DDB1 prepares brown adipocytes for cold-induced thermogenesis

Abstract

Brown adipose tissue (BAT) plays a key role in thermogenesis during acute cold exposure. However, it remains unclear how BAT is prepared to rapidly turn on thermogenic genes. Here, we show that damage-specific DNA binding protein 1 (DDB1) mediates the rapid transcription of thermogenic genes upon acute cold exposure. Adipose- or BAT-specific Ddb1 knockout mice show severely whitened BAT and significantly decreased expression of thermogenic genes. These mice develop hypothermia when subjected to acute cold exposure at 4 °C and partial lipodystrophy on a high-fat diet due to deficiency in fatty acid oxidation. Mechanistically, DDB1 binds the promoters of Ucp1 and Ppargc1a and recruits positive transcriptional elongation factor b (P-TEFb) to release promoter-proximally paused RNA polymerase II (Pol II), thereby enabling rapid and synchronized transcription of thermogenic genes upon acute cold exposure. Our findings have thus provided a regulatory mechanism of how BAT is prepared to respond to acute cold challenge.

Graphical abstract

Keywords

brown adipose tissue / thermogenesis / DDB1 / RNA polymerase II / promoter-proximal pausing / UCP1 / PGC1α

Introduction

The increasing prevalence of obesity and associated metabolic diseases worldwide has become a big challenge to public health, and new strategies are in urgent need to treat these diseases. Brown adipose tissue (BAT) is a major organ for non-shivering thermogenesis in mammals [1]. Upon cold exposure, BAT oxidizes metabolic fuels and generates heat via the activity of uncoupling protein 1 (UCP1) [2]. In rodent models, activation of BAT regulates glucose homeostasis and insulin sensitivity by increasing glucose and lipid clearance [3–5]. In many mouse models, enhanced BAT activity leads to resistance to weight gain [6]. Adult humans have metabolic active BAT [7–11]. Activated BAT is associated with accelerated lipid metabolism and improved insulin sensitivity [12–14]. Therefore, targeting BAT represents a promising strategy to treat metabolic diseases [3].

The thermogenesis of BAT is controlled at multiple levels [6]. Upon cold exposure, sympathetic nerves release norepinephrine to activate PKA in brown adipocytes, which phosphorylates CREB and ATF2 to activate the transcription of Ucp1 and Ppargc1a (encoding PGC1α) [15]. PGC1α then cooperates with other transcriptional factors to induce the transcription of thermogenic genes including Ucp1 [6]. Other epigenetic and transcriptional factors also play essential roles in the transcription of thermogenic genes [16, 17]. However, it remains unclear how the thermogenic genes are rapidly transcribed upon cold challenge.

Promoter-proximal pausing of RNA polymerase II (Pol II) is a regulatory mechanism for transcription of immediate early response genes involved in stimulus-responsive pathways [18, 19]. In Pol II pausing, the pre-initiation complex is formed, but Pol II pauses after a synthesis of 20–60 nt of mRNA [20]. In response to various stimuli or developmental cues, the positive transcription elongation factor b (P-TEFb) is released from its inhibitory complex, phosphorylates Pol II at Ser2 of its C-terminal repeat, and activates Pol II, thereby enabling rapid and synchronized expression of the downstream genes [21, 22]. Emerging roles of Pol II pausing have been demonstrated in mammalian embryonic stem cells [23–25], but not much is known in other cell lineages.

Damage-specific DNA binding protein 1 (DDB1) is well recognized as a component of the Cullin4 (CUL4)-RING E3 ubiquitin ligase complex that regulates a variety of physiological events using a subset of WD40 proteins as adaptors [26–28]. Recently, we showed that DDB1 works both in CUL4-dependent and CUL4-independent manner to regulate adipogenesis [29, 30]. In the very early stage of adipogenesis, DDB1 works independently of CUL4 by recruiting P-TEFb to the immediate early response genes to initiate the transcriptional cascade [29]. In the late stage, DDB1 complexes with CUL4 and WDTC1 to ubiquitinate and degrade MED20, a subunit of the Mediator complex, to inhibit adipogenesis [30]. Here, to investigate the role of DDB1 in mature adipocytes, we crossed Ddb1^f/f mice with AdipoQ-Cre or Ucp1-Cre mice and generated adipose- or BAT-specific knockout mice of Ddb1. We characterized the mice and found that depletion of DDB1 in BAT greatly disrupted the thermogenic function. We showed that the thermogenic genes were subjected to regulation by promoter-proximal pausing of Pol II and that DDB1 was required for the release of paused Pol II. Our studies provide a mechanism for how thermogenic genes are rapidly turned on upon acute cold exposure.

Results

DDB1 is required for the maintenance of the brown phenotype of BAT

To explore whether and how DDB1 regulates the physiological function of mature adipocytes, we crossed Ddb1^f/f mice with AdipoQ-Cre mice and generated adipocyte-specific Ddb1 knockout mice, designated as Ddb1-AKO. DDB1 was largely depleted in BAT, inguinal WAT (iWAT), and gonadal WAT (gWAT) of Ddb1-AKO mice (Fig. 1a). As shown in Fig. 1b, the BAT of Ddb1-AKO mice appeared severely whitened compared with that in Ddb1^f/f mice. Hematoxylin and eosin (H&E) analysis revealed that the lipid droplets were dramatically enlarged in the BAT of Ddb1-AKO mice (Fig. 1c). A close examination by electron microscopy revealed that DDB1-deficient brown adipocytes exhibited not only enlarged lipid droplets but also smaller mitochondria that appeared to be darker and contained less cristae (Fig. 1d). Loss of DDB1 also significantly decreased the mitochondrial DNA content in BAT (Fig. 1e). Furthermore, both mRNA and protein levels of BAT marker genes were significantly decreased in the BAT of Ddb1-AKO mice (Fig. 1f and g).

