TMEM120A maintains adipose tissue lipid homeostasis through ER CoA channeling

Abstract

Efficient fatty acid (FA) re-esterification is essential for lipid homeostasis in adipocytes, yet the mechanisms coordinating Coenzyme A (CoA) availability at the endoplasmic reticulum (ER)-a major site of lipid synthesis-remain unclear. Here, we identify TMEM120A as an ER-resident CoA-binding protein that regulates intracellular FA metabolism. TMEM120A interacts with the ER-localized acyl-CoA synthetase ACSL1 and ACSL3 to promote long-chain acyl-CoA synthesis and channeling into the ER, thereby facilitating FA re-esterification and lipid cycling during lipolysis. By relieving acyl-CoA-mediated feedback inhibition of lipolysis, TMEM120A enhances lipid turnover while protecting against ER stress and lipotoxicity. Adipocyte-specific deletion of Tmem120a in mice impairs lipolysis-induced energy expenditure and exacerbates inflammation and metabolic dysfunction under high-fat diet conditions. These findings establish TMEM120A as a critical regulator of ER CoA handling and lipid flux, revealing a previously unrecognized mechanism that links intracellular CoA dynamics to systemic energy balance and metabolic health.

Fig. 1. Tmem120a is predominantly expressed in adipocytes and downregulated in obesity.

A mRNA levels of Tmem120a in various tissues. n = 3. B, C Protein expression of TMEM120A in various tissues. n = 4. D, E UMAP plot of total 18,542 nuclei isolated from gWAT of WT and Tmem120a AKO mice fed an NCD. Red color indicates higher expression levels, whereas grey color indicates lower expression levels. (Two replicates for the WT condition and one replicate for the Tmem120a AKO condition). Clusters are colored by cell types: Adipocyte, mesothelial cell (MC), lymphatic endothelial cell (LEC), vascular endothelial cell (VEC), adipocyte progenitor cell (APC), smooth muscle cell (SMC), macrophages, B cell and T cell. F mRNA levels of Tmem120a in floating adipocytes and F4/80⁺ macrophages and PDGFRα⁺ progenitor cells from SVF isolated from iWAT and gWAT. n = 3. G Single-nucleus RNA analysis of adipocytes, APC, Trem2⁺ macrophages from gWAT of mice fed an NCD or HFD for 12 weeks (SRP426501). H Analysis of Tmem120a expression in gWAT of mice fed an NCD or HFD for 8 weeks, determined by publicly available transcriptomic data (GSE182930). n = 3. I mRNA levels of Tmem120a in BAT, iWAT, and gWAT of WT mice fed an NCD or HFD for 8 weeks. n = 10. J Expression of TMEM120A across different cell types in human white adipose tissue, based on the public dataset GSE176171. Adipose-derived stromal/stem progenitor cell (ASPC). mRNA levels (K) and correlation analysis (L) of TMEM120A gene expression in human subcutaneous adipose tissue of normal (BMI < 22 kg/m²) and obese (BMI > 26 kg/m²) patients. n = 12. M Pearson correlation between BMI and average TMEM120A expression in human adipocytes (GSE176171). Each point represents an individual sample. The black dashed line indicates the linear regression fit, and the shaded area represents the 95% confidence interval. N Analysis of TMEM120A expression in human adipose tissue ((MHL): n  =  14; (MHO): n  =  25; (MUO): n  =  27), determined by publicly available transcriptomic data (GSE156906). Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by an unpaired two-sided Student’s t-test (A, C, F–I, K, N), or two-tailed Pearson correlation (L, M).

Fig. 2. TMEM120A is an ER localized CoA-binding protein lacking fatty acid elongase activity.

