC > 0.5) and downregulated genes in blue (adjusted P  <  0.05 and logFC < −0.5). n = 3 biological replicates/group. Statistical analysis was performed by a paired T test and corrected with the False Discovery Rate (FDR). (I) Gene ontology analysis of all differentially expressed genes showing enriched biological processes following depletion of NAA40 after 48 h of siRNA treatment. n = 3 biological replicates/group. Data information: all data are presented as mean ± SEM.

Figure EV2. Lipid synthesis upon acetyltransferase depletion is not associated with ACLY in AML12 hepatocytes.

(A) Relative abundance of ¹²C and ¹³C in histone acetylation peptides measured by MS in scramble cells 48 h after siRNA treatment and supplementation with ¹³C₆-Glucose. n = 3 biological replicates/group. (B) Relative abundance of ¹²C and ¹³C methyl fatty acids measured by MS in scramble cells 48 h after siRNA treatment and supplementation with ¹³C₆-Glucose. n = 3 biological replicates/group. (C) Normalised levels (% label/ scramble) incorporation in indicated metabolites in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells measured by MS, 48 h after siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test. All comparisons between KD and Scramble that are non-significant are not indicated. (D) RT-qPCR analysis of Myst-1, Gcn5, Naa40, Naa10 and Acly mRNA levels in scramble, ACLY-KD and the indicated double-KD cells after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ***P ≤  0.001, ****P ≤ 0.0001. (E) Quantification of relative lipid droplets in scramble, ACLY-KD and the indicated double-KD cells after 48 h of siRNA treatment. n = 6–8 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ****P ≤ 0.0001. (F) Representative images of lipid droplets Nile red (red) and nuclei by DAPI (blue) of scramble, ACLY-KD and the indicated double-KD cells after 48 h of siRNA treatment. Scale bar = 25 μm. n = 6–8 biological replicates/group. Data information: (A–D) are presented as mean ± SEM; (E) is presented as a violin plot.

Figure EV3. Analysis of histone acetylation marks, methyl fatty acids, free fatty acids and aqueous metabolites upon acetyltransferase depletion in chromatin-enriched ¹³C-labelled cells.

(A) Normalised levels (%RA/scramble) of the triply acetylated H4K5K8K12K16 peptide in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells measured by MS, 48 h after siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; **P  ≤  0.01, ns non-significant. (B) Relative abundance of ¹²C and ¹³C in acetylated histone peptides in AML12 cells 7 days after supplementation with ¹³C₆-Glucose. n = 3 biological replicates/group. (C) Representative immunoblots of GCN5, NAA40 and β-actin in scramble, GCN5-KD, and NAA40-KD cells at 6, 12, 24 and 48 h after siRNA treatment. n = 3 biological replicates/group. (D) Normalised levels (%label 6 h/%label 0 h) of singly labelled H3K9ac|K14ac in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (E) Normalised levels (%label 6 h/%label 0 h) of singly labelled H3K9me1K14ac in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (F) Normalised levels (%label 6 h/%label 0 h) of singly labelled H3K9me2K14ac in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (G) Normalised levels (%label 6 h/%label 0 h) of singly labelled H3K18ac | K23 in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (H) Normalised levels (%label 6 h/%label 0 h) of singly labelled H2AK5ac|K9ac in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (I) Normalised levels (%label 6 h/%label 0 h) of doubly labelled H3K9acK14ac in scramble, GCN5-KD, and NAA40-KD cells in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (J) Normalised levels (%label 6 h/%label 0 h) of methyl eicosenoate (20:1) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; **P  ≤  0.01. (K) Normalised levels (%label 6 h/%label 0 h) of Methyl eicosatetranoate (20:4) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ***P ≤  0.001. (L) Normalised levels (%label 6 h/%label 0 h) of free palmitate (16:0) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (M) Normalised levels (%label 6 h/%label 0 h) of free oleate (18:1) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (N) Normalised levels (%label 6 h/%label 0 h) of free eicosanoate (20:0) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (O) Normalised levels (%label 6 h/%label 0 h) of free eicosenoate (20:1) in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; ns non-significant. (P) Normalised levels (%label 6 h/%label 0 h) of the indicated aqueous metabolites in scramble, GCN5-KD, and NAA40-KD cells. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test. All comparisons between KD and Scramble that are non-significant are not indicated. Data information: all data are presented as mean ± SEM. Source data are available online for this figure.

Figure EV4. Lipid synthesis upon acetyltransferase depletion is associated with ACSS2-driven HDAC-dependent acetate formation.

(A) Quantification of H4K8Ac using ImageJ MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells with and without sodium butyrate, 48 h after siRNA treatment. n = 6 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05, **P  ≤  0.01, ns non-significant. (B) Quantification of relative lipid droplets in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells with and without sodium butyrate, 48 h after siRNA treatment. n = 6 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test; *P  ≤  0.05. (C) Representative images of lipid droplets by Nile red staining (red), H4K8Ac (green) and nuclei by DAPI (blue) in scramble, MYST1-KD, GCN5-KD, NAA40-KD, and NAA10-KD cells with and without sodium butyrate, 48 h after siRNA treatment; Scale bar = 25 μm. n = 6 biological replicates/group. (D) RT-qPCR analysis of Acss2 mRNA levels in the indicated treatment groups. n = 3 biological replicates/group. Statistical analysis was performed using a one-way ANOVA with post hoc Dunnett’s multiple-comparisons test; ***P ≤  0.001, ****P ≤ 0.0001, ns non-significant. (E) RT-qPCR analysis of Myst-1, Gcn5, Naa40, Naa10 and Acss2 mRNA levels in scramble, ACSS2-KD and the indicated double-KD cells after 48 h of siRNA treatment. n = 3 biological replicates/group. Statistical analysis was performed using two-way ANOVA with post hoc Tukey’s multiple-comparisons test. ****P ≤ 0.0001. Data information: all data are presented as mean ± SEM.

Figure EV5. MS-MS spectra for the differentially acetylated forms of the histone H4 tail.

(A) MS-MS spectra for differentially acetylated H4K5K8K12K16-1ac histone peptide. (B) MS-MS spectra for differentially acetylated H4K5K8K12K16-2ac histone peptide. (C) MS-MS spectra for differentially acetylated H4K5K8K12K16-3ac histone peptide.