We have also examined the adipogenesis markers in the BAT of Ddb1^f/f and Ddb1-AKO mice. As shown in Supplementary Fig. S1a, there was not much difference in the protein levels of PERILIPIN, PPARγ, C/EBPα, and CD36. The DNA content of BAT was also not different between the two strains (Supplementary Fig. S1b). These data indicate that the phenotype of BAT in Ddb1-AKO mice is not due to defects in adipogenesis.

To ensure that the whitening of BAT in the Ddb1-AKO was not secondary to changes in WAT, we crossed Ddb1^f/f mice with Ucp1-Cre mice and generated BAT-specific Ddb1 knockout mice (Ddb1-BKO) (Fig. 1h). Very similar to Ddb1-AKO mice, Ddb1-BKO mice showed whitened BAT with enlarged lipid droplets and less mitochondrial content (Fig. 1i–l). The expression levels of BAT-specific genes were largely reduced in the BAT of Ddb1-BKO mice (Fig. 1m and n). These results indicate that DDB1 maintains the brown phenotype of BAT in a cell-autonomous manner.

To further evaluate the effect of DDB1 on BAT, we extracted mRNA from BAT of Ddb1-AKO and their littermate controls and performed RNA sequencing (RNA-seq). Among the 12,290 genes analyzed, 1827 genes were downregulated in the Ddb1-AKO mice by more than 1.5-fold, including the BAT-specific genes, such as Cox8b, Cidea, and Ucp1, and the mitochondrial genes, such as mt-Co1 and mt-Cytb (Fig. 2a). Gene ontology analysis of the suppressed genes in the Ddb1-AKO mice revealed a strong enrichment of genes involved in mitochondrial oxidative phosphorylation, tricarboxylic acid (TCA) cycle, cellular respiration, electron transport chain, respiratory chain, and mitochondrial protein complex (Fig. 2b). A close examination revealed that almost all the genes of mitochondrial complex I–V were downregulated in the BAT of Ddb1-AKO mice (Fig. 2c), which was further confirmed by quantitative real-time polymerase chain reaction (qRT-PCR) analysis (Supplementary Fig. S1d). In contrast, glycolysis genes did not show much difference between Ddb1^f/f and Ddb1-AKO mice (Supplementary Fig. S1c). Furthermore, some of the key transcriptional regulators of BAT, including Ebf2, Cebpb, Ppara, Atf2, Esrra, and Hdac3 [6, 31, 32], were significantly decreased in both -Ddb1-AKO and Ddb1-BKO mice (Fig. 2d). The mRNA level of Adrb3 was significantly increased in DDB1-deficient BAT (Fig. 2d), implying a compensatory effect. These data suggest that DDB1 might be a master transcriptional regulator of thermogenic genes.

DDB1 is required for cold-induced thermogenesis

We then directly tested the role of DDB1 in the thermogenic functions of BAT. First, we examined the body temperature of pups of Ddb1-AKO and control littermates on postnatal day 3. As shown in Fig. 3a and b, the skin temperature of the Ddb1-AKO mice (32.7 ± 0.1 °C) was significantly lower than that (34.3 ± 0.2 °C) of the control littermates.

We then studied the role of DDB1 in adult mice. When Ddb1^f/f and Ddb1-AKO mice were housed at 22 °C, both strains could maintain their core temperatures around 37 °C (Fig. 3c). When they were switched to 4 °C, the Ddb1^f/f mice could maintain their core temperatures around 35 °C; however, the Ddb1-AKO mice failed to do so, and their core temperatures decreased to 16.4 ± 1.9 °C after 4 h of cold exposure (Fig. 3c). At this point, the Ddb1-AKO mice became moribund and we had to stop the experiment. After the 4-h cold exposure, the triglyceride content in BAT was higher (Fig. 3d), but the blood glucose level was significantly lower (Fig. 3e) in the Ddb1-AKO mice, indicating an impaired utilization of fatty acids and increased utilization of glucose.

To further test the function of BAT in response to cold challenge, we treated mice with CL316,243, an agonist of the β3-adrenergic receptor, and monitored oxygen consumption in a metabolic cage. While Ddb1^f/f mice exhibited increased oxygen consumption after CL316,243 injection, Ddb1-AKO mice barely showed any response (Fig. 3f), which further confirms that lack of DDB1 results in dysfunctional BAT.

We then isolated primary stromal vascular fractions (SVFs) from BAT of Rosa-Cre^ERT2–Ddb1^f/f mice, and induced differentiation into brown adipocytes, followed by treatment with 4-hydroxytamoxifen (4-OHT) to induce deletion of Ddb1 and analysis of oxygen consumption rate (OCR). As shown in Fig. 3g and h, the OCR in DDB1-deficient cells was significantly lower than that in control cells under both basal and forskolin-stimulated conditions.

We next performed the same experiments in Ddb1-BKO mice. Consistent with the results obtained from Ddb1-AKO mice, both pups and adults of Ddb1-BKO mice had defects in thermogenesis in response to cold challenge (Supplementary Fig. S2a–f), which further confirms the critical role of DDB1 in cold-induced thermogenesis.

We have also examined the role of DDB1 in the browning of iWAT. As shown in Supplementary Fig. S3a and b, after 10 consecutive days of administration of CL316,243, both mRNA and protein levels of Ucp1, Ppargc1a, and the mitochondrial genes were dramatically upregulated in the iWAT of Ddb1^f/f mice; however, such effect was largely blocked in Ddb1-AKO mice. H&E staining and immunostaining using anti-UCP1 antibody revealed that mice lacking DDB1 showed less extent of browning with dramatically reduced the expression of UCP1 (Supplementary Fig. S3c).