A Schematic diagram representing the generation of doxycycline (Dox)-inducible TMEM120A overexpression model. Confirmation of TMEM120A overexpression model in C3H10T1/2 adipocytes by qPCR (B) and immunoblot analysis (C, D). n = 3. E Representative images of fixed C3H10T1/2 adipocytes overexpressing GFP-tagged TMEM120A, stained with either anti-Calnexin or anti-Lamin A/C. Nuclei were counterstained with DAPI. n = 8. Scale bar, 5 μm. F Colocalization analysis of TMEM120A with either anti-Calnexin and anti-Lamin A/C. n = 8. G Immunoblot analysis of TMEM120A in cell lysate (C), nuclear (N), mitochondrial (Mito), and microsomal (Mic) fractions from C3H10T1/2 adipocytes. Cells were harvested in isolation buffer (IB; 225 mM mannitol, 75 mM sucrose, 0.1 mM EGTA, 30 mM Tris-HCl, pH 7.4) supplemented with protease and phosphatase inhibitors. Adipocytes were homogenized using a syringe. Nuclei were pelleted at 600 x g and lysed in RIPA buffer. The supernatant (containing mitochondria and microsomes) was collected, and crude mitochondria were isolated at 10,000 x g and resuspended in mitochondria resuspending buffer (MRB; 250 mM mannitol, 5 mM HEPES, 0.5 mM EGTA, pH 7.4) supplemented with protease and phosphatase inhibitors. The remaining supernatant (cytosolic fraction containing microsomes) was collected, and microsomal fractions were obtained by ultracentrifugation at 100,000 x g and resuspended in MRB. 3 µg of protein was loaded per lane. H Schematic diagram illustrating in vitro fatty acid elongation assay. Microsomal fraction was incubated with 100 μM Tris-HCl (pH 7.4), 200 μM palmitoyl-CoA, 500 μM NADPH, and 60 μM ¹³C-malonyl-CoA for 1 h. I Measurement of in vitro fatty acid elongation activity. n = 5. J. pLDDT confidence mapping of AlphaFold3–modeled TMEM120A homodimers in complex with CoA. Structures are colored by pLDDT score: 0–50 (orange), 50–70 (yellow), 70–90 (light blue), 90–100 (dark blue). TMEM120A–CoA, the core CoA-binding pocket is predominantly light/dark blue, reflecting high confidence and agreement with cryo-EM data (PDB 7F3T, 7N0L). Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by an unpaired two-sided Student’s t-test (B, D, F, I).

Fig. 3. TMEM120A physically interacts with ACSL1 and ACSL3 at the ER and enhances fatty acid re-esterification.

A IP analysis with control and overexpressed TMEM120A^WT or TMEM120A^W193A and endogenous ACSL1 or ACSL3 of C3H10T1/2 adipocytes. B–E Immunoblot analysis of ACSL1, ACSL3 and TMEM120A in mitochondrial and microsomal fraction of C3H10T1/2 adipocytes treated with either vehicle control (VC) or isoproterenol (ISO) for 2 h. n = 3. F–J Immunoblot analysis of ACSL1, ACSL3 and TMEM120A in mitochondrial and microsomal fraction of control or Tmem120a knockdown C3H10T1/2 adipocytes. n = 3. K Bimolecular fluorescence complementation (BiFC) analysis of TMEM120A and ACSL3 in HEK293T cells. BiFC signals are shown in green, while ER-Tracker Red is shown in red. BiFC signals correspond to ER stained with ER-Tracker Red. n = 5. Scale bar, 10 μm. L Assessment of fatty acid re-esterification rates by measuring ¹³C-palmitate incorporation into triacylglycerol (TG) species in C3H10T1/2 adipocytes overexpressing TMEM120A or control cells. Cells were treated with either VC or ISO (10 μM) and ¹³C-palmitate (100 μM) for 2 h. Data are presented relative to the isotope negative control (INC, %), in which cells were incubated with unlabeled palmitate (100 μM). n = 6. M Quantification of CoA levels in ER-enriched microsomal fractions from C3H10T1/2 adipocytes overexpressing TMEM120A or control cells, using an enzymatic kit and targeted LC-MS/MS. n = 3. N, O Immunoblot analysis of ER stress markers in controls or Tmem120a knockdown C3H10T1/2 adipocytes treated with A922500 (DGAT1 inhibitor, 5 μM) for 1 h, followed by treating with ISO (10 μM) for 4 h. n = 3. P Measurement of ROS levels using H₂DCFDA (10 uM) in controls or Tmem120a knockdown C3H10T1/2 adipocytes treated with either vehicle or A922500 (DGAT1 inhibitor, 5 μM) for 1 h, followed by treating with ISO (10 μM) for 4 h. n = 3. Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by a two-way ANOVA followed Sidak’s multiple comparisons test (C–E, G–J, L, O, P) and an unpaired two-sided Student’s t-test (M).

Fig. 4. TMEM120A promote lipolysis by reducing cytosolic LC-acyl-CoA levels in adipocytes.

