Mitochondrial biogenesis in white adipose tissue mediated by JMJD1A-PGC-1 axis limits age-related metabolic disease

Summary Mitochondria play a vital role in non-shivering thermogenesis in both brown and subcutaneous white adipose tissues (BAT and scWAT, respectively). However, specific regulatory mechanisms driving mitochondrial function in these tissues have been unclear. Here we demonstrate that prolonged activation of β-adrenergic signaling induces epigenetic modifications in scWAT, specifically targeting the enhancers for the mitochondria master regulator genes Pgc1a/b. This is mediated at least partially through JMJD1A, a histone demethylase that in response to β-adrenergic signals, facilitates H3K9 demethylation of the Pgc1a/b enhancers, promoting mitochondrial biogenesis and the formation of beige adipocytes. Disruption of demethylation activity of JMJD1A in mice impairs activation of Pgc1a/b driven mitochondrial biogenesis and limits scWAT beiging, contributing to reduced energy expenditure, obesity, insulin resistance, and metabolic disorders. Notably, JMJD1A demethylase activity is not required for Pgc1a/b dependent thermogenic capacity of BAT especially during acute cold stress, emphasizing the importance of scWAT thermogenesis in overall energy metabolism.

The histidine 1120 of the Jumonji domain is highly conserved and essential for binding to Fe (II) (Figures 1C and S1B). 22This is reported to be required for its demethylase activity. 12To verify the functional significance of this specific residue, a mutation was introduced, replacing H1120 with phenylalanine (H1120F).Substituting histidine with phenylalanine or tyrosine is a common approach to assess the significance of a histidine residue in enzyme function due to their structural and chemical similarities to histidine.The wild type and its mutant versions were subsequently expressed and purified from insect Sf9 cells. 15Notably, only the wild-type recombinant protein exhibited significant H3K9 demethylase activity (Figures S1C and S1D), highlighting the crucial role of histidine 1120 in facilitating the demethylation process.
Next, we examined the role of H1120 in a cellular assay system using immortalized scWAT cells where the endogenous mouse Jmjd1a was knocked down using shRNA.These cells were then transduced with either wild type hJMJD1A or a mutant form carrying a mutation converting the conserved histidine to a tyrosine residue (H1120Y) which would also be predicted to eliminate demethylase activity. 19The cells transduced with wild type hJMJD1A displayed a high density of mitochondria after differentiation, indicating successful mitochondrial biogenesis (Figure 1D).However, cells transduced with the H1120Y mutant exhibited a significantly lower number of mitochondria and a reduced oxygen consumption rate (OCR) compared to those transduced with WT-hJMJD1A (Figures 1D and 1E).Importantly, the transduction of H1120Y-hJMJD1A not only impaired mitochondrial biogenesis (Figure 1D), but also blunted induction of the key thermogenic gene Ucp1. 19These results suggest that histone demethylation mediated by JMJD1A is crucial for activation of thermogenic gene expression and mitochondrial biogenesis during the differentiation of cultured preadipocytes into functional beige adipocytes.
To elucidate the physiological significance of JMJD1A-mediated histone demethylation in mitochondrial biogenesis in scWAT in live animals, we generated two lines of mice carrying a demethylation-defective form of JMJD1A with a histidine-1122 to tyrosine (H1122Y, or HY) mutation, which corresponds to the H1120Y mutation in human JMJD1A (Figures 1F, S1E, and S1F).Genotyping analysis of the offspring from Jmjd1a HY/+ parents revealed the number of mice with different genotypes.In line #1, there were 149 Jmjd1a +/+ mice, 241 Jmjd1a HY/+ mice, and 54 Jmjd1a HY/HY mice.In line #2, there were 65 Jmjd1a +/+ mice, 116 Jmjd1a HY/+ mice, and 18 Jmjd1a HY/HY mice (Figure S1G).These results indicated that fewer Jmjd1a HY/HY pups were obtained than the expected Mendelian ratio.Immunoblot analysis using a specific anti-JMJD1A antibody (Figure S1H) showed that the expression of JMJD1A protein in scWAT was comparable between Jmjd1a +/+ and Jmjd1a HY/HY (Figure S1I).However, it was observed that obtaining Jmjd1a HY/HY males in line #1 was more challenging compared to line #2.This difficulty in obtaining Jmjd1a HY/HY male in line #1 could be attributed to a phenomenon known as sex reversal, which had been previously reported in Jmjd1a knock-out mice. 23Therefore, female mice from line #1 and male mice from line #2 were used in the subsequent experiments to ensure an adequate representation of the Jmjd1a HY/HY genotype.
To determine the roles of JMJD1A demethylation activity in cold-induced mitochondrial biogenesis during beiging, mice were exposed to cold temperature for 1 week (Figure 1G).It was observed that Jmjd1a HY/HY mice had significantly lower levels of mitochondrial DNA-encoded mt-Nd1 gene compared to Jmjd1a +/+ mice after cold exposure (Figure 1H).Additionally, there was a decreasing trend in the levels of mt-Nd2 and mt-Nd4 genes in Jmjd1a HY/HY mice compared to Jmjd1a +/+ mice after cold exposure (Figure 1H).Under thermoneutral condition, the mRNA expression of nuclear-encoded mitochondrial Ndufs8, Atp5j2, and Cox5a genes were comparable between Jmjd1a +/+ and Jmjd1a HY/HY mice (Figure 1I).Upon cold exposure, these genes were significantly induced in Jmjd1a +/+ mice, indicating an activation of mitochondrial biogenesis.In contrast, the induction of these genes was significantly lower in Jmjd1a HY/HY compared to Jmjd1a +/+ mice (Figure 1I).In addition, the levels of mitochondrial proteins, including UQCRC2, NDUFB8, and COXIV, were lower in scWAT of Jmjd1a HY/HY mice Figure 1.Cold induced mitochondrial biogenesis in scWAT requires histone demethylation (A) Schematic illustration of the chronic cold exposure experiment.C57BL6/J (wild-type; WT) mice were housed at 8 C or 30 C for 1 week after acclimation in thermoneutral conditions for 1 week.(B) Mitochondrial DNA content measured by qPCR using primers for mt-Nd1, mt-Nd2 and mt-Nd4 in the subcutaneous white adipose tissue (scWAT) of WT mice housed at 8 C (n = 5) or 30 C (n = 5) for 1 week.(C) Schematic representation of the domain structure of wild-type (WT) or H1120Y mutant hJMJD1A.(D) Representative transmission electron microscope images of differentiated im-scWAT cells expressing WT-hJMJD1A or HY-hJMJD1A (scale bar = 0.5 mm).N indicates nucleus.(E) Oxygen consumption rate (OCR) of immortalized (im)-scWAT cells overexpressing WT-hJMJD1A or HY-hJMJD1A, as measured by a metabolic flux assay.Arrows indicate the time of addition of oligomycin (Oligo), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), and rotenone/antimycin A (Rot/ Anti).Basal, maximum, and uncoupled respiration were calculated (mean G SEM of five technical replicates).(F) Schematic diagram of the targeting strategy for the Jmjd1a H1122Y mutation using the CRISPR/Cas9 system.The gRNA-targeting sequence is underlined, the protospacer adjacent motif (PAM) sequence is colored green, and the mutated sequences of the donor oligo DNA are colored red.(G) Schematic illustration of the chronic cold exposure experiment.Jmjd1a +/+ (+/+) and Jmjd1a HY/HY (HY/HY) mice were housed at 8 C for 1 week after acclimation in thermoneutral conditions for 1 week.(H) Mitochondrial DNA content measured by qPCR using primers for the mt-Nd1, mt-Nd2 and mt-Nd4 genes in scWAT from WT (n = 7) and Jmjd1a HY/HY (n = 5) mice exposed to chronic cold exposure (8 C) for 1 week.(I) The mRNA levels of Ndufs8, Atp5j2 and Cox5a were determined by qPCR of scWAT from Jmjd1a +/+ or Jmjd1a HY/HY mice with or without chronic cold exposure (8 C) for 1 week.(J) Immunoblot analysis of oxidative phosphorylation (OXPHOS) proteins, COX-IV, ATP5A, UQCRC2, SDHB, and NDUFB8, in tissue homogenates of scWAT from Jmjd1a +/+ (n = 7) and Jmjd1a HY/HY mice (n = 5) with chronic cold exposure (8 C) for 1 week (left); band intensities were calculated using ImageJ software (right).Data are expressed as mean G SEM (B, E, H, I, J).Representative of two independent experiments (E).Welch's t test (B, E, H, I, J) was performed for comparison.*p < 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significant.n.s not significant.See also Figure S1.The uncropped images of the blots are shown in Figure S9.compared to Jmjd1a +/+ mice (Figure 1J).These results collectively indicate that histone demethylation activity of JMJD1A is required for coldinduced mitochondrial biogenesis in scWAT in vivo.