Loss of DDB1 in BAT disrupts whole-body lipid metabolism

We then sought to examine the effect of whitened BAT on whole-body metabolic homeostasis. We first subjected Ddb1-AKO mice and their control littermates to chow and high-fat diet (HFD) for 16 weeks. While Ddb1^f/f and Ddb1-AKO mice showed no difference in their body weight on chow diet, the body weight of HFD-fed Ddb1-AKO mice was significantly lower than Ddb1^f/f mice starting from week 10 (Fig. 4a). However, the HFD-fed Ddb1-AKO mice showed decreased capability to clear glucose (Fig. 4b), and they were insulin resistant (Fig. 4c). On week 16 of HFD feeding, Ddb1-AKO mice showed higher liver weight and decreased weights of iWAT and gWAT (Fig. 4d). The Ddb1-AKO mice showed higher plasma levels of insulin (Fig. 4e) and free fatty acids (Fig. 4f) and significantly elevated liver triglyceride content (Fig. 4g), resembling the phenotype of partial lipodystrophy.

To further characterize these mice, we subjected them to metabolic cage analysis. As shown in Fig. 4h and i, the respiratory exchange ratio of Ddb1-AKO mice during daytime was significantly higher than that in Ddb1^f/f mice under both chow- and HFD-fed conditions, indicating that Ddb1-AKO mice have a defect in utilizing fatty acids to maintain body temperature. These results were supported by higher levels of plasma-free fatty acids in Ddb1-AKO mice (Fig. 4f) and further explained the ectopic lipid storage in the liver (Fig. 4g).

We have also performed the same experiments in Ddb1-BKO mice and their control littermates. Similarly, Ddb1-BKO mice also gained less weight on HFD but developed insulin resistance and partial lipodystrophy (Supplementary Fig. S4a–g), confirming that loss of DDB1 in BAT disrupts whole-body lipid metabolism.

DDB1 binds promoters of early response thermogenic genes to control their transcription

We then went on to interrogate the underlying mechanism of how DDB1 controls thermogenesis in BAT. DDB1 typically functions as a component of the CUL4 E3 ligase complex [27], but we have previously shown that DDB1 can also act in a CUL4-independent manner [29]. To examine whether the function of DDB1 in regulating thermogenesis was dependent on the CUL4 E3 ligase complex, we crossed Cul4a^f/f and Cul4b^f/y mice with AdipoQ-Cre mice to generate adipose tissue-specific knockout of Cul4a (Cul4a-AKO) and Cul4b (Cul4b-AKO) mice, respectively. Supplementary Fig. S5a–c and e–g shows that neither Cul4a-AKO nor Cul4b-AKO mice showed any defect in BAT morphology or expression of BAT-specific genes. And these mice showed no difference from littermate controls in response to acute cold exposure (Supplementary Fig. S5d and h). Therefore, DDB1 might function in a CUL4-independent manner to regulate the thermogenesis of BAT.

We then explored whether DDB1 would directly regulate the transcription of thermogenic genes. We exposed Ddb1^f/f and Ddb1-BKO mice to 4 °C for 4 h, extracted mRNAs from BAT, and subjected them to RNA-Seq analysis. In Ddb1^f/f mice, 810 genes were early response genes that were upregulated more than 2-fold by cold exposure (Fig. 5a). Among them, the fold induction of 415 genes in Ddb1-BKO mice was less than 50% of that in Ddb1^f/f mice, and they were designated as DDB1 downstream genes (Fig. 5a). Gene ontology analysis revealed that these genes were involved in fatty acid metabolism, regulation of cytokine production, fat cell differentiation, glycerolipid metabolism, fatty acid elongation, and positive regulation of transcription from Pol II promoter (Fig. 5b). Fig. 5c shows the heat map of some of the representative genes. To further confirm the results, we performed a qRT-PCR analysis and found that cold induction of Ucp1, Ppargc1a, Dio2, Elovl3, Fgf21, and Ffar4 was largely blunted in the BAT of both Ddb1-AKO and Ddb1-BKO mice (Fig. 5d).

To directly test the effect of DDB1 on the transcription of its downstream genes, we isolated primary SVFs from BAT of Ddb1^f/f and Ddb1-AKO mice, induced differentiation into mature adipocytes, and treated with forskolin. As shown in Fig. 6a, Ucp1 and Ppargc1a were dramatically induced by forskolin in Ddb1^f/f cells, but such induction was significantly reduced in DDB1-deficient cells.

We then performed ChIP-Seq to visualize the binding of DDB1 on its downstream genes. Indeed, DDB1 bound the promoters of its downstream genes, and its binding intensity was not changed by forskolin treatment (Fig. 6b and c). Fig. 6d shows that DDB1 directly bound the promoter of Ucp1. De novo motif search revealed that DDB1-binding sites overlapped with that of bZIP_CREB family of transcription factors (Fig. 6e), which has been shown to play a critical role in thermogenesis [6, 33]. Immunoprecipitation assay revealed that endogenous DDB1 interacted with CREB in the presence or absence of forskolin (Fig. 6f).

Furthermore, when A-CREB, a dominant-negative form of CREB, was overexpressed in brown adipocytes, it indeed significantly decreased forskolin-induced expression of Ucp1 and Ppargc1a (Fig. 6g). We then performed dual luciferase reporter assay and found that depletion of DDB1 significantly decreased the CRE-reporter activity in response to forskolin (Fig. 6h), indicating that DDB1 directly regulates the transcriptional activity of CREB.

DDB1 facilitates the release of paused Pol II on the thermogenic genes

We next sought to know how DDB1 controls the transcription of CREB downstream thermogenic genes. As we have previously shown that DDB1 recruits P-TEFb to turn on the immediate early response genes in adipogenesis [29], we examined whether DDB1 would function in a similar way to control the transcription of the early response thermogenic genes. We first performed immunoprecipitation using the anti-CREB antibody in control and DDB1-deficient brown adipocytes. Fig. 7a shows that immunoprecipitation of CREB pulled down Pol II and the two subunits of P-TEFb, CDK9 and Cyclin T1, and treatment with forskolin enhanced their interaction. However, when DDB1 was depleted from these cells, the interaction between CREB and Pol II or P-TEFb was largely abolished (Fig. 7a), indicating a critical role of DDB1 in recruiting P-TEFb to CREB downstream genes.

