O-GlcNAcylation of ATP-citrate lyase couples glucose supply to lipogenesis for rapid tumor cell proliferation

Abstract

Elevated lipid synthesis is one of the best-characterized metabolic alterations in cancer and crucial for membrane expansion. As a key rate-limiting enzyme in de novo fatty acid synthesis, ATP-citrate lyase (ACLY) is frequently up-regulated in tumors and regulated by posttranslational modifications (PTMs). Despite emerging evidence showing O-GlcNAcylation on ACLY, its biological function still remains unknown. Here, we observed a significant upregulation of ACLY O-GlcNAcylation in various types of human tumor cells and tissues and identified S979 as a major O-GlcNAcylation site. Importantly, S979 O-GlcNAcylation is required for substrate CoA binding and crucial for ACLY enzymatic activity. Moreover, it is sensitive to glucose fluctuation and decisive for fatty acid synthesis as well as tumor cell proliferation. In response to EGF stimulation, both S979 O-GlcNAcylation and previously characterized S455 phosphorylation played indispensable role in the regulation of ACLY activity and cell proliferation; however, they functioned independently from each other. In vivo, streptozocin treatment- and EGFR overexpression-induced growth of xenograft tumors was mitigated once S979 was mutated. Collectively, this work helps comprehend how cells interrogate the nutrient enrichment for proliferation and suggests that although mammalian cell proliferation is controlled by mitogen signaling, the ancient nutrition-sensing mechanism is conserved and still efficacious in the cells of multicellular organisms.

Keywords: ATP-citrate lyase; CoA association; O-GlcNAcylation; fatty acid synthesis; tumor cell proliferation.

Fig. 1.

ACLY is highly O-GlcNAcylated in tumor cells and patient tumor tissues. (A) Detection of ACLY O-GlcNAcylation in human cell lines. Whole lysates from indicated cells were immunoprecipitated with ACLY antibody and then subjected to western blot (WB) analysis with antibodies against ACLY and O-GlcNAcylation, respectively. O-GlcNAcylation of ACLY was normalized to total ACLY and quantified. Mean ± SD, n = 3. Student’s t test: ^***P < 0.001; ^****P < 0.0001. A.U., arbitrary units. (B) Stoichiometric analysis of O-GlcNAcylated ACLY. O-GlcNAcylated proteins from A549 cells were biotinylated and precipitated by using the Click-iT O-GlcNAc enzymatic labeling system as described in Materials and Methods. All samples were subjected to WB analysis for ACLY. To quantitatively evaluate O-GlcNAcylated ACLY, a series of lysate dilutions from input indicated as 100%, 50%, 25%, 12.5%, 6.25%, 3.125%, and 0% were prepared and loaded. According to the loading controls, stoichiometry of O-GlcNAcylated ACLY was ∼6.25 to 12.5%. (C) Evaluation of O-GlcNAcylated proteins with Alkyne-5 K-PEG resolvable mass tag. In A549 cells, both labeled and unlabeled lysates were subjected to WB analysis for ACLY. Based on band densities, O-GlcNAcylation level of ACLY was estimated. Mean ± SD, n = 3. (D–F) Analysis of ACLY O-GlcNAcylation in patient tissues. 30 paired lung tumor tissues (T) and their matching adjacent normal tissues (N) were subjected to immunoprecipitation (IP) and WB for the analysis of ACLY O-GlcNAcylation. Representative blots (D) and associated quantification (E) of ACLY O-GlcNAcylation were shown. Mean ± SD, n = 30; Student’s t test: ^****P < 0.0001. ACLY O-GlcNAcylation of tumor tissues that were categorized into different malignant stages (I–III) was also quantified (F). Mean ± SD, n = 10; One-way ANOVA: ns, not significant. A.U., arbitrary units.

Fig. 2.