Cold-induced mitochondrial biogenesis regulated by JMJD1A facilitates energy expenditure
To investigate the contribution of JMDJ1A-mediated mitochondrial biogenesis in scWAT to energy metabolism, Jmjd1a +/+ and Jmjd1a HY/HY mice were acclimated to 30 C for 1 week and then housed at either 8 C or 30 C for an additional week as a control (Figure 2A).It was observed that Jmjd1a HY/HY mice had higher scWAT weight compared to Jmjd1a +/+ mice at 8 C (Figure 2B).Body weights of Jmjd1a +/+ mice were increased by cold exposure, while those of Jmjd1a HY/HY mice were unchanged.Gonadal WAT (gWAT) weights were similar between both genotypes at both 30 C and 8 C. On the other hand, BAT weights of Jmjd1a +/+ and Jmjd1a HY/HY mice were increased by cold exposure suggesting that cold-induced BAT hyperplasia was not affected by the defect of JMJD1A catalytic activity (Figure S2A).Histological analysis of scWAT using hematoxylin & eosin (HE) staining and immunostaining for UCP1, a marker of thermogenic adipocytes, revealed that Jmjd1a +/+ mice developed abundant clusters of UCP1-positive adipocytes with multilocular lipid droplets specifically at 8 C, indicating the presence of thermogenic adipocytes (Figure 2C, left and middle panels).In contrast, Jmjd1a HY/HY mice showed reduced level of UCP-1 positive adipocytes or multilocular lipid droplets in scWAT at 8 C suggesting impaired thermogenic capacity (Figures 2C, left and middle panels and S2B, left panel).Immunostaining for TOMM20, a mitochondrial marker, further supported these findings.Jmjd1a +/+ mice showed a strong signal in scWAT at 8 C, whereas scWAT from Jmjd1a HY/HY mice showed a weaker TOMM20 signal (Figures 2C, right panels and S2B, right panel).This indicates there was a more robust increase in mitochondria in Jmjd1a +/+ vs. Jmjd1a HY/HY mice.Importantly, the differences in scWAT phenotype between Jmjd1a +/+ and Jmjd1a HY/HY mice were not observed under thermoneutral condition (Figures S2B and S2C).In gWAT, adipocyte sizes of Jmjd1a HY/HY mice were significantly larger than those of Jmjd1a +/+ mice at 30 C, whereas at 8 C, there were no (E) qPCR analysis revealed comparable expression of the mitochondrial regulatory genes Pgc1a and Pgc1b, the mitochondrial genes Ndufs8, Sdhb and Atp5j2, the adipogenic gene Pparg, and the thermogenic genes Ucp1, Elovl3 and Dio2 in the BAT of 22-week-old Jmjd1a +/+ (n = 7) and Jmjd1a HY/HY (n = 7) mice under control or acute cold exposure (4 C for 8 h) conditions.
significant changes in the adipocyte size between the two genotypes (Figure S2D).These results indicate that JMJD1A-mediated mitochondrial biogenesis is specifically impaired under cold stress, resulting in a reduced presence of thermogenic adipocytes and compromised mitochondrial content in scWAT of Jmjd1a HY/HY mice.
Cold-induced hyperplasia of BAT was not impaired by the absence of JMJD1A catalytic activity with the increased BAT weight upon chronic cold exposure in Jmjd1a HY/HY mice (Figure S2A, right panel).Additionally, the UCP1 signals in BAT were comparable between Jmjd1a +/+ and Jmjd1a HY/HY mice at 30 C, indicating similar basal thermogenic capacity (Figure S2E).Interestingly, upon exposure to 8 C for 1 week, both Jmjd1a HY/HY and Jmjd1a +/+ mice exhibited similar levels of enhanced UCP1 signals in BAT, indicating the activation of thermogenesis in response to chronic cold stress was comparable.Under thermoneutral condition, BAT in Jmjd1a HY/HY mice showed increased lipid accumulation compared to Jmjd1a +/+ mice.However, after chronic cold exposure, the levels of lipid droplets in BAT reached similar levels between the two genotypes (Figure S2E).This suggests that the differences in lipid accumulation in BAT between Jmjd1a HY/HY and Jmjd1a +/+ mice at thermoneutrality are overcome upon chronic cold exposure, possibly due to the activation of thermogenic processes that promote lipid utilization.Furthermore, mRNA levels of genomic DNA-encoded mitochondrial genes, such as Ndufs8 and Sdhb, as well as mitochondrial regulatory genes including Pgc1a, Pgc1b, in BAT showed no significant differences between two genotypes after chronic cold exposure (Figure S2F).This suggests that the transcriptional regulation of these mitochondrial genes and regulators in BAT during cold exposure is not affected by the JMJD1A histone demethylation activity.
To assess the impact of cold-induced mitochondrial biogenesis on whole body energy expenditure, we measured OCR in Jmjd1a HY/HY and Jmjd1a +/+ mice.Norepinephrine (NE)-induced OCR after chronic cold exposure was significantly lower in Jmjd1a HY/HY compared to Jmjd1a +/+ mice (Figure 2D).This suggests that the deficiency in JMJD1A histone demethylation activity impairs the ability of scWAT to increase energy expenditure in response to chronic cold exposure.In contrast, under thermoneutral condition, there was no significant difference in OCR between the two genotypes (Figure S2G).This indicates that the role of JMJD1A-mediated mitochondrial biogenesis in energy expenditure is specific to the chronic cold-induced activation of scWAT.Taken together, these results highlight the importance of JMJD1A histone demethylation activity in facilitating cold-induced mitochondrial biogenesis and enhancing energy expenditure specifically in scWAT.Because BAT induced thermogenesis was similar between the Jmjd1a +/+ and Jmjd1a HY/HY mice, these results also demonstrate that scWAT beige conversion plays a significant role in whole body OCR (Figures 2B-2D, see below Figures 3A-3C).

Demethylation activity of JMJD1A is dispensable for mitochondrial biogenesis and thermogenesis in BAT
The above findings indicate that the demethylation activity of JMJD1A is dispensable for mitochondrial biogenesis and thermogenesis in BAT.Because global Jmjd1a-deficient mice failed to maintain body temperature in acute phase of cold exposure, 14,15 we assessed the thermogenic capacity and energy expenditure in BAT in Jmjd1a +/+ vs Jmjd1a HY/HY mice.The mice were exposed to acute cold stress by keeping them at 4 C for 8 hours after one week of acclimation at 30 C (Figure 3A).Interestingly, we found that the body temperatures of Jmjd1a HY/HY mice were similar to those of Jmjd1a +/+ mice throughout the 8 hour-cold exposure period (Figure 3B).We also measured the actual tissue temperature of BAT, as well as rectal and skeletal muscle temperature, and observed no significant differences between Jmjd1a HY/HY and Jmjd1a +/+ mice after NE injection (Figure 3C).Jmjd1a HY/HY mice have broader lipid droplet area in BAT than Jmjd1a +/+ mice at 30 C, while there was no difference in accumulated lipid between two genotypes at 8 C (Figures 3D and S3A).Furthermore, the expression of mitochondrial genes (Ndufs8, Sdhb, Atp5j2) and the adipogenic gene (Pparg) in BAT were comparable between the two genotype groups (Figure 3E).Importantly, mitochondrial regulatory genes (Pgc1a, Pgc1b) and thermogenesis genes (Ucp1, Elovl3, Dio2) were also similar in both genotypes under acute cold stress, indicating that demethylation activity of JMJD1A is not required for the induction of these genes in BAT (Figure 3E).Immunoblot analysis confirmed similar levels of induction of Ucp1 expression after 8 h-cold exposure in both genotypes (Figures 3F  and S3B).Also, the H1122Y mutation did not affect b-AR stimulated JMJD1A phosphorylation at S265, which is necessary for the catalyticindependent induction of thermogenic genes in BAT 14,15 (Figure 3G).These results extend our previous studies using cultured brown adipocytes 14,15 and demonstrate that the catalytic activity of JMJD1A is not necessary for the thermogenic function in mouse BAT in mice during acute cold exposure (Figure 3H).