To directly test whether the early response thermogenic genes were regulated by Pol II pausing and to study the role of DDB1 in the process, we performed ChIP-Seq analysis of Pol II and Pol II-S2P in BAT of Ddb1^f/f and Ddb1-BKO mice at 22 °C or 4 °C. Pol II-S2P is the transcriptionally active form of Pol II that is phosphorylated by P-TEFb at Ser2 in its C-terminal repeats [34]. At 22 °C, Pol II bound the proximal promoter of Ucp1 in the BAT of both Ddb1^f/f and Ddb1-BKO mice but showed a higher binding intensity on the gene body in Ddb1^f/f mice (Fig. 7b). Consistently, the binding of Pol II-S2P on the gene body of Ucp1 is much higher in Ddb1^f/f mice (Fig. 7b). When the mice were subjected to acute cold exposure, the binding of both Pol II and Pol II-S2P dramatically increased in Ddb1^f/f mice but not in Ddb1-BKO mice (Fig. 7b). Similar observations were made on the binding of Pol II and Pol II-S2P on Ppargc1a (Supplementary Fig. S6a). We have also performed a global analysis of Pol II and Pol II-S2P binding on the 415 DDB1 downstream genes and found that DDB1 was indeed required for the binding of Pol II and Pol II-S2P on the proximal promoters of the early response thermogenic genes (Fig. 7c–f; Supplementary Fig. S6b–e). These results indicate that the early response thermogenic genes are subject to regulation by Pol II pausing and that DDB1 plays an essential role in the release of paused Pol II.

To further confirm that the early response thermogenic genes are regulated by promoter-proximal pausing of Pol II, we treated SVF-derived brown adipocytes with JQ1, an inhibitor of BRD4, to block the recruitment of P-TEFb and release of paused Pol II [35, 36]. Fig. 7g shows that treatment with JQ1 significantly decreased forskolin-induced transcription of Ucp1 and Ppargc1a. We then pretreated wild type (WT) mice with JQ1 followed by acute cold exposure at 4 °C and found that JQ1-treated mice showed significantly lower core temperatures (Fig. 7h and i). These mice also showed decreased blood glucose levels and increased BAT triglyceride content after cold exposure (Fig. 7j). Cold-induced transcription of Ucp1 and Ppargc1a was significantly reduced by JQ1 treatment (Fig. 7k). These data provided further evidence that cold-induced thermogenesis is regulated by Pol II pausing.

To summarize our work, we propose the following working model (Fig. 8). In WT brown adipocytes, DDB1 binds the proximal promoters of early response thermogenic genes including Ucp1 and Ppargc1a. The pre-initiation complex is formed on these promoters, but Pol II is paused under normal conditions. Upon cold exposure, P-TEFb is released from its inhibitory complex and recruited to the proximal promoters of the thermogenic genes by DDB1. P-TEFb then phosphorylates and activates Pol II, resulting in productive transcriptional elongation of the thermogenic genes to maintain body temperature. In DDB1-deficient brown adipocytes, P-TEFb cannot reach the proximal promoters of the thermogenic genes and fails to turn on the transcription of these genes. BAT thus shows a defect in fatty acid oxidation and fails to produce enough heat to main body temperature.

Discussion

Ever since the discovery of metabolically active BAT in adult humans, BAT has emerged as a potential therapeutic target to treat obesity and related metabolic disorders [3]. Many transcriptional factors and chromatin remodeling factors have been identified to regulate thermogenesis [16, 17], but it remains unclear how the cold signal is integrated with the transcription of thermogenic genes. Here, we show that the early response thermogenic genes are regulated by promoter-proximal pausing of Pol II and that DDB1 plays an essential role in releasing paused Pol II and turning on the transcription of these genes.

The transcription of the thermogenic genes is regulated by various factors [16]. Traditionally, the rate-limiting step of transcription is the formation of the pre-initiation complex. Indeed, histone modifiers such as EHMT1, JMJD1A, and HDAC3 [32, 37, 38] and transcriptional factors such as CREB, ATF2, PRDM16, EBF2, and ZFP516 have also been reported to be required for activation of the thermogenic genes [15, 31, 39, 40]. However, accumulating evidence shows that many immediate early response genes are subjected to regulation by promoter-proximal pausing of Pol II, in which the pre-initiation complex has been formed before stimulation [18]. Here, we show that promoter-proximal pausing of Pol II plays a critical role in cold-induced thermogenesis. Through RNA-Seq and ChIP-Seq analysis, we demonstrate that the transcription of the early response thermogenic genes including Ucp1 and Ppargc1a is controlled by the release of paused Pol II upon cold exposure. Furthermore, blocking the release of Pol II by JQ1 results in cold-induced hypothermia in mice.

DDB1 is typically recognized as a component of the CUL4-E3 complex [27]. We have previously reported a CUL4-independent function in adipogenesis, namely that DDB1 complexes with P-TEFb to initiate the transcriptional cascade of adipogenesis [29]. Here, we identify another CUL4-independent function of DDB1, which is to regulate thermogenesis. The adipose- or BAT-specific Ddb1 knockout mice exhibited whitened BAT and developed hypothermia when subjected to acute cold exposure. In contrast, knocking out either Cul4a or Cul4b in adipose tissues has no effect on thermogenesis. Based on the CUL4-independent function of DDB1 in regulating adipogenesis and thermogenesis, we speculate that DDB1 might work independently of CUL4 to regulate other physiological events.

