Diabetes Volume 66, December 2017 2987

RNA Binding Ybx2 Regulates RNA Stability During Cold-Induced Brown Fat Activation

Dan Xu,1,2 Shaohai Xu,3 Aung Maung Maung Kyaw,2 Yen Ching Lim,1 Sook Yoong Chia,2 Diana Teh Chee Siang,2 Juan R. Alvarez-Dominguez,4 Peng Chen,3 Melvin Khee-Shing Leow,5,6,7 and Lei Sun2,8

Diabetes 2017;66:2987–3000 | https://doi.org/10.2337/db17-0655

Recent years have seen an upsurge of interest in brown and inducible/beige adipocytes. Classic BAT is located as a adipose tissue (BAT) to combat the epidemic of obesity discernible depot in the interscapular region in small mam- and diabetes. How its development and activation are regu- mals and infants. Beige/inducible adipocytes exist in lated at the posttranscriptional level, however, has yet to be defined anatomical white adipose tissue (WAT) depots, par- fully understood. RNA binding (RBPs) lie in the cen- ticularly in subcutaneous WAT, and express a program ter of posttranscriptional regulation. To systemically study more like WAT at thermoneutrality. In response to pro-

fi > STUDIES OBESITY the role of RBPs in BAT, we pro led 400 RBPs in different longed cold exposure, chronic treatment of b-adrenergic adipose depots and identified Y-box binding protein receptor agonist, or intensive exercise, the number of beige 2 (Ybx2) as a novel regulator in BAT activation. Knock- adipocytes dramatically increases, accompanied by en- down of Ybx2 blocks brown adipogenesis, whereas its overexpression promotes BAT marker expression in brown hanced Ucp1 levels and mitochondria biogenesis, a process “ ” and white adipocytes. Ybx2-knockout mice could form BAT known as browning (2,5,6). but failed to express a full thermogenic program. Integrative Understanding the detailed mechanisms underlying BAT analysis of RNA sequencing and RNA-immunoprecipitation differentiation and function is an area of immense research study revealed a set of Ybx2’s mRNA targets, including interest. A vast array of factors has been identified that Pgc1a, that were destabilized by Ybx2 depletion during regulate BAT development and activity by acting at the cold-induced activation. Thus, Ybx2 is a novel regulator transcriptional level (6–16). How these processes are regu- that controls BAT activation by regulating mRNA stability. lated at the posttranscriptional level, however, has yet to be fully understood. RNA binding proteins (RBPs) comprise a largeanddiversegroup(17,18)thatlieatthecenterof Obesity has reached an epidemic scale in many countries, posttranscriptional regulation by governing the fate of resulting in a steep escalation in health care expenditures mRNA transcripts from biogenesis, stabilization, and trans- and a growing burden of chronic obesity-related morbidities lation to RNA decay. Several RBPs have been reported to (1). An attractive approach to improve metabolic health is modulate adipocyte development and lipid metabolism. to augment the mass and activity of brown adipose tissue SFRS10 (splicing factor arginine/serine-rich10) inhibits li- (BAT) (2–7). There are at least two types of thermogenic pogenesis by controlling the alternative splicing of LPIN1, a adipocytes in mammals, namely, classic brown adipocytes key regulator in lipid metabolism (19,20). Sam68 (the

1School of Laboratory Medicine and Life Science, Wenzhou Medical University, Corresponding author: Dan Xu, [email protected], or Lei Sun, sun.lei@duke- Wenzhou, Zhejiang, China nus.edu.sg. 2 Cardiovascular and Metabolic Disorders Program, Duke-NUS Medical School, Received 9 June 2017 and accepted 12 September 2017. Singapore This article contains Supplementary Data online at http://diabetes 3School of Chemical & Biomedical Engineering, Nanyang Technological Univer- .diabetesjournals.org/lookup/suppl/doi:10.2337/db17-0655/-/DC1. sity, Singapore 4Department of Stem Cell and Regenerative Biology, Harvard Stem Cell Institute, © 2017 by the American Diabetes Association. Readers may use this article as Harvard University, Cambridge, MA long as the work is properly cited, the use is educational and not for profit, and the 5Clinical Nutrition Research Centre, Singapore Institute for Clinical Sciences, work is not altered. More information is available at http://www.diabetesjournals Agency for Science, Technology and Research (A*STAR), Singapore .org/content/license. 6Department of Endocrinology, Tan Tock Seng Hospital, Singapore 7Office of Clinical Sciences, Duke-NUS Medical School, Singapore 8Institute of Molecular and Cell Biology, Singapore 2988 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