Chronic cold stress-induced Pgc1 in scWAT requires histone demethylation of H3K9me2
To gain further insights into the impact of JMJD1A demethylation activity on thermogenesis, we conducted RNA-seq analysis using scWAT from Jmjd1a HY/HY and wild-type mice exposed to either cold (8 C) or thermoneutral (30 C) conditions for one week (Figure 4A).Principal component analysis (PCA) clearly demonstrated the distinct transcriptome profiles for each group (Figure S4A).Analysis of differentially expressed genes in wild-type scWAT between 8 C and 30 C revealed 1,815 genes that were increased by cold exposure, including key .Continued (F) Immunoblot analysis of UCP1 in tissue homogenates of BAT from Jmjd1a +/+ and Jmjd1a HY/HY mice.Actin was used as a loading control.(G) Immortalized scWAT cells from Jmjd1a +/+ and Jmjd1a HY/HY mice were treated with 10 mM isoproterenol (ISO) for 20 min.Homogenates of these cells were subjected to immunoprecipitation (IP) with anti-mJMJD1A, followed by immunoblotting (IB) with anti-phospho-JMJD1A S265 or anti-mJMJD1A antibodies.(H) Model of adaptation to cold environment through JMJD1A phosphorylation and histone demethylation activity.In the acute phase, JMJD1A contributes to thermogenesis in BAT via phosphorylation and chromatin conformation changes.In the chronic phase, JMJD1A promotes thermogenesis through histone demethylation in scWAT.Jmjd1a HY/HY mice have an impaired adaptive potential to chronic cold, although they retain the adaptability to acute cold.Data are the mean G SEM (B, C, E).Repeated-measures ANOVA with a post-hoc Welch's t test (B) or Welch's t test (C, E) were performed for comparison.*p < 0.05 was considered statistically significant.n.s not significant.See also Figure S3.Uncropped images of the blots are shown in Figure S9.Oxidative phosphorylation

Pgc1a
Relative mRNA levels in scWAT Figure 4.The demethylation activity of JMJD1A is required for the expression of thermogenesis and mitochondrial genes in response to chronic cold (A) Schematic illustration of the chronic cold exposure experiment.(B) Volcano plots comparing the gene expression in scWAT of Jmjd1a +/+ mice before and after cold exposure (8 C) (left) and scWAT of Jmjd1a +/+ and Jmjd1a HY/HY mice after cold exposure (right).(C) Venn diagram showing the cold-induced genes (1,815 genes, defined in (B), left panel, up-regulated genes) and JMJD1A demethylation activity-dependent genes (1,115 genes, defined in (B), right panel, down-regulated genes).Overlapping genes were identified as JMJD1A demethylation-dependent cold-induced genes (857 genes).
beige-selective genes such as Ucp1, Cidea, and Cpt1b, which were defined as cold-induced genes (Figure 4B, left).Under cold condition, 1,115 genes were expressed at lower levels in Jmjd1a HY/HY compared to wild-type mice, and these were defined as JMJD1A demethylation activity-dependent genes (Figure 4B, right).By comparing the genes from these two defined groups, we identified 857 cold-induced JMJD1A demethylation-dependent genes (Figure 4C).Pathway enrichment analysis showed that these genes were related to oxidative phosphorylation (OXPHOS) and thermogenesis (Figure 4D).Gene set enrichment analysis (GSEA) further supported these findings, indicating that inactivation of JMJD1A catalytic activity prevented the induction of thermogenesis and OXPHOS gene expression in scWAT under chronic cold exposure (Figures S4B and S4C).Specifically, the JMJD1A H1122Y mutation resulted in the loss of induction of OXPHOS genes, including both genomic DNA-encoded and mitochondrial DNA-encoded genes (Figures 4E, 4F, S4D, and S4E), as well as thermogenesis genes (Figure 4G).In search for the regulatory genes that might be responsible for cold-induced mitochondrial biogenesis, we specifically examined regulatory genes associated with mitochondrial gene expression, as defined in WikiPathways.Our analysis revealed that inactivation of JMJD1A catalytic activity abolished induction of Pgc1b, Pgc1a, estrogen related receptor a (Esrra) and Gabpa expression in scWAT under cold conditions (Figure 4H).ESRRA and GABPA were known to regulate the transcription for mitochondrial genes through the DNA-binding and cooperation with PGC-1. 24These findings strongly suggest that demethylation activity of JMJD1A is crucial for the activation of Pgc1 genes and the subsequent mitochondrial biogenesis during scWAT beiging in response to chronic cold exposure.Furthermore, we examined the expression of genes involved in mitochondrial fission and fusion, as these processes play a role in maintaining mitochondrial quality. 25,26However, we did not observe any differences in the mRNA expression of mitofission-related genes (Drp1, Mff, Fis1, and Mief1) or mitofusionrelated genes (Mfn1, Mfn2, Opa1 and Oma1) between Jmjd1a +/+ and Jmjd1a HY/HY (Figures S4F and S4G).Taken together, these results indicate that the catalytic activity of JMJD1A specifically affects Pgc1 function and mitochondrial biogenesis in scWAT beiging under chronic cold exposure, rather than influencing mitochondrial dynamics through fission and fusion processes.