It is notable that disrupting BAT function by depleting DDB1 did not lead to obesity as expected, although BAT activation is associated with lower body mass [11]. Instead, both Ddb1-AKO and Ddb1-BKO mice showed decreased body weight on HFD, but they were less healthy because both strains showed decreased glucose clearance rate and developed insulin resistance, resembling a phenotype of partial lipodystrophy. These mice also showed a decreased capability for oxidizing fatty acids, elevated plasma levels of free fatty acids, and ectopic lipid storage in the liver. Usually, to maintain body temperature, WAT will upregulate lipolysis to release fatty acids to provide fuels for BAT. However, the Ddb1-AKO and Ddb1-BKO mice had dramatically decreased expression of the thermogenic genes, and they could not efficiently utilize fatty acids. Consequently, WAT would increase the lipolysis rate, resulting in increased plasma-free fatty acids and decreased size of WAT. The excessive plasma free fatty acids would then be ectopically stored in the liver, leading to fatty liver and insulin resistance. In previous studies, Ucp1^−/− mice are not obese [2], and mice lacking PGC1α in adipose tissues develop insulin resistance without extra weight gain on HFD [41]. Recent human clinical studies show that BAT activation contributes to a small amount of energy metabolism that is unlikely to cause weight loss but improves glucose metabolism [42]. Taken together, BAT activation might be more effective to improve glucose homeostasis than weight loss.

It is also notable that the transcription of Ucp1 at 22 °C is already lower in the BAT of Ddb1-AKO or Ddb1-BKO mice than that in the control mice. Ideally, to study the effect of DDB1 on the transcription of Ucp1, it would be better to start from the same basal level of Ucp1, especially using an inducible knockout system. As we have reported before that inducible knockout of Ddb1 in Rosa-Cre^ERT2–Ddb1^f/f mice causes lethality [29], an adipose-specific inducible knockout model might be more suitable in future studies.

Materials and Methods

Materials

We obtained CL316,243, dexamethasone, isobutylmethylxanthine (IBMX), forskolin, isoproterenol, pioglitazone, bovine insulin, urea, sodium dodecyl sulfate (SDS), dithiothreitol (DTT), dimethylsulfoxide (DMSO), and Triton X-100 from Sigma-Aldrich; Dulbecco’s modified Eagle’s medium (DMEM) with low (1 g/L) or high (4.5 g/L) glucose, fetal and neonatal bovine serum, blasticidin, and puromycin from Thermo Fisher Scientific; donkey anti-rabbit IgG or anti-mouse IgG conjugated to horseradish peroxidase from Jackson Immuno Research; protease inhibitor cocktail from Roche Applied Science; and all other chemicals from local suppliers unless otherwise specified.

Culture, immortalization, and differentiation of BAT adipocytes

For immortalization of BAT preadipocytes, SVFs were isolated from interscapular BAT in newborn Ddb1^f/f, Rosa-Cre^ERT2–Ddb1^f/f, or AdipoQ-Cre–Ddb1^f/f mice, cultured and immortalized as previously described [43, 44]. Briefly, cells were cultured in medium A (DMEM high glucose, 20 mM 4-(2-Hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES) pH 7.4, 10% (v/v) fetal calf serum (FCS), 100 U/ml penicillin, and 100 mg/ml streptomycin) at 37 °C in an atmosphere of 8.8% CO₂. To differentiate into mature brown adipocytes, cells were cultured to 100% confluence and maintained in medium A for another 2 days. On day 0 of differentiation, cells were treated with medium A containing 1 nM T3, 0.1 μg/ml insulin, 0.125 mM indomethacin, 5 μM dexamethasone, and 0.5 mM IBMX. On days 2, 4, and 6, the medium was changed to medium A containing 1 nM T3 and 0.1 μg/ml insulin. On day 8, fully differentiated brown adipocytes were achieved. To induce deletion of Ddb1 in SVF-derived brown adipocytes of Rosa-Cre^ERT2–Ddb1^f/f mice, cells were treated with 8 μM 4-OHT for 4 days.

Dual luciferase reporter assay

The CRE reporter was constructed by inserting a 4 × CRE sequence into a pGL3-basic vector (Promega), designated as CRE-Fluc. On day 4 of differentiation, cells were treated with 4-OHT (8 μM) to induce deletion of Ddb1. On day 6, BAT adipocytes were replated at 4 × 10⁴ cells per well in a 12-well plate. On day 7, cells were transfected with 1.5 μg CRE-Fluc and 0.01 μg Renilla luciferase (Rluc) expressing vector with X-tremeGENE HP (Roche). On day 9, cells were treated with forskolin (10 μM) for 8 h before harvest for dual luciferase reporter assay following the manufacturer’s instructions (Promega, E1960). All the measurements were done in triplicates.

Retrovirus production and infection

For retrovirus production, HEK293T cells were set up on day 0 at 2.5 × 10⁵ cells per 60-mm dish. On day 2, A-CREB/pMSCV-IRES-GFP II (pMIG II, Addgene, 52107) was co-transfected with pCL-Eco (Addgene, 12371) at 1:1. On day 3, fresh medium was changed. On days 4 and 5, media containing retrovirus particles were collected, centrifuged at 1,500 g for 5 min, aliquoted, and stored at −80 °C until use. For retroviral infections, BAT adipocytes were infected with retrovirus in a medium containing 8–10 mg/ml polybrene on days 4 and 6 of differentiation.

Mice and diets

All mice were housed in colony cages at 22 °C with 12-h light/12-h dark cycles. The dark cycle began at 7 p.m. All animal studies were performed with the approval of the Institutional Animal Care and Research Advisory Committee at Fudan University and Xiamen University.