Src-associated substrate during mitosis of 68 kDa) is re- Retrovirus Transduction quired for WAT adipogenesis by regulating mTOR alter- Amurinestemcellvirus(MSCV)–based retroviral vector native splicing (21). Knockout of KSRP (KH-type splicing (MSCV-pgkGFP-U3-U6P-Bbs vector) (28) was used to gen- regulatory protein) promotes browning of WAT by reducing erate short hairpin (sh)RNAs to infect preadipocytes; miR-150 expression (22). IGF2 mRNA binding protein 2 XZ201 vector (29) was used to overexpress Ybx2 for gain- (IGF2BP2) is a widely expressed RBP, and a single nucleo- of-function studies. All of the retroviruses were packaged in tide polymorphism in its intron is associated with type 2 293T cells with the pCL-eco packaging vector and then used diabetes by genome-wide association studies (23). Knockout to transduce preadipocytes in the presence of 4 mg/mL of IGF2BP2 results in resistance to diet-induced obesity, Polybrene (Sigma-Aldrich), followed by induction of differ- largely resulting from an enhanced translational efficiency entiation. FuGENE 6 Transfection Reagent (Promega) was of Ucp1 and other mitochondria mRNAs in the knockout used for plasmid transfection according to the manufac- BAT (24). Paraspeckle component 1 (PSPC1) was recently turer’s instructions. identified as an essential RBP for adipose differentiation RNA Immunoprecipitation in vitro and in vivo by regulating the export of adipogenic Primary brown and white adipocytes were infected with RNA from the nucleus to the cytosol (25). Despite these retroviral Ybx2 and differentiated for 4 days. RNA immu- advances, our understanding of RBPs in adipocytes, partic- noprecipitation (RIP) was performed using the Magna RIP ularly in brown adipocytes, is still at its early stage, and the kit (Merck Millipore) according to the manufacturer’sin- functions of most RBPs remain unknown. structions. RNA samples retrieved from anti-Ybx2 (Abcam) In this study, we systemically profiled 413 RBPs in and IgG control with the Magna RIP kit were used for RNA different fat depots, during white fat browning and brown sequencing (RNA-seq). adipogenesis, and identified 5 BAT-enriched RBPs. We dem- onstrated the role of Y-box binding protein 2 (Ybx2) in the RNA Pull-Down development and activation of BAT in vitro and in vivo, which RNA pull-down was performed according to our published could be, at least partially, explained by stabilizing mRNA. protocol with a few modifications(30).Inthisstudy,we used tissue lysate from mouse BAT instead of primary cell RESEARCH DESIGN AND METHODS culturepreparedasdescribedabove.Thetissuelysatewas prepared as described in RNA IMMUNOPRECIPITATION.Therestof Animal Studies the experiment followed our published protocol (30). The animal experimental protocols in this study were ap- proved by the Singapore SingHealth Research Facili- Extracellular Flux Analysis ties Institutional Animal Care and Use Committee. Ybx2 Primary brown preadipocytes were seeded in an X-24 cell heterozygous mice (NSA [CF-1] background) were originally culture plate, infected by retroviral constructs, as indicated imported from Dr. Paula Stein (University of Pennsylvania). in the text, followed by induction of differentiation. C57BL6 mice were obtained from The Jackson Laboratory Differentiated cells were analyzed by Extracellular Flux and subsequently bred in house. All mice were maintained Analyzer (Seahorse Bioscience) according to the manufac- at the animal vivarium at Duke-NUS Medical School. For turer’s instructions. Oxygen consumption rates were nor- cold challenge experiments, animals were housed individu- malized by protein concentration. ally in a 4°C chamber for 6 h. The rectal body temperature Animals were kept at 4°C for 6 h before experiments. was recorded with a probe thermometer (Advance Technol- BAT and skeletal muscle (gastrocnemius) were harvested ogy) at a constant depth. and minced with a micromincer (Glen Mills Inc.). The Glucose tolerance tests and insulin tolerance tests were minced tissue was kept in ice-chilled mitochondrial respira- performed as previously described (26), and EchoMRI was tion media (MiR05) (EGTA, 0.5 mmol/L; MgCl2$6H2O, used to measure fat and lean mass. For the in vivo insulin- 3 mmol/L; lactobionic acid, 60 mmol/L; taurine, 20 mmol/L; signaling study, Ybx2 knockout (KO) and wild-type (WT) KH2PO4, 10 mmol/L; HEPES, 20 mmol/L; D-sucrose, mice were fasted for 6 h at room temperature or 4°C. Then 110 mmol/L; and BSA, 1 g/L). Tissue lysate, 2 mg and the mice were injected with insulin (1 unit/kg body wt). 10 mg, respectively, was immediately loaded into Oroboros Mice were sacrificed after 5 min, and BAT was collected. Respirometry together with substrates, including glutamate Lipolysis assay was performed as previously described (26). (10mmol/L),malate(2mmol/L),pyruvate(5mmol/L),and ADP (5 mmol/L). The oxygen consumption rate (OCR) was Cell Culture monitored at basal level and when the samples were treated 293T cells for retroviral packing were cultured in DMEM with different drugs: oligomycin (5 mmol/L), carbonyl cya- containing 10% FBS (HyClone). Primary brown and white nide-p-trifluoromethoxy phenylhydrazone (FCCP; 1 mmol/L), preadipocytes were isolated from 3- to 4-week-old C57BL6 and antimycin (5 mmol/L). mice. The procedure for preadipocytes isolation, culture, and differentiation and Oil Red O staining was described Western Blot previously (26). Human primary interscapular brown adipo- Western blotting was performed to detect target proteins using cytes were obtained from Zenbio Inc. and cultured and Ybx2 (Abcam), Gapdh (Abcam), Ucp-1(Abcam), Pgc1a (Santa differentiated as previously described (27). Cruz Biotechnology), Cidea (Santa Cruz Biotechnology), diabetes.diabetesjournals.org Xu and Associates 2989