Cold-sensitive JMJD1A demethylates H3K9me2 in the enhancers of the mitochondrial master regulator Pgc1, inducing gene expression
To gain insight into the direct target genes of JMJD1A in beige adipocytes, we utilized previously published chromatin immunoprecipitation sequencing (ChIP-seq) data for JMJD1A in cultured beige adipocytes. 19This data was integrated with H3K27ac ChIP-seq data from beige adipocytes in scWAT under two conditions; exposure of mice to cold (referred to as cold beige) and following exposure to warm conditions after induction of beige adipocytes formation (referred to as warm beige). 21Initially, we classified H3K27ac peaks into 3 categories: cold beige-specific peaks (10,385 regions with fold change [FC] of H3K27ac cold beige/H3K27ac warm beige >2), warm beige-specific peaks (5,744 regions with an FC of cold beige H3K27ac/warm beige H3K27ac < 0.5), and common H3K27ac peaks (17,795 regions with an FC of cold beige H3K27ac/warm beige H3K27ac from 0.5 to 2) (Figure S5A).Next, we compared cold beige-specific H3K27ac peaks (10,385 peaks) with JMJD1A binding sites (39,364 peaks), and identified 2,089 overlapping peaks that were annotated to promoters of 1,647 genes (defined as 2 kb upstream to 1 kb downstream of TSS) (Figure S5B).Further comparison of the 857 genes cold-induced JMJD1A demethylation-dependent (as shown in Figure 4C) with 1,647 genes that exhibited both JMJD1A binding and cold beige-specific H3K27ac peaks lead to identification of 228 genes, including Pgc1b and Pgc1a.These 228 genes were defined as JMJD1A direct target genes, as they met the criteria of being cold-induced, JMJD1A demethylation-dependent, exhibiting cold beige-specific H3K27ac peaks, and having JMJD1A binding (Figure S5B; Table S1).
To identify putative enhancers of mitochondrial regulatory genes, particularly Pgc1b, that are epigenetically regulated under cold conditions, we conducted transposase-accessible chromatin sequencing (ATAC-seq) analysis using cultured beige adipocytes derived from Jmjd1a +/+ and Jmjd1a HY/HY .The ATAC-seq data revealed periodicity signals matching nucleosome units (Figure S6A) and their distribution along the transcription start site (TSS) (Figure S6B).Interestingly, the distribution of ATAC-seq peaks showed a similar pattern between Jmjd1a +/+ or Jmjd1a HY/HY cells, with a higher proportion of peaks (27.2% for Jmjd1a +/+ and 22.3% for Jmjd1a HY/HY cultured adipocytes, respectively) located at promoter-transcription start sites (TSS) among all the peaks (Figure S6C).This indicates that the chromatin accessibility profile, as indicated by the ATAC-seq peaks, was not significantly affected by loss of demethylation activity of JMJD1A (Figure S6C).By comparing open chromatin peaks in cultured preadipocytes and adipocytes that were differentiated into beige adipocytes in both genotypes, we identified total of 10,048 peaks in Jmjd1a +/+ and 12,500 in Jmjd1a HY/HY cultured adipocytes.These peaks were specifically present in differentiated beige adipocytes (Figure S6D).
Gene ontology (GO) analysis of the 5,224 and 6,061 genes associated with the adipocyte differentiation-specific peaks for Jmjd1a +/+ (10,048 peaks) or Jmjd1a HY/HY (12,500 peaks) respectively (Figure S6D) demonstrated an enrichment of cold-induced thermogenesis-related genes in adipocytes for both genotypes (Figure S6E).Moreover, we observed a significant enrichment of the EBF2, CEBP, and NFI transcription factor binding motifs within the identified peaks specific to beige adipocytes in both genotypes.This is significant because these factors (E) Heatmap showing the down-regulated genomic DNA-encoded genes, including the OXPHOS genes, determined by RNA-seq analysis of Jmjd1a +/+ or Jmjd1a HY/HY scWAT at 30 C or 8 C. A color intensity scale is provided for reference.(F-H) Heatmap showing mitochondrial DNA-encoded genes (F), thermogenesis genes (G) and mitochondrial regulatory genes (H).Mitochondrial regulatory genes were defined as ''mitochondrial gene expression''-related genes ways (https://www.wikipathways.org/).See also Figure S4.are well-studied transcription factors involved in the induction of thermogenic genes [27][28][29][30] (Figure S6F).These results indicate that our ATACseq data successfully captured regions of open chromatin that undergo dynamic changes during the differentiation of beige adipocytes.