Ddb1^f/f mice were generous gifts from Dr. Yong Cang at Shanghai Tech University [26]. Cul4a^f/f mice were generous gifts from Dr. Nengming Xiao at Xiamen University [29]. Cul4b^f/y mice were generous gifts from Dr. Yaoqin Gong at Shandong University [45]. These mice were bred with AdipoQ-Cre [46] or Ucp1-Cre [47] transgenic mice to generate adipose- or BAT-specific knockout mice.

The chow diet (Xietong Organism, Nanjing, China) contains 12% of calories from fat, 67.4% from carbohydrates, and 20.6% from protein. The HFD (Research Diet, D12492) contained 60% calories from fat, 20% calories from carbohydrate, and 20% calories from protein.

Histology

Adipose tissues were fixed for 20–48 h in 4% (wt/vol) paraformaldehyde in phosphate buffered saline (PBS). The fixed tissues were embedded in paraffin and sectioned at 5 μm. Slides were then counterstained with H&E. Immunohistochemistry was performed using anti-UCP1 (Abcam, ab10983).

Electron microscopy

Mice were perfused with 25 ml of a solution containing 4% (wt/vol) paraformaldehyde, 1% (wt/vol) glutaraldehyde, and 250 mM sucrose in 0.1 M cacodylate buffer (pH 7.4), and BAT was isolated and fixed as previously described [48]. The samples were postfixed with 1% (wt/vol) OsO₄, embedded, and sectioned. Specimens were visualized on a JFC1600 transmission electron microscope.

Quantification of mitochondrial DNA

BAT was isolated and digested at 55 °C overnight in 0.1 M Tris (pH 8.0), 0.2 M NaCl, 5 mM ethylenediaminetetraacetic acid (EDTA), 0.4% SDS, and 0.2 mg/ml protease K. DNA was phenol-chloroform extracted, precipitated with isopropanol, dried up, and resuspended in 10 mM Tris (pH 8.0) and 1 mM EDTA. DNA was then subjected to qRT-PCR, and the ratio of Mt-Co1 and Ndufv1 was used to analyze the relative mitochondrial DNA content [49].

Gene expression analysis

Total RNA was isolated, and qRT-PCR measurements were performed as described [50]. The primers are listed in Supplementary Table S1. All reactions were done in triplicates. The relative amount of each mRNA was calculated by using the comparative threshold cycle (C_T) method. Cyclophilin or 36B4 mRNA was used as the invariant control.

RNA-Seq analysis of global gene expression profiling was conducted and analyzed as previously described [29]. Genes with fragments per kilobase per million (FPKM) no less than 1 were included in the analysis.

Thermo imaging of skin temperature

Mice were genotyped on postnatal day 2. On postnatal day 3 before the onset of hair growth, three pairs of pulps were transferred to a 6-well plate, and skin temperatures were measured by a thermal imaging camera (T300 InfraRed Camera; FLIR Systems).

Acute cold exposure at 4 °C

Mice were maintained at 22 °C and singly housed a week before the experiment. On the day of experiment, mice were briefly fasted for 4 h and then switched to a 4 °C cold room. Core temperature was measured at 0, 0.5, 1, 2, 3, and 4 h after cold exposure using a digital thermometer with a rectal thermocouple probe (Physitemp, Model BAT-12).

Metabolic cage study

Metabolic cage analysis was performed in a home-cage system Phenomaster (TSE Systems). Mice were singly housed a week before the experiments and allowed to get acclimated to the metabolic cages for 2 days. Mice were then monitored for 4 days on food intake, body weight, oxygen consumption, carbon dioxide production, and locomotor activities.

To test the effect of CL316,243 on oxygen consumption, each mouse received an intraperitoneal injection of CL316,243 at 10 mg/kg. Oxygen consumption was monitored at 10 min intervals from 30 min before to 3 h after the injection.

Oxygen consumption analysis of brown adipocytes

OCR was determined at 37 °C using the OROBOROS Oxygraph-2K module (OROBOROS Instruments GmbH). Differentiated brown adipocytes in a 6-cm dish were pretreated with or without 10 μM forskolin for 1 h before the assay. During OCR measurement, cells were subsequentially treated with oligomycin (0.25 μM), phenylhydrazone (FCCP, 5 μM), and Rotenone (0.1 μM). OCR was normalized to the protein content.

Oral glucose tolerance test and insulin tolerance test

Oral glucose tolerance test and insulin tolerance test were performed as previously described [29]. Briefly, for the oral glucose tolerance test, mice were fasted for 16 h (from 5 p.m. to 9 a.m.) and orally gavaged with 1 or 2 mg/kg glucose as indicated in the figure legends. For insulin tolerance test, mice were briefly fasted for 6 h (from 8 a.m. to 2 p.m.) and intraperitoneally injected with insulin 0.5 or 1 U/kg body weight as indicated in the figure legends. Blood was collected from the tail vein at 0, 15, 30, 60, 90, and 120 min after gavage with glucose or injection with insulin, and blood glucose was measured by a Bayer Contour Glucometer.

Metabolic parameters

Plasma insulin was measured using a commercial kit (EZassay, MS200). Liver triglyceride and cholesterol were extracted as previously described [51] and measured by commercial kits from Wako Chemicals. To measure triglyceride content in BAT, BAT was homogenized in PBS containing 0.5% SDS, gradually heated up to 95 °C, and kept at 95 °C for another 5 min. The cooled samples were subjected to triglyceride measurement using a commercial kit from Wako Chemicals. The content of triglyceride was normalized to protein content, which was quantified by a Pierce BCA kit (Thermo Fisher Scientific).

Chromosome immunoprecipitation sequencing

Brown adipocytes and BAT tissue were used for chromosome immunoprecipitation sequencing (ChIP-Seq) analysis. For ChIP with anti-DDB1 antibodies, cells were fixed with 1% glutaraldehyde for 15 min at room temperature. For ChIP with anti-Pol II and anti-Pol II-S2P antibodies, BAT was collected from three mice in each group, grinded in liquid nitrogen, and fixed with 1% formaldehyde for 15 min at 37 °C. Fixation was stopped by adding glycine to a final concentration of 0.14 M and incubating at RT for 10 min, followed by two washes with cold PBS. ChIP, ChIP-Seq library preparation, and data analysis were performed as previously described [29].