S979 is the major O-GlcNAcylation site on ACLY. (A and B) Construction of an endogenous-Acly-mutated cell line expressing ACLY S979A. (A) Schematic of gene editing by PE. (B) Verification of S979 mutation on ACLY. In A549 cells carrying S979 mutation, ACLY was immunoprecipitated for the evaluation of O-GlcNAcylation by WB after PUGNAc (PUG, 100 μM) treatment or transfection of GFP-tagged OGT for 48 h. Representative blots (Top) as well as associated quantifications (Bottom) were shown. Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^***P < 0.001; ns, not significant. A.U., arbitrary units. (C) Examination of S455 phosphorylation, K540/546/554 (3K) acetylation and O-GlcNAcylation of ACLY. A549 cells transfected with Flag-tagged ACLY WT or S979A were analyzed for S455 phosphorylation (pS455), 3K acetylation (Ac-3K), and O-GlcNAcylation of ACLY. Representative blots (Left) as well as associated quantifications (Right) were shown. Mean ± SD, n = 3. Student’s t test: ns, not significant. A.U., arbitrary units. (D) Analysis of ACLY ubiquitination. A549 cells, which were transfected with Flag-tagged ACLY WT or S979A and treated with MG132 (10 μM) for 6 h, were analyzed for ACLY ubiquitination by IP and WB. Representative blots (Left) as well as associated quantifications (Right) were shown. Mean ± SD, n = 3. Student’s t test: ns, not significant. A.U., arbitrary units. (E) Examination of ACLY O-GlcNAcylation and S455 phosphorylation. A549 cells transfected with Flag-tagged ACLY WT, S455A, or S455D were analyzed for ACLY O-GlcNAcylation and S455 phosphorylation by IP and WB. Representative blots (Left) as well as associated quantifications (Right) were shown. Mean ± SD, n = 3. One-way ANOVA: ns, not significant. A.U., arbitrary units. (F) Acetylation and O-GlcNAcylation on ACLY were analyzed. A549 cells transfected with Flag-tagged ACLY WT, 3KR, or 3KQ were analyzed for ACLY O-GlcNAcylation by IP and WB. Representative blots (Left) as well as associated quantifications (Right) were shown. Mean ± SD, n = 3. One-way ANOVA: ns, not significant. A.U., arbitrary units. (G) Ubiquitination and O-GlcNAcylation on ACLY were analyzed. A549 cells, which were transfected with Flag-tagged ACLY WT or 3KR and treated with MG132 (10 μM) for 6 h, were analyzed for ACLY O-GlcNAcylation by IP and WB. Representative blots (Left) as well as associated quantifications (Right) were shown. Mean ± SD, n = 3. Student’s t test: ns, not significant. A.U., arbitrary units.

Fig. 3.

S979 O-GlcNAcylation enhances ACLY activity by promoting CoA association. (A) Analysis of ACLY activity. A549 cells transfected with Flag-tagged ACLY WT or S979A were either treated with PUGNAc (PUG, 100 μM) or cotransfected with GFP-tagged OGT for 48 h. ACLY obtained via Flag IP was used for enzymatic activity analysis, and citrate consumption and acetyl-CoA production were also examined in above cells. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (B) Evaluation of CoA and citrate affinity to ACLY. A549 cells treated as in (A) were subjected to Flag IP prior to the analysis of ACLY activity. Curves depicted the kinetics of ACLY activity at various concentrations of CoA or citrate (Left). Michaelis constant (K_m) values were demonstrated (Right). Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^***P < 0.001; ^****P < 0.0001. ns, not significant. (C and D) In vitro assay for ACLY O-GlcNAcylation and activity. His-tagged ACLY WT or S979A purified from Escherichia coli were incubated with or without GST-tagged OGT in the presence of UDP-GlcNAc. Subsequently, ACLY O-GlcNAcylation (C) and activity (D) were examined by WB and Activity Assay Kit, respectively. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (E) BLI assay for CoA binding to ACLY. His-tagged ACLY WT and S979A protein purified from E. coli were incubated with or without GST-tagged OGT in the presence of UDP-GlcNAc prior to loading to Ni-NTA biosensors. The sensorgram depicting CoA binding to ACLY (Left) and associated K_D values (Right) were demonstrated. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant. (F) ITC assay for CoA binding to ACLY. His-tagged ACLY WT or S979A protein prepared as in (C and D) were injected into calorimetric cells before CoA injection. Theoretical titration curves (Left) as well as associated K_D values (Right) were displayed. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ns, not significant.

Fig. 4.