I
To identify enhancer regions that are associated with beige specific genes influenced by both cold and thermoneutral temperatures, we integrated H3K27ac and H3K4me1 ChIP-seq dataset from beige adipocytes in scWAT of NuTRAP mice, a system allowing histone modification analysis from single cell types in animal tissues (explained in more detail below), 21 with our ATAC-seq dataset from cultured beige adipocytes.This analysis revealed predicted enhancer regions of several genes, including Pgc1b (À38, À48, and À73 kb from transcription start site [TSS]), Pgc1a (À42, À272 and À314 kb) (mitochondrial regulatory genes) (Figure 5A), Ucp1 (À4.8 kb), Ppara (À10 kb) and Cidea (À13.5 kb) (thermogenesis genes) (Figure S7A).Notably, the predicted enhancer regions for Ucp1, Ppara and Cidea were consistent with previous reports. 14To validate the predicted enhancer of Pgc1b and Pgc1a, we analyzed published HiC-data and confirmed that these enhancers (À38 and À48 kb of Pgc1b and À42, À272, and À314kb of Pgc1a) physically interacted with the TSS in 3T3-L1 adipocytes 31 (Figure S7B).Furthermore, information in the database EnhancerAtlas also indicated that the distal regions of Pgc1b were annotated as enhancers in BAT 32 (Figure 5A).
To assess the H3K9me2 levels of these differentiation regulated enhancers in beige adipocytes in vivo, we utilized NuTRAP::Ucp1-Cre mice, which allowed us to purify the nuclei specifically from beige adipocytes 21,33 within a large background of white adipocytes.These mice expressed mCherry-labeled nuclei and EGFP-labeled ribosomes exclusively in beige adipocytes (Figure 5B).Following cold exposure, we isolated mCherry-labeled nuclei from WAT using flow cytometry (Figures 5C and S7C).Isolated nuclei were cross-linked, immunoprecipitated with anti-H3K9me2 antibody followed by ChIP-qPCR.The results indicated that the H3K9me2 levels at the enhancers of Pgc1b (À48 and À38 kb) and Pgc1a (À314, À272 and À42 kb) in mCherry-positive-beige cell nuclei were significantly reduced as compared to the H3K9me2 levels in mCherry-negative nuclei (non-beige cell nuclei) (Figure 5D), indicating that demethylation of H3K9me2 occurs in beige adipocytes in both in vitro and in vivo settings.These data also underscore that active epigenetic marks H3K27ac and H3K4me1 on Pgc1a/b enhancers overlapped with reduced H3K9me2 levels in beige adipocytes.Furthermore, the ChIP-qPCR analysis provided additional evidence of increased H3K27ac levels (Figures 5E and S7D) and recruitment of JMJD1A (Figure 5F) in the enhancers of Pgc1b and Pgc1a, while simultaneously showing a decrease in H3K9me2 levels to these regions during beige differentiation of cultured preadipocytes derived from scWAT (Figures 5G and S7E).Similar results were obtained for the Ucp1, Ppara, and Cidea enhancers.By contrast, there were minimal effects observed on histone modification or JMJD1A recruitment to the control region of Prdx5 gene (Figures 5E-5G).Consistent with a key role for the enzyme activity of JMJD1A during adipocyte differentiation, we observed lower expression of Ucp1, Cidea, Ppara, and Pgc1b in differentiated im-scWAT adipocytes derived from Jmjd1a HY/HY mice compared to Jmjd1a +/+ mice (Figure 5H).Additionally, H3K9me2 levels remained higher in Jmjd1a HY/HY im-scWAT cells at the enhancers of Pgc1b, Ucp1, Ppara, and Cidea after beige adipocyte differentiation (Figures 5I and S7F).Interestingly, despite the lower expression and the higher histone methylation of these key adipogenic genes, ATAC-seq analysis from differentiated Jmjd1a HY/HY im-scWAT cells before and after differentiation revealed an open chromatin state in these enhancers in both genotypes (Figure 5A, track 8 and 10), indicating that these enhancers are open upon differentiation and that H3K9me2 regulation occurs through changes in gene function not directly related to chromatin accessibility.The precise biological mechanism that H3K9me2 contributes to in this context remains unclear.However, these results indicate that the expression of Pgc1 genes, key regulators of mitochondrial biogenesis, and other thermogenic genes are epigenetically controlled by JMJD1A during beige adipogenesis.
Impaired mitochondrial biogenesis observed in Jmjd1a HY/HY mice leads to obesity and metabolic disorders in aged mice Next, we followed male and female Jmjd1a +/+ and Jmjd1a HY/HY mice over the course of a sixty-week time course at room temperature (23 C).5][36] Both genotypes of mice gained weight over time, however both male and female Jmjd1a HY/HY gained more weight than the control starting at around 30 weeks of age (Figures 6A-6C).Because mice typically exhibit a partial cold-induced thermogenic response characterized by increased energy expenditure and activation of thermogenic adipocytes at 23 C relative to thermoneutrality, we hypothesized that the Jmjd1a HY/HY mice may gain more weight because of a defect in mitochondrial biogenesis and beige associated thermogenesis.Consistent with this hypothesis, mitochondrial DNA qPCR analysis revealed that Jmjd1a HY/HY mice had lower mitochondrial DNA content compared to Jmjd1a +/+ mice at 53 weeks of age (Figure 6D).Histological analysis showed that the Jmjd1a +/+ mice developed a small population of scWAT that express both UCP1 and TOMM20, whereas similarly aged Jmjd1a HY/HY did not exhibit the same phenotype (Figures 6E and S8A).Indirect calorimetry analysis demonstrated that Jmjd1a HY/HY mice had significantly lower oxygen consumption and carbon dioxide production during both light and dark phases as compared to Jmjd1a +/+ mice (Figures 6F and S8B, left and middle).These results suggest that mitochondrial biogenesis and beige adipocyte conversion in scWAT contributes to an increase in energy expenditure that significantly contributes to the management of body weight over time.However, there was no significant difference in respiratory quotient (RQ) between the two genotype groups, indicating that fuel substrates utilized by both genotypes were not significantly different (Figures S8B, right and S8C).Analysis of body composition using X-ray computed tomography (CT) scanning revealed marked fat deposition in scWAT and visceral adipose tissue (vWAT) of Jmjd1a HY/HY mice at 43 weeks of age (Figure 6G).Anatomical analysis confirmed that Jmjd1a HY/HY mice had a greater body weight at the time of CT scanning (Figure S8D).Moreover, the total adipose tissue weight, including both vWAT and scWAT, in Jmjd1a HY/HY mice was approximately 3.8 times higher compared to Jmjd1a +/+ mice (Figure S8E).Specifically, the weights of scWAT and vWAT, including mesenteric WAT (mWAT) and epididymal WAT (eWAT), were significantly higher in Jmjd1a HY/HY mice compared to control mice (Figures 6H, S8F, and S8G).Female mice also showed an increasing trend in BAT and gWAT weights (Figures 6H, S8H, and S8I).Histological analysis further revealed lipid accumulation in BAT and adipocyte hypertrophy in scWAT, visceral mWAT, and gWAT in Jmjd1a HY/HY mice (Figures 6I and S8J-S8L).Notably, the presence of crown-like structures (CLS) was observed in visceral eWAT of Jmjd1a HY/HY mice, suggesting these mice also suffered from low grade chronic inflammation (Figure 6I).Consistent with this observation, the mRNA expression levels of inflammation-related genes, Ccl2, Ccl5, Tnf-a, F4/80 and Cd11b in eWAT were significantly higher in Jmjd1a HY/HY mice (Figure S8M).These findings suggest that impaired mitochondrial biogenesis in scWAT due to JMJD1A deficiency contributes to obesity, altered body composition, and possibly metabolic disorders.
In fact, the aged Jmjd1a HY/HY mice had significantly higher levels of plasma insulin, cholesterol, and triglyceride compared to Jmjd1a +/+ mice (Figures 6J-6L).Additionally, Jmjd1a HY/HY mice showed lower level of non-esterified fatty acid (NEFA) after fasting (Figure S8N).However, the levels of plasma glucose and total ketone bodies were comparable between the two genotype groups (Figures S8O and S8P).Plasma leptin, a hormone associated with satiety and adiposity, was approximately 3 times higher in Jmjd1a HY/HY mice consistent with adipose tissue dysfunction (Figure S8Q).During the glucose tolerance test, Jmjd1a HY/HY mice had higher plasma glucose and insulin levels compared to control mice (Figures 6M and 6N).Additionally, the insulin tolerance test showed higher plasma glucose levels in Jmjd1a HY/HY mice compared to Jmjd1a +/+ mice (Figure 6O).These findings indicate that the glucose intolerance in Jmjd1a HY/HY mice is attributable to insulin insensitivity.The presence of CLS in the visceral eWAT of Jmjd1a HY/HY mice (Figure 6I) suggests that adipose tissue inflammation may contribute, at least in partially, to the insulin insensitivity observed in Jmjd1a HY/HY mice.Based on these results, it can be concluded that Jmjd1a HY/HY mice developed a range of biochemical abnormalities that are hallmarks of metabolic diseases that accumulate over time.The study highlights the importance of mitochondrial biogenesis in scWAT, especially in response to mild cold exposure (room temperature), as a protective mechanism against age related obesity and metabolic disorders.
To investigate if the metabolic abnormalities developed prior to the noticeable differences in body weight, we performed body composition measurements on young mice aged 8 to 14 weeks, which did not exhibit any body weight difference compared to Jmjd1a +/+ mice (Figure 6P).CT analysis revealed significantly higher adipose tissue content, including visceral and subcutaneous WAT, in Jmjd1a HY/HY mice (Figures 6Q-6S).Notably, at 15 weeks of age, Jmjd1a HY/HY mice showed similar glucose tolerance despite having larger adipose tissue compared to Jmjd1a +/+ mice (Figure 6T).These observations suggest that young Jmjd1a HY/HY mice experience normal weight obesity and display metabolic abnormalities as they age.Overall, the findings of this study demonstrate that Jmjd1a HY/HY mice exhibit a spectrum of metabolic abnormalities associated with aging, resembling characteristic features of metabolic diseases.These results underscore the crucial role of mitochondrial biogenesis in scWAT, particularly in response to mild cold exposure, as a protective mechanism against obesity and metabolic disorders.They also are consistent with the obesity/metabolic disease paradigm that excess body weight that occurs during aging precedes the appearance of significant metabolic disorders.(F) The mean of oxygen consumption (VO 2 ) and carbon dioxide production (VCO 2 ) was calculated every 12 h.VO 2 and VCO 2 were normalized to mouse body weight (42 weeks old, Jmjd1a +/+ : n = 5, Jmjd1a HY/HY : n = 4).(G) Representative computed tomography (CT) images of the lumber L4 vertebra of Jmjd1a +/+ and Jmjd1a HY/HY female mice.Visceral fat, subcutaneous fat, and muscle are shown in purple, yellow, and blue, respectively.(H) Weights of BAT, scWAT, mesenteric (m)WAT, gonadal (g)WAT from female mice (53 weeks old, Jmjd1a +/+ : n = 5, Jmjd1a HY/HY : n = 4), and epididymal (e)WAT from male mice (57 weeks old, Jmjd1a +/+ : n = 6, Jmjd1a HY/HY : n = 5).(I) H&E stained sections of BAT (scale bar, 100 mm), scWAT, and mWAT from Jmjd1a +/+ and Jmjd1a HY/HY female mice, and eWAT from male mice (scale bar, 50 mm).(J-L) Plasma insulin (J), cholesterol (K), and triglycerides (L) were measured in Jmjd1a +/+ and Jmjd1a HY/HY mice in the fed state.(M and N) Glucose tolerance test (GTT).Plasma glucose (M) and insulin (N) concentrations were measured after the intraperitoneal injection of 2 g/kg glucose (49 weeks old, Jmjd1a +/+ : n = 5, Jmjd1a HY/HY : n = 4).(O) Insulin tolerance test (ITT).Plasma glucose concentration was measured after the administration of 0.75 U/kg insulin (52 weeks old, Jmjd1a +/+ : n = 5, Jmjd1a HY/HY : n = 4).(P-S) Changes in body weight (P) and fat mass determined by X-ray CT (Q-S) in 8-to 14-week-old Jmjd1a +/+ and Jmjd1a HY/HY female mice (Jmjd1a +/+ : n = 9, Jmjd1a HY/HY : n = 7).(T) GTT was conducted as described in (M) using 15-week-old female mice (Jmjd1a +/+ : n = 11, Jmjd1a HY/HY : n = 4).Data are the mean G SEM ( A, B, D, F, H, J-T).Repeated-measures ANOVA with a post-hoc Welch's -test (A, B, M-T) or Welch's t test was performed for comparison in (D, F, H, J-L).*p < 0.05 and **p < 0.01, and ***p < 0.001 were considered statistically significant.n.s indicates no significance.See also Figure S8.
To further investigate the link between JMJD1A and metabolic disease, we utilized datasets of human metabolic parameters and adipose tissue gene expression from METSIM (METabolic Syndrome In Men), a cohort study of Finnish men. 37,38Our analysis revealed negatively correlation between JMJD1A expression and BMI, waist circumference, hip circumference, serum triglycerides, and serum cholesterol (Figures 7A-7F).This implies that JMJD1A may play a role in enhancing energy metabolism in adipose tissue and potentially preventing obesity in humans.