Western blot

Total proteins of adipose tissues were extracted, and Western blot was carried out as previously described. The following antibodies were used: anti-DDB1 (1:50,000, Abcam, ab109027), anti-UCP1 (1:1000, Abcam, ab10983), anti-PPARγ (1:500, Santa Cruz, sc-7273), anti-C/EBPα (1:1000, CST, 8178s), anti-CD36 (1:1000, Sino Biological Inc., 80263-T48), anti-PERILIPIN (1:1000, CST, 9349s), anti-CREB (1:1000, CST, 9197S), anti-pCREB (1:1000, CST, 9198S), anti-COXII (1:1000, Proteintech, 55070-1-AP), anti-COXIV (1:1000, Proteintech, 11424-1-AP), anti-PGC1a (1:1000, Proteintech, 66369-1-Ig), anti-CyclinT1 (1:1000, CST, 81464), anti-CDK9 (1:1000, Santa Cruz, sc-13130), anti-RNA Polymerase II (1:1000, Bethyl, A300-653A), anti-GAPDH (1:5000, Proteintech, 60004-1), and anti-Flag M2 (1:1,000, Sigma, F1804). Membranes were developed in a ChemStudio imaging system (Analytik Jena AG).

Quantification and statistical analysis

All the statistical analysis was performed using the Student’s two-tailed paired t-test. The value represents mean ± SEM. Statistical details of all experiments can be found in the figure legends, including the exact number of cell samples or mice. Asterisks (*) indicate the levels of statistical significance. *P < .05; **P < .01; ***P < .001. No data were excluded from any of the experiments.

References

[1] Cannon B, Nedergaard J. Brown adipose tissue: function and physiological significance. Physiol Rev 2004; 84: 277–359.

[2] Enerback S, Jacobsson A, Simpson EMet al. Mice lacking mitochondrial uncoupling protein are cold-sensitive but not obese. Nature 1997; 387: 90–4.

[3] Peirce V, Vidal-Puig A. Regulation of glucose homoeostasis by brown adipose tissue. Lancet Diabetes Endocrinol 2013; 1: 353–60.

[4] Stanford KI, Middelbeek RJ, Townsend KLet al. Brown adipose tissue regulates glucose homeostasis and insulin sensitivity. J Clin Invest 2013; 123: 215–23.

[5] Shamsi F, Wang CH, Tseng YH. The evolving view of thermogenic adipocytes – ontogeny, niche and function. Nat Rev Endocrinol 2021; 17: 726–44.

[6] Harms M, Seale P. Brown and beige fat: development, function and therapeutic potential. Nat Med 2013; 19: 1252–63.

[7] Cypess AM, Lehman S, Williams Get al. Identification and importance of brown adipose tissue in adult humans. N Engl J Med 2009; 360: 1509–17.

[8] Nedergaard J, Bengtsson T, Cannon B. Unexpected evidence for active brown adipose tissue in adult humans. Am J Physiol Endocrinol Metab 2007; 293: E444–52.

[9] Saito M, Okamatsu-Ogura Y, Matsushita Met al. High incidence of metabolically active brown adipose tissue in healthy adult humans: effects of cold exposure and adiposity. Diabetes 2009; 58: 1526–31.

[10] Virtanen KA, Lidell ME, Orava Jet al. Functional brown adipose tissue in healthy adults. N Engl J Med 2009; 360: 1518–25.

[11] van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NMet al. Cold-activated brown adipose tissue in healthy men. N Engl J Med 2009; 360: 1500–8.

[12] Chondronikola M, Volpi E, Borsheim Eet al. Brown adipose tissue improves whole-body glucose homeostasis and insulin sensitivity in humans. Diabetes 2014; 63: 4089–99.

[13] Chondronikola M, Volpi E, Borsheim Eet al. Brown adipose tissue activation is linked to distinct systemic effects on lipid metabolism in humans. Cell Metab 2016; 23: 1200–6.

[14] Lee P, Smith S, Linderman Jet al. Temperature-acclimated brown adipose tissue modulates insulin sensitivity in humans. Diabetes 2014; 63: 3686–98.

[15] Contreras C, Nogueiras R, Dieguez Cet al. Hypothalamus and thermogenesis: heating the BAT, browning the WAT. Mol Cell Endocrinol 2016; 438: 107–15.

[16] Inagaki T, Sakai J, Kajimura S. Transcriptional and epigenetic control of brown and beige adipose cell fate and function. Nat Rev Mol Cell Biol 2016; 17: 480–95.

[17] Villarroya F, Peyrou M, Giralt M. Transcriptional regulation of the uncoupling protein-1 gene. Biochimie 2017; 134: 86–92.

[18] Adelman K, Lis JT. Promoter-proximal pausing of RNA polymerase II: emerging roles in metazoans. Nat Rev Genet 2012; 13: 720–31.

[19] Levine M. Paused RNA polymerase II as a developmental checkpoint. Cell 2011; 145: 502–11.

[20] Jonkers I, Lis J. Getting up to speed with transcription elongation by RNA polymerase II. Nat Rev Mol Cell Biol 2015; 16: 167–77.

[21] Peterlin BM, Price DH. Controlling the elongation phase of transcription with P-TEFb. Mol Cell 2006; 23: 297–305.

[22] Zhou Q, Li T, Price DH. RNA polymerase II elongation control. Annu Rev Biochem 2012; 81: 119–43.

[23] Robinson C, Lowe M, Schwartz Aet al. Mechanisms and developmental roles of promoter-proximal pausing of RNA polymerase II. J Stem Cell Res Ther 2016; 6: 330.