A Schematic of the single-molecule FRET sensor containing Cerulean (CFP) mutant in fusion with ABHD5, a flexible region containing GGS repeats (black line), Citrine (YFP), and PLIN5^384−417. B C3H10T1/2 adipocytes stably expressing the ABHD5^R299N LC-acyl-CoA sensor and TMEM120A (WT or W193A mutant) or ABHD5^R299N LC-acyl-CoA sensor alone were imaged using laser scanning confocal microscopy at 30-second interval. Isoproterenol (ISO) was treated at the indicated time, and the nFRET signals were calculated. n = 4. Data are from four independent experiments imaged for each condition and are representative of 4 individual experiments. C Confocal imaging of C3H10T1/2 adipocytes stably expressing the ABHD5^R299N LC-acyl-CoA sensor and TMEM120A (WT or W193A mutant) or ABHD5^R299N LC-acyl-CoA sensor alone at the indicated time. Scale bar, 10 μm. n = 4. D, E Immunoblot analysis of TMEM120A overexpressing C3H10T1/2 adipocytes or controls treated with either vehicle control (VC) or ISO (10 μM) for 4 h. n = 3. F Intracellular cAMP levels in TMEM120A overexpressing C3H10T1/2 adipocytes or controls treated with either VC or ISO (10 μM) for 4 h. n = 3. Glycerol (G) and FFA (H) in the treatment media following 4 h treatment with ISO in TMEM120A (WT or W193A mutant) overexpressing C3H10T1/2 adipocytes or controls. n = 3. Glycerol (I) and FFA (J) in the treatment media following 4 h treatment of either VC or SR-3420 (20 μM) in TMEM120A (WT or W193A mutant) overexpressing C3H10T1/2 adipocytes or controls. n = 3. Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by a two-way ANOVA followed Sidak’s multiple comparisons test (E–J).

Fig. 5. Adipocyte-specific deletion of TMEM120A impairs CL-induced lipolysis and reduces energy expenditure in vivo.

A Schematic diagram showing the establishment of adipocyte-specific TMEM120A KO mice (TMEM120A AKO). Male mice at 8 weeks of age were used for the experiments. B Experimental design of WT and TMEM120A AKO mice at 8 weeks of age treated intraperitoneally with CL316,243 (1 mg/kg) for 4 h. C mRNA levels of Tmem120a in adipose tissues of WT and TMEM120A AKO mice treated with either vehicle control (VC) or CL316,243 (CL). n = 6. FFA (D) and glycerol (E) in serum of WT and TMEM120A AKO mice treated with either VC or CL for 4 h. n = 6. F–H Immunoblot analysis of iWAT and gWAT of WT and TMEM120A AKO mice treated with either VC or CL for 4 h. n = 6. Significant effects of genotype in protein levels of iWAT (TMEM120A/Tubulin: VC:WT vs. VC:AKO, p < 0.0001; CL:WT vs. CL:AKO, p < 0.0001; p-HSL/HSL: CL:WT vs. CL:AKO, p < 0.0001; p-PLIN/PLIN: CL:WT vs. CL:AKO, p = 0.0094) Significant effects of genotype in protein levels of gWAT (TMEM120A/Tubulin: VC:WT vs. VC:AKO, p < 0.0001; CL:WT vs. CL:AKO, p < 0.0001; p-HSL/HSL: CL:WT vs. CL:AKO, p = 0.0047; p-PLIN/PLIN: CL:WT vs. CL:AKO, p = 0.0012) I-L. Indirect calorimetry analysis. Arrows indicate the injection of CL. VO₂: rate of oxygen consumption; VCO₂: rate of carbon dioxide production; RER: respiratory exchange ratio. n = 6. Significant effects of genotype in VO₂ (Total VO₂, p = 0.0189; Dark VO₂, p = 0.0047). Significant effects of genotype in VCO₂ (Total VCO₂, p = 0.0004; Light VCO₂, p = 0.0018; Dark VCO₂, p = 0.0027). Significant effects of genotype in EE (Total EE, p = 0.0081; Light EE, p = 0.0387; Dark EE, p = 0.0023). Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by an unpaired two-sided Student’s t-test (I–L) or a two-way ANOVA followed Sidak’s multiple comparisons test (C–E, H).

Fig. 6. Single nucleus RNAseq reveals upregulation of genes involved in oxidative stress and inflammation by TMEM120A KO in adipocytes.