S979 O-GlcNAcylation of ACLY promotes lipogenesis. (A–C) Functional analysis of ACLY S979 O-GlcNAcylation in lipid synthesis. A549 cells were treated as in Fig. 2B and assayed for free fatty acid (A), total triglyceride (B), and total cholesterol (C). Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^**P < 0.01; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (D–G) Evaluation of influences from ACLY S979 mutation on lipid synthesis by metabolomics. In A549 cells that were depleted for endogenous ACLY and rescued by either ACLY WT or S979A as described in SI Appendix, Fig. S3E, levels of acyl-CoA (D), phospholipids (E), sphingolipids (F), and neutral lipids (G) were examined by ultra-high-performance liquid chromatography–high-resolution mass spectrometry (UHPLC-HRMS). Quantification of different types of acyl-CoA (D). Mean ± SD, n = 4. Student’s t test: ^*P < 0.05; ^**P < 0.01; ^***P < 0.001; ^****P < 0.0001. A.U., arbitrary units. Heatmap depicting phospholipids (E), sphingolipids (F), and neutral lipids (G) in ACLY WT versus S979A cells. z-score was calculated from the average value of each group (n = 4). z-score >0 and <2 indicate increase (red), and z-score <0 and >–2 indicate decrease (blue).

Fig. 5.

S979 O-GlcNAcylation of ACLY coordinates fatty acid synthesis with glucose supply. (A) Analysis of ACLY O-GlcNAcylation in response to alterations in glucose supply. A549 cells, which were depleted for endogenous ACLY and transfected with Flag-tagged ACLY WT or S979A, were treated with various concentrations of glucose (0, 5, 10, or 25 mM) for 6 h. Flag-tagged ACLY was immunoprecipitated for the evaluation of O-GlcNAcylation by WB. Representative blots (Top) and associated quantifications (Bottom) were shown. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ^***P < 0.001; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (B) Evaluation of CoA and citrate affinity to ACLY. In A549 cells, which were depleted for endogenous ACLY and transfected with Flag-tagged ACLY WT or S979A, were treated with various concentrations of glucose (5, 10, or 25 mM) for 6 h. Flag-ACLY was immunoprecipitated for the analysis of enzymatic activity. Curves depicted the kinetics of ACLY activity at various concentrations of CoA or citrate (Left). Michaelis constant (K_m) values were demonstrated (Right). Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^***P < 0.001; ^****P < 0.0001; ns, not significant. (C and D) Assay of ACLY activity and lipid synthesis. In A549 cells treated as in (B), ACLY activity (C), free fatty acid, total triglyceride, and total cholesterol (D) were examined. Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (E) Examination of ACLY O-GlcNAcylation and activity at various concentrations of [U-¹³C] glucose. ACLY WT and S979A rescue cells that were generated as described in SI Appendix, Fig. S3E were treated with different concentrations of [U-¹³C] glucose (10 or 25 mM) for 24 h and subjected to the analysis of ACLY O-GlcNAcylation and activity, respectively. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (F) Evaluation of influence from ACLY S979 mutation on de novo fatty acid synthesis by isotope tracing. A549 cells treated as in (E) were cultured in the presence of [U-¹³C] glucose (10 or 25 mM) for 24 h prior to metabolic flux analysis for fatty acid synthesis by UHPLC-HRMS. Mean ± SD, n = 4. Two-way ANOVA: ^*P < 0.05; ^****P < 0.0001; ns, not significant. (G) Functional analysis for ACLY S979 O-GlcNAcylation in the control of tumor cell proliferation. A549 cells carrying S979 mutation described in Fig. 2 A and B were cultured at different concentrations (10 or 25 mM) of glucose for various time. Cell proliferation curves were demonstrated. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant.

Fig. 6.