[24] Williams LH, Fromm G, Gokey NGet al. Pausing of RNA polymerase II regulates mammalian developmental potential through control of signaling networks. Mol Cell 2015; 58: 311–22.

[25] Min IM, Waterfall JJ, Core LJet al. Regulating RNA polymerase pausing and transcription elongation in embryonic stem cells. Genes Dev 2011; 25: 742–54.

[26] Cang Y, Zhang J, Nicholas SAet al. Deletion of DDB1 in mouse brain and lens leads to p53-dependent elimination of proliferating cells. Cell 2006; 127: 929–40.

[27] Iovine B, Iannella ML, Bevilacqua MA. Damage-specific DNA binding protein 1 (DDB1): a protein with a wide range of functions. Int J Biochem Cell Biol 2011; 43: 1664–7.

[28] Jackson S, Xiong Y. CRL4s: the CUL4-RING E3 ubiquitin ligases. Trends Biochem Sci 2009; 34: 562–70.

[29] Wang X, Wang HY, Hu GSet al. DDB1 binds histone reader BRWD3 to activate the transcriptional cascade in adipogenesis and promote onset of obesity. Cell Rep 2021; 35: 109281.

[30] Tang WS, Weng L, Wang Xet al. The Mediator subunit MED20 organizes the early adipogenic complex to promote development of adipose tissues and diet-induced obesity. Cell Rep 2021; 36: 109314.

[31] Shapira SN, Lim HW, Rajakumari Set al. EBF2 transcriptionally regulates brown adipogenesis via the histone reader DPF3 and the BAF chromatin remodeling complex. Genes Dev 2017; 31: 660–73.

[32] Emmett MJ, Lim HW, Jager Jet al. Histone deacetylase 3 prepares brown adipose tissue for acute thermogenic challenge. Nature 2017; 546: 544–8.

[33] Hilton C, Karpe F, Pinnick KE. Role of developmental transcription factors in white, brown and beige adipose tissues. Biochim Biophys Acta 2015; 1851: 686–96.

[34] Bowman EA, Kelly WG. RNA polymerase II transcription elongation and Pol II CTD Ser2 phosphorylation: a tail of two kinases. Nucleus 2014; 5: 224–36.

[35] Filippakopoulos P, Qi J, Picaud Set al. Selective inhibition of BET bromodomains. Nature 2010; 468: 1067–73.

[36] Liu W, Ma Q, Wong Ket al. Brd4 and JMJD6-associated anti-pause enhancers in regulation of transcriptional pause release. Cell 2013; 155: 1581–95.

[37] Abe Y, Rozqie R, Matsumura Yet al. JMJD1A is a signal-sensing scaffold that regulates acute chromatin dynamics via SWI/SNF association for thermogenesis. Nat Commun 2015; 6: 7052.

[38] Ohno H, Shinoda K, Ohyama Ket al. EHMT1 controls brown adipose cell fate and thermogenesis through the PRDM16 complex. Nature 2013; 504: 163–7.

[39] Dempersmier J, Sambeat A, Gulyaeva Oet al. Cold-inducible Zfp516 activates UCP1 transcription to promote browning of white fat and development of brown fat. Mol Cell 2015; 57: 235–46.

[40] Kajimura S, Seale P, Tomaru Tet al. Regulation of the brown and white fat gene programs through a PRDM16/CtBP transcriptional complex. Genes Dev 2008; 22: 1397–409.

[41] Kleiner S, Mepani RJ, Laznik Det al. Development of insulin resistance in mice lacking PGC-1alpha in adipose tissues. Proc Natl Acad Sci USA 2012; 109: 9635–40.

[42] Marlatt KL, Ravussin E. Brown adipose tissue: an update on recent findings. Curr Obes Rep 2017; 6: 389–96.

[43] Fasshauer M, Klein J, Kriauciunas KMet al. Essential role of insulin receptor substrate 1 in differentiation of brown adipocytes. Mol Cell Biol 2001; 21: 319–29.

[44] Fasshauer M, Klein J, Ueki Ket al. Essential role of insulin receptor substrate-2 in insulin stimulation of Glut4 translocation and glucose uptake in brown adipocytes. J Biol Chem 2000; 275: 25494–501.

[45] Jiang B, Zhao W, Yuan Jet al. Lack of Cul4b, an E3 ubiquitin ligase component, leads to embryonic lethality and abnormal placental development. PLoS One 2012; 7: e37070.

[46] Eguchi J, Wang X, Yu Set al. Transcriptional control of adipose lipid handling by IRF4. Cell Metab 2011; 13: 249–59.

[47] Kong X, Banks A, Liu Tet al. IRF4 is a key thermogenic transcriptional partner of PGC-1alpha. Cell 2014; 158: 69–83.

[48] Zhang Y, Fang F, Goldstein JLet al. Reduced autophagy in livers of fasted, fat-depleted, ghrelin-deficient mice: reversal by growth hormone. Proc Natl Acad Sci USA 2015; 112: 1226–31.

[49] Amthor H, Macharia R, Navarrete Ret al. Lack of myostatin results in excessive muscle growth but impaired force generation. Proc Natl Acad Sci USA 2007; 104: 1835–40.

[50] Zhao TJ, Sakata I, Li RLet al. Ghrelin secretion stimulated by β1-adrenergic receptors in cultured ghrelinoma cells and in fasted mice. Proc Natl Acad Sci USA 2010; 107: 15868–73.

[51] Guan HP, Goldstein JL, Brown MS, et al. Accelerated fatty acid oxidation in muscle averts fasting-induced hepatic steatosis in SJL/J mice. J Biol Chem 2009; 284: 24644–52.

RIGHTS & PERMISSIONS

The Author(s) 2022. Published by Oxford University Press on behalf of Higher Education Press.