A mRNA levels of Tmem120a in adipose tissues of WT and TMEM120A AKO mice. n = 6. B Volcano plot of differentially expressed genes (DEGs) in adipocytes between TMEM120A AKO and WT mice. Male mice at 20 weeks of age were used for this experiment. Dashed lines indicate an adjusted p value = 0.05 and |log₂ fold change | = 0.25. Genes (dots) are colored by associated pathways. C, D Selected upregulated and downregulated pathways in adipocytes of TMEM120A AKO mice compared to WT. E Violin plots showing peroxisome biogenesis and β-oxidation scores in adipocytes across conditions (WT, n = 7,264; AKO, n = 2,127 nuclei). F Enrichment plots for three representative pathways upregulated in macrophages of TMEM120A AKO mice compared to WT. Each plot shows the enrichment score (green line), the positions of genes in the ranked list (black ticks), and the maximum enrichment score (horizontal red dashed line). Normalized enrichment score (NES) and adjusted p value (p.adj) are indicated in each panel. G Violin plot showing pathway activity scores in macrophages across conditions (WT, n = 868; AKO, n = 69 nuclei). H Dot plot showing pathways enriched in TMEM120A AKO compared to WT mice for each cell type. The size of each dot represents the adjusted p value (p.adj), and the color indicates the NES. Only pathways with adjusted p value < 0.05 are shown. I mRNA levels of genes involved in oxidative phosphorylation, reactive oxygen species, lipid metabolism, and insulin sensitivity. n = 4. J Tissue weights of WT and TMEM120A AKO mice. n = 6. K Body composition analysis of WT and TMEM120A AKO mice. WT: n = 5; AKO: n = 4. L Intraperitoneal-glucose tolerance test (IP- GTT) of WT and TMEM120A AKO mice. n = 6. Data are presented as mean values  ±  SEM. Each point represents a biological replicate. p values were determined by an unpaired two-sided Student’s t-test (A, E, G, I–L).

Fig. 7. Adipocyte-specific deletion of TMEM120A exacerbates metabolic dysfunction in a diet–induced obesity mouse model.

A Experimental design of WT and TMEM120A AKO mice, initiated at 8 weeks of age and subjected to a HFD for 12 weeks. Indirect calorimetry was conducted during week 12, and glucose tolerance test (GTT) was performed after 12 weeks of HFD feeding. B Body weight monitoring of WT and TMEM120A AKO mice. n  =  6. C Tissue weights of adipose tissue depots and the liver of WT and TMEM120A AKO mice. n = 6. D Body composition of WT and TMEM120A AKO mice. n = 6. E Representative H&E staining of iWAT, gWAT, and liver. Scale bar = 50 μm. F Area of adipocytes in iWAT and gWAT from WT and TMEM120A AKO mice. The number of adipocytes measured for WT and TMEM120A AKO are 500. Hepatic TG (G) and serum FFA (H) levels of WT and TMEM120A AKO mice. n = 6. I, J Regression plots of energy expenditure against body mass (ANCOVA using body mass as a covariate, two-sided without adjustment, n = 6) and ANCOVA-adjusted energy expenditure (predicted energy expenditure at the mean body mass 47.63 g). Fasting glucose levels (K) and intraperitoneal- (IP-) GTT (L) of WT and TMEM120A AKO mice. n = 6. M, N Immunoblot of ER stress and inflammation markers in gWAT of WT and TMEM120A AKO mice fed a HFD for 12 weeks. n = 6. O. Immunofluorescence staining of F4/80 in gWAT of WT and TMEM20A AKO mice and quantification of F4/80 fluorescence intensity. n  =  14. Scale bar = 50 μm. Each point represents a technical replicate. P, Q Immunoblot analysis of apoptosis marker in gWAT of WT and TMEM120A AKO mice. n = 6. R Flow cytometric analysis of CD45⁺CD11b⁺CD11c⁺ (M1-like) and CD45⁺CD11b⁺CD206⁺ (M2-like) macrophages in SVFs isolated from gWAT of WT and TMEM120A AKO mice. n = 3. Data are presented as mean values  ±  SEM. Each point represents a biological replicate except (O). p values were determined by an unpaired two-sided Student’s t-test (B–D, F–H, K, L, N, O, and Q) and a two-way ANOVA followed Sidak’s multiple comparisons test (R).