DISCUSSION
In our current study, we extend our previous studies 14,15,19 to show that the expression of key mitochondrial regulators, such as Pgc1a and b, in scWAT are regulated by the histone demethylase JMJD1A in an external temperature-dependent manner in mice.We generated a mouse model with a single amino acid change in the Jumonji domain that eliminates iron binding which is critical to the a-ketoglutarate dependent demethylase activity of all Jumonji domain proteins.We used this Jmjd1a HY/HY mouse model to demonstrate that JMJD1A plays a pivotal role in modulating H3K9me2 levels at enhancers of key mitochondrial genes such as Pgc1a/b to promote their expression, thereby facilitating mitochondrial biogenesis and the development of beige adipocytes during cold exposure in mice (Figure 8A).Our previous studies 14,15 (also reviewed in [16][17][18] ) combined with our current observations show that JMJD1A regulates thermogenesis in beige scWAT through a sequential two-step mechanism, starting with the phosphorylation of JMJD1A at amino acid S265, followed by the demethylation of histone H3K9, ultimately leading to the induction of mitochondrial biogenesis and beige adipogenesis.Notably, our data suggests that inactivation of JMJD1A's catalytic activity does not influence chromatin opening (Figure S6C), suggesting that the demethylation of H3K9me2 promotes gene transcription by facilitating the recruitment of coactivators 19 and other histone modifications, such as H3K9 acetylation to open sites.
Our study revealed that defects in JMJD1A-mediated demethylation is associated with obesity and metabolic disorders as mice age on a normal chow diet.Interestingly, young mice fed a normal diet showed no difference in body weight but had increased body fat mass (Figures 6Pand 6Q).These mice exhibited decreased scWAT beiging and thermogenesis due to reduced mitochondrial biosynthesis under mild cold exposure (room temperature).These phenotypes are recognized as ''normal weight obesity'' or ''sarcopenic obesity'' in humans which are linked to a higher risk of future metabolic abnormalities and insulin resistance. 39Mechanistically, our study emphasizes the critical role of JMJD1A in histone demethylation of key mitochondrial regulatory genes, PGC-1a and PGC-1b, for preventing obesity and metabolic abnormalities.
Aging is associated with a decline in mitochondrial quantity and function, contributing to the development of obesity over the life span in humans. 40Additionally, the activity of brown adipose and beige adipose tissues declines as humans age as well. 41This has sparked a growing interest in approaches to enhance mitochondrial function in adipose tissue to address and prevent obesity and metabolic diseases. 10The regulation of mitochondrial activity through histone demethylation holds promise as a potential avenue to achieve this objective.
JMJD1A also functions as a chromatin scaffold in brown adipocytes in BAT, which is independent of its enzyme activity. 15Importantly, our study further demonstrates that the absence of catalytic activity of JMJD1A does not influence brown adipocyte thermogenic functions, and thus, it has no impact on cold tolerance during acute cold stress (Figure 3), nor does it appear to be involved in the hyperplasia observed under chronic cold stress (Figures S2A and S2F).These findings indicate that JMJD1A's contribution to acute thermogenesis in brown adipocytes is Pearson's correlation coefficient was used to determine the correlation between JMJD1A expression level and the indicated parameters.These datasets were obtained from GSE70353. 37,38ndependent of histone demethylation activity.15]19 This critical result emphasizes the complementary roles played by JMJD1A in regulating thermogenesis in the two different types of thermogenic adipocytes, brown and beige.Although JMJD1A catalytic activity has no impact on brown adipocyte thermogenic function, following the Jmjd1a HY/HY over time also revealed a key role for JMJD1A in regulating body weight as mice age (Figure 6).The absence of JMJD1A's catalytic activity led to impaired mitochondrial biogenesis in scWAT, resulting in increased weight gain that proceeded the development of metabolic abnormalities including glucose intolerance (Figure 6).These observations highlight the significance of JMJD1A in regulating epigenetic remodeling, promoting mitochondrial biogenesis in scWAT, which in turn prevents excess weight gain and glucose intolerance as mice age.These findings also provide significant support for the model that obesity associated metabolic disease is a consequence of excess weight gain.
Thermogenic adipocytes, brown adipocytes in BAT and beige adipocytes in scWAT, play a significant role in energy expenditure.Previous studies have highlighted the importance of thermogenesis in beige scWAT in maintaining systemic energy metabolism.For example, one study demonstrated that the loss of CD81, a marker of beige adipocyte progenitor cells, impairs the differentiation of beige adipocytes, but not brown adipocytes, leading to diet-induced obesity and insulin resistance. 42In contrast, the specific deficiency of PRDM16, a transcriptional regulator that promotes the differentiation of both brown and beige adipocytes, in brown adipocytes alone did not result in obesity. 43y elucidating the specific contribution of Jmjd1a in beige scWAT thermogenesis, our study provides additional evidence for the importance of beige adipocytes in maintaining energy balance and metabolic homeostasis.
Previous in vivo studies using Jmjd1a knock-in mice with a S265A mutation (preventing phosphorylation of JMJD1A) have shown reduced energy expenditure in both BAT and scWAT (Figure 8B), 14 however, these mouse models were not sufficient for directly investigating the twostep regulatory mechanism mediated by JMJD1A in vivo.Therefore, we engineered a new complementary mouse model with a selective inactivation of the catalytic activity of JMJD1A, and results presented here provide more direct in vivo evidence supporting the involvement of the two-step mechanism in the induction of genes involved in mitochondrial biogenesis such as Pgc1a and Pgc1b in scWAT.Figure 8B presents an overview of the distinct yet overlapping phenotypes revealed by our studies using the three mouse models: step 1 signal sensing deficient mice (knock-in mice with a S265A mutation), step 2 epigenetic rewriting defective point mutant mice (knock-in mice with an H1122Y mutation).The combined studies expand our understanding beyond the regulation of beige-selective genes and highlights the broader role of JMJD1A in promoting mitochondrial biogenesis in response to b-AR signaling and in thermogenic regulation during cold exposure and aging.
In summary, our data highlight the importance of JMJD1A in epigenetic regulation of mitochondrial biogenesis in scWAT downstream of b-AR signaling.This process plays a critical role in regulating energy expenditure and also in preventing obesity associated metabolic disease as mice age.Thus, the concept of tissue-specific regulation of mitochondrial biogenesis through the rewritable epigenome provides a new perspective for the development of therapeutic interventions aimed at improving energy metabolism and combating metabolic disorders associated with obesity.For oil red O (ORO) staining, cells at specified stages of differentiation were rinsed with PBS and fixed in 3.7% formaldehyde in H 2 O for 10 min.After two washes with PBS and one wash with 60% isopropanol for 1 min, the cells were stained for 15 min in freshly diluted ORO solution (0.18% (wt/vol) ORO in 60% isopropanol).The stain was removed, and the cells were washed twice with PBS and photographed as previously described. 58,59THOD DETAILS

Generation of Jmjd1a HY/HY mice
To generate Jmjd1a HY/HY mice, the following procedure was followed: a CRISPR-Cas9 expression vector (5 ng/mL), sgRNA targeting Jmjd1a exon 22 (5'-TACATCTAAGTGAAGATTTG -3'), and single-stranded donor DNA (10 ng/mL) containing the H1122Y mutation were injected into fertilized C57BL/6J eggs.The detailed method for harvesting fertilized eggs has been previously described. 60The injected single-cell embryos were transferred to pseudopregnant ICR mice.PCR-based screening and Sanger sequencing were performed to identify the founder mice.Genotyping was performed using clipped toes as the source of DNA.To amplify the WT allele, a specific set of primers (5'-GGAAA TATGGGACCACAAATCTTC-3' for WT allele) and a reverse primer (5'-AGACAGTAAGCCCAGCTTCAA-3for both WT and Jmjd1a HY allele) were used.To amplify the Jmjd1a HY allele, a set of primers (5'-GGAAATATGGGACCACTAACTTGT-3' for Jmjd1a HY allele), along with the common reverse primer for both the WT and Jmjd1a HY allele, were used.