ACLY S979 O-GlcNAcylation is required for EGF-induced cell responses. (A–D) Analysis of ACLY O-GlcNAcylation and activity. A549 cells, which were depleted for endogenous ACLY and transfected with Flag-tagged ACLY WT, were treated with EGF, or subjected to glucose deprivation (A and B) or using GLUT1 inhibitor BAY-876 (50 μM) (C and D) for 6 h. Flag-tagged ACLY was immunoprecipitated for the evaluation of ACLY O-GlcNAcylation (A and C) and activity (B and D). Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (E) Analysis of ACLY O-GlcNAcylation and S455 phosphorylation in response to EGF treatment. A549 cells, which were depleted for endogenous ACLY and transfected with Flag-tagged ACLY WT, S979A, or S455A, were treated with EGF for various time (0, 6, or 48 h). Flag-tagged ACLY was immunoprecipitated for the evaluation of O-GlcNAcylation or S455 phosphorylation by WB. Representative blots (Left) and associated quantifications (Right) were shown. Mean ± SD, n = 3. Two-way ANOVA: ^*P < 0.05; ^**P < 0.01; ^***P < 0.001; ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (F) Assay of ACLY activity. In A549 cells treated as in (E), ACLY activity was examined. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (G) Assay of lipid synthesis. In A549 cells, which were depleted for endogenous ACLY and rescued by either ACLY WT, S979A, or S455A and were treated with EGF for various time (0 or 48 h). free fatty acid, total triglyceride, and total cholesterol were examined. Mean ± SD, n = 3. Two-way ANOVA: ^****P < 0.0001; ns, not significant. A.U., arbitrary units. (H) Functional analysis for ACLY S979 O-GlcNAcylation in the control of tumor cell proliferation. ACLY WT, S979A, or S455A rescue cells were treated with EGF for various time. Cell proliferation curves were demonstrated. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ^****P < 0.0001.

Fig. 7.

S979 O-GlcNAcylation of ACLY is needed for EGF-induced tumor growth. (A) Influence from oleate on cell proliferation. ACLY WT and S979A rescue cells described in SI Appendix, Fig. S3E were cultured in the presence or absence of oleate (20 µM) for various time. Cell proliferation curves were demonstrated. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ns, not significant. (B and C) Assay for tumor formation. ACLY WT and S979A rescue cells described in SI Appendix, Fig. S3E were injected into athymic nude mice and xenograft tumors were sampled and photographed after 21 d. Representative images (B) and quantification of tumor volume (C) were displayed. Mean ± SD, n = 9. Student’s t test: ^****P < 0.0001. (D) Analysis of lipid synthesis in xenograft tumors. In tumor xenografts described in (B), levels of free fatty acid, total triglyceride, and total cholesterol were detected. Mean ± SD, n = 9. Student’s t test: ^****P < 0.0001. A.U., arbitrary units. (E) Immunohistochemistry (IHC) analysis for cell proliferation. Nine pairs of mice injected with ACLY WT or S979A rescue cells were costained with hematoxylin and eosin (H&E) and Ki-67. Representative images depicting tumor tissues (Scale bar, 50 μm) (Left). Quantification of Ki-67 staining (Right). Mean ± SD, n = 9. Student’s t test: ^***P < 0.001. (F) Influence from EGFR overexpression on cell proliferation. ACLY WT and S979A rescue cells described in SI Appendix, Fig. S3E were stably overexpressed with EGFR plasmid. Cell proliferation curves were demonstrated. Mean ± SD, n = 3. Two-way ANOVA: ^**P < 0.01; ns, not significant. (G and H) Assay for tumor formation. ACLY WT and S979A rescue cells described in SI Appendix, Fig. S3E were injected into athymic nude mice and xenograft tumors were sampled and photographed after 21 d. Representative images (G) and quantification of tumor volume (H) were displayed. Mean ± SD, n = 5. Two-way ANOVA: ^*P < 0.05; ^**P < 0.01; ns, not significant. (I) Analysis of lipid synthesis in xenograft tumors. In tumor xenografts described in (G), levels of free fatty acid, total triglyceride, and total cholesterol were detected. Mean ± SD, n = 5. Two-way ANOVA: ^****P < 0.0001; ns, not significant. (J) Immunohistochemistry (IHC) analysis for cell proliferation. Five pairs of mice injected with ACLY WT or S979A rescue cells were costained with hematoxylin and eosin (H&E) and Ki-67. Representative images depicting tumor tissues (Scale bar, 100 μm) (Left). Quantification of Ki-67 staining (Right). Mean ± SD, n = 5. Two-way ANOVA: ^***P < 0.001; ns, not significant.