Cold exposure experiments
Mice were individually caged in a thermostatic chamber iB-230 (TAITEC, Aichi, Japan) and allowed to acclimate to thermal neutrality ( 30 C) for 1 week prior to the initiation of the cold exposure experiments.For acute cold exposure, the mice were transferred to different thermostatic chambers and maintained at 4 C for 8 h periods.For chronic cold exposure, the mice were placed in a chamber maintained at 8 C for 1 week.

Body temperature
Mice were restrained by hand, and a thermocouple RET-3 rectal probe (Physitemp, Clifton, NJ, USA) was gently inserted approximately 2 cm into the anus of the mice and held for 3 s to obtain the rectal temperature.The rectal temperature was monitored using a digital thermometer BDT-100 (BRC, Aichi, Japan) connected to the thermocouple RET-3 rectal probe.

Indirect calorimetry
Mice were placed in an individual acrylic metabolic chamber either set at RT or in a thermostatic chamber maintained at the indicated temperatures with unrestricted access to water and food and maintained on a 12-h light/dark cycle.Oxygen consumption (VO 2 ) and carbon dioxide production (VCO 2 ) were monitored using an O 2 /CO 2 metabolic measurement system (MK-5000RQ; Muromachi Kikai, Japan).Measurements were taken every 3 min for 2 consecutive days.The respiratory quotient (RQ) was calculated as the ratio of VCO 2 to VO 2 .The mean values of 240 measurements of the O 2 and CO 2 concentrations were used to calculate the total O 2 consumption and CO 2 production in the mice during each 12-h period.
To analyze the thermogenic function induced by norepinephrine (NE, Sigma-Aldrich), mice were intraperitoneally injected with NE (1 mg/kg body weight) and oxygen consumption was measured for 1 h using an indirect calorimeter (MK-5000RQ).Daily energy expenditure was also measured using an MK-5000RQ.The chamber volume was 720 ml, and the airflow to the chamber was 400 ml/min for mice acclimated to 30 C and 800 ml/min for mice acclimated to 8 C. Samples were collected every 3 min, and a standard gas reference was taken every 30 min.

Metabolite measurements
Plasma glucose levels were determined using the Glucose CII test (Wako, Osaka, Japan).Plasma non-esterified fatty acids (NEFA) and total ketone body levels were determined using the NEFA C test (Wako) and Autokit Total Ketone Bodies test (Wako), respectively.Plasma triglyceride and cholesterol levels were measured using the triglyceride E-test (Wako) and cholesterol E-test (Wako), respectively.Plasma insulin and leptin levels were determined using a mouse insulin ELISA kit (U-type, Wako) and a mouse leptin immunoassay kit (Morinaga Institute of Biological Science, Inc.), respectively, according to the manufacturer's instructions.
For the glucose tolerance tests (GTTs), mice were fasted overnight for 16 h, glucose (2 g/kg body weight) was administrated intraperitoneally, and the plasma glucose concentration was determined using the Glucose CII test (Wako).The plasma insulin concentration was measured during the GTT using an LBIS mouse insulin ELISA kit (U-type, Wako).For the insulin tolerance test (ITT), non-fasted mice were injected intraperitoneally with 0.75 U/kg insulin (Sigma-Aldrich) approximately 2 h after the start of the light cycle.

Computed tomography analysis
To measure the fat content of the mice, computed tomography (CT) was performed using a LaTheta LCT-200 (Hitachi-Aloka, Tokyo, Japan) system under anesthesia with 0.3 mg/kg medetomidine, 4 mg/mL midazolam, and 5 mg/mL butorphanol.To quantify the visceral and subcutaneous fat depots, the area between the proximal end of the lumbar vertebra L1 and the distal end of L6 was scanned.The weight of the fat depots was calculated using LaTheta software.

Histological analysis
Excised BAT and scWAT samples were fixed in 10% (v/v) neutral buffered formalin for 48 h at 4 C and embedded in paraffin.Immunostaining was performed on deparaffinized 3 mm sections, which were subsequently rehydrated, and the endogenous peroxidase activity was quenched via treatment with 0.3% hydrogen peroxide for 20 min.Antigen retrieval was performed by incubating the slides in an autoclave at 120 C for 5 min in antigen retrieval solution (pH 9.0; 415291, Nichirei Bio Sciences, Tokyo, Japan).Sections were incubated overnight at 4 C with rabbit anti-UCP1 antibody (ab10983, Abcam) or anti-TOMM20 antibody (11802-1-AP, Proteintech) at a dilution of 1:2000.UCP1 or TOMM20 signals were amplified with Simple Stain Mouse MAX PO (414311, Nichirei Bio Sciences), and the color was developed using a 3,3'Diaminobenzidine (DAB) substrate.The sections were counterstained with hematoxylin.Adipocytes were quantified with regards to area and number using the ImageRI_Adipocyte_Tools' (https://dev.mri.cnrs.fr/projects/imagej-macros/wiki/Adipocytes_Tool)tool with a minimum size of 40 mm 2 and a maximum size of 40000 mm 2 .More than 110 cells were counted in each individual during quantification.

Tissue temperature measurements
The mouse tissue temperature was measured according to a method described in a previous report. 61Briefly, anesthesia was induced in mice through the intraperitoneal injection of medetomidine hydrochloride (0.3 mg/kg), midazolam (4 mg/kg), and butorphanol (5 mg/kg).The mice were then placed on a self-regulating heating pad (NHP-M30N, Nissinrika, Tokyo, Japan) to maintain their rectal temperature at 37 C. To measure the tissue temperature, a thermocouple RET-3 rectal probe (Physitemp, Clifton, NJ, USA) was implanted into the rectum, whereas needle microprobes MT-29-1 (Physitemp, Clifton, NJ, USA) were inserted into the interscapular BAT and the skeletal muscle on the back.The tissue temperatures were recorded simultaneously using the TC-2000 Meter (Sable Systems International, North Las Vegas, NV, USA).Once the tissue temperature had stabilized for over 5 min, an intraperitoneal infusion of 1 mg/kg NE was administered.Changes in tissue temperature were continuously measured every second for 1 h, and the mean value per minute was calculated.

Transmission electron Microscopy
Differentiated cells expressing WT-hJMJD1A or HY-hJMJD1A were incubated in a solution containing 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 1 h.The cells were then washed three times with 0.1 M cacodylate buffer.Subsequently, cells were treated with 1% OsO 4 in 0.1 M cacodylate buffer for 15 min.After dehydration in a series of ethanol concentrations ranging from 50-100%, the dehydrated cells were embedded in Epon resin.Thin sections were cut using a microtome (EM UC-7; Leica, Wetzlar, Germany), stained with 2% uranyl acetate, 1% lead citrate, 1% lead acetate, and 1% lead nitrate, and images were obtained using a transmission electron microscope (JEM1400, JEOL, Tokyo, Japan).
For immunoprecipitation with anti-JMJD1A antibody (IgG-F0618), 15 2 mg of the protein lysates were incubated with 3 mg of antibody.After a 2-h incubation at 4 C, Proteins G Sepharose 4 Fast Flow beads (GE Healthcare, Chicago, IL, USA) were added and incubated for another 1 h.The beads were washed three times with cell lysis buffer.Immunoprecipitates were subjected to immunoblotting using an anti-JMJD1A antibody (IgG-F0231) 15 and anti-phospho-JMJD1A S265 (#11890-2). 15For immunoblotting, proteins were separated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis, transferred to poly vinylidene di-fluoride (PVDF) membranes, and incubated with antibodies overnight at 4 C. Immunoblots were visualized by chemiluminescence using Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific), and luminescence images were analyzed and quantified using ImageJ software. 46aluation of mtDNA copy number The mitochondrial DNA (mtDNA) copy number was evaluated by determining the ratio of mtDNA to nuclear DNA (nDNA).mtDNA was extracted from scWAT using phenol/chloroform after digestion with proteinase K (150 mg/ml) in DNA lysis buffer (50 mM Tris-HCl [pH 7.9], 20 mM EDTA, 1% SDS, 100 mM NaCl) at 55 C overnight.The relative amounts of mtDNA were determined using quantitative real-time PCR relative to the 18S rRNA level.Mitochondrial coding genes encoding NADH dehydrogenase 1, 2, and 4 (mt-Nd1, mt-Nd2, and mt-Nd4, respectively) were selected to represent the mtDNA copy number.

Quantitative RT-PCR
Total RNA was extracted from cells using a Sepasol reagent (Nacalai Tesque) or from tissues using the ISOGEN reagent (Nippon Gene, Tokyo, Japan), according to the manufacturer's instructions.Extracted RNA (2 mg) was converted to cDNA using the SuperScript III First-Strand Synthesis System and oligo (dT) primers (Thermo Fisher Scientific).Real-time PCR was performed in 384-well plates using SYBR Green PCR buffer and the ABI PRISM 7900HT Sequence Detection System (Thermo Fisher Scientific).All reactions were performed in triplicate.mRNA expression is presented as a fold change relative to the indicated control after normalization to Cyclophilin 15 or Rpl32. 62All primer sequences used in this study are listed in Table S3.

Chromatin immunoprecipitation assay
For ChIP using the anti-H3K27ac and anti-H3K9me2 antibodies, cells were cross-linked with 0.5% formaldehyde for 10 min at RT as previously described. 58,63For ChIP using the anti-JMJD1A antibodies, cells were cross-linked with 1.5mM ethylene glycol bis (succinimidyl succinate) (Thermo Fisher Scientific) for 30 min, followed by a second cross-linking with 0.5% formaldehyde for 10 min at RT, as previously described. 14,19ross-linked cells were homogenized by passing through a 22 G needle 10 times in ice-cold hypotonic buffer (10 mM HEPES-KOH [pH 7.5], 1.5 mM MgCl 2 , 10 mM KCl, 1 mM EDTA, 1 mM EGTA) with protease inhibitors.Nuclear fractions were lysed in cell lysis buffer C (23 mM Tris-HCl [pH 8.0], 3.0 mM EDTA, 0.9% Triton X-100, 134mM NaCl, 0.2% SDS) with protease inhibitors, and fragmented to approximately 500 bp using a Branson SONIFIER 250 (Emerson, St. Louis, MO, USA).The sonicated nuclear fraction was incubated overnight with each antibody conjugated to Dynabeads Protein G (Thermo Fisher Scientific).After washing, decrosslinking, and elution, the immunoprecipitated DNA was purified using the QIAquick PCR purification kit (Qiagen, Venlo, Netherlands), and the concentration was determined using the Qubit double-stranded DNA High Sensitivity Assay Kit (Thermo Fisher Scientific).

RNA-sequencing (RNA-seq)
RNA-seq libraries were prepared using the TruSeq RNA Sample Purification Kit (Illumina, San Diego, CA, USA), according to the manufacturer's protocol.Deep sequencing was performed on a HiSeq 2500 sequencer (Illumina, San Diego, CA, USA) using paired-end 50-base reads.

RNA-seq data analysis
For RNA-seq data analysis, FastQC (developed by the Bioinformatics Group at the Babraham Institute; https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to examine the quality of the sequencing reads from the FASTQ files.Adapter sequences and low quality bases were trimmed using fastp. 47After adapter trimming and filtering low-quality reads, the sequencing reads were aligned to the mm10 mouse genome using STAR. 48Fragments per kilobase of exon per million (FPKM) tables of RNA-seq data were calculated using StringTie 49 with the GENOCODE comprehensive gene annotation set M25 Release.For the integrated analysis of RNA-seq and JMJD1A ChIP-seq data in cultured beige adipocytes 14 (GSE107901), sequencing reads were aligned to the mouse genome mm9, and FPKM was calculated using GFOLD 50 with the GENOCODE comprehensive gene annotation set M1 release.

Figure 3
Figure 3. Continued (F) Immunoblot analysis of UCP1 in tissue homogenates of BAT from Jmjd1a +/+ and Jmjd1a HY/HY mice.Actin was used as a loading control.(G) Immortalized scWAT cells from Jmjd1a +/+ and Jmjd1a HY/HY mice were treated with 10 mM isoproterenol (ISO) for 20 min.Homogenates of these cells were subjected to immunoprecipitation (IP) with anti-mJMJD1A, followed by immunoblotting (IB) with anti-phospho-JMJD1A S265 or anti-mJMJD1A antibodies.(H) Model of adaptation to cold environment through JMJD1A phosphorylation and histone demethylation activity.In the acute phase, JMJD1A contributes to thermogenesis in BAT via phosphorylation and chromatin conformation changes.In the chronic phase, JMJD1A promotes thermogenesis through histone demethylation in scWAT.Jmjd1a HY/HY mice have an impaired adaptive potential to chronic cold, although they retain the adaptability to acute cold.Data are the mean G SEM (B, C, E).Repeated-measures ANOVA with a post-hoc Welch's t test (B) or Welch's t test (C, E) were performed for comparison.*p < 0.05 was considered statistically significant.n.s not significant.See also FigureS3.Uncropped images of the blots are shown in FigureS9.

Figure 5 .
Figure 5. Histone demethylation in the enhancer region of mitochondrial regulator and thermogenesis genes by JMJD1A (A) Genome browser representation for H3K27ac, H3K4me1 in cold or warm beige adipocytes, and JMJD1A and ATAC-seq in im-scWAT cells on days 0 and 8 on Pgc1b or Pgc1a genomic regions.Distal enhancers of Pgc1b and Pgc1a predicted by Enhancer Atlas 2.0 are also presented on the genome browser.(B) Schematic representation of the transgene in NuTRAP::Ucp1-Cre mice.

Figure 5 .Figure 6 .
Figure 5. Continued (C) NuTRAP::Ucp1-Cre mice were exposed to cold at 8 C for 2 weeks before analysis.(D) ChIP-qPCR analysis of H3K9me2 using sorted mCherry + nuclei from beige adipocytes and mCherry À nuclei from the other cells.(E-G) ChIP-qPCR analysis of H3K27ac (E), JMJD1A (F) and H3K9me2 (G) on indicated Pgc1b, Pgc1a, Ucp1 and Prdx5 enhancers during beige adipogenesis.(H) mRNA expression of indicated genes in differentiated im-scWAT cells (day 8) derived from Jmjd1a +/+ or Jmjd1a HY/HY mice.(I) ChIP-qPCR showing the decrease in H3K9me2 levels on Pgc1b and Ucp1 genes during beige adipogenesis, which was perturbed in the im-scWAT from Jmjd1a HY/HY mice.Values of fold enrichment at day 0 of differentiation are set to 1. Data are the mean G SEM (D-I).Representative of two (D-G) or three (H and I) independent experiments.Welch's t test was performed for comparison in (D-I).*p < 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significant.See also Figures S5-S7.

Figure 7 .
Figure 7. Correlation between the expression of JMJD1A and the risk of obesity and metabolic disorder in human cohorts (A-F) Correlation between the expression level of JMJD1A in human adipose tissue and age (A), BMI (B), waist circumference (C), hip circumference (D), serum triglycerides (E) and serum cholesterol (F).Pearson's correlation coefficient was used to determine the correlation between JMJD1A expression level and the indicated parameters.These datasets were obtained from GSE70353.37,38

Figure 8 .
Figure 8. Summary of this study (A) Schematic model of cold-induced mitochondrial biogenesis via epigenetic reprogramming of mitochondrial regulatory genes during scWAT beiging by JMJD1A.(B) Distinct roles of serine phosphorylation (step 1) and histone demethylation (step 2) of JMJD1A in BAT activation and scWAT beiging.(Top) JMJD1A WT mice can activate BAT and induce WAT beiging via histone demethylation independent and dependent mechanisms.(Middle) JMJD1A demethylation inactive mice (HY) are unable to induce mitochondrial biogenesis in scWAT while BAT activation is unaffected.(Bottom) JMJD1A phosphorylation defective mice (SA) are unable to activate either BAT or scWAT beiging.