Pparg (Santa Cruz Biotechnology), phosphorylated (p)-AKT we discovered six BAT-enriched RBPs that were induced (Cell Signaling), AKT (Cell Signaling), Cpt1a (Proteintech), during browning and brown adipogenesis, including Pgc1b, MCad (Santa Cruz Biotechnology), b-actin (Sigma-Aldrich), Larp4, Rbpms2, Grsf1, Akap1, and Ybx2 F (Fig. 1A–D), for and tubulin (Cell Signaling) antibodies. further investigation. b Analysis and Gene Set Enrichment Pgc1 is not a typical RBP so we excluded it from our Analysis subsequent experiments. The tissue enrichment and dy- fi Gene lists were analyzed for enrichment of Gene Ontology namic regulation of the ve other candidates during WAT (GO) terms by using Database for Annotation, Visualization browning were successfully validated by real-time PCR E and Integrated Discovery (DAVID) Functional Annotation across 15 major mice organs (Fig. 1 ) and in iWAT after G Tools (31,32). Gene set enrichment analysis (33) was per- the animals were housed for 7 days at 4°C (Fig. 1 ). In BAT, fi formed by using default parameters with the preranked only Ybx2 was signi cantly induced by cold treatment F gene sets. (Fig. 1 ). To test whether these RBPs were repressed by BAT and beige fat inactivation, we housed mice at thermo- RNA Decay Analysis neutrality (30°C) for 7 days to induce “whitening” of BAT Brown preadipocytes were cultured and differentiated and iWAT. All five RBPs were downregulated during for 5 days. We treated cells with 5 mg/mL actinomycin D BAT and iWAT “whitening” (Fig. 1H and I). We next ex- (Sigma-Aldrich) and harvested RNA at different times as amined their expression during an in vitro differentiation fi indicated in the gures. We took the same proportion of time course of primary brown and white adipocyte culture. RNA from each sample at a different time, conducted re- All five RBPs were upregulated during brown and white verse with random primers and real-time PCR. adipogenesis, with a higher expression level in brown adi- CTsfromeachsamplewereusedtocalculatetheremaining pocytes (Fig. 1J). fi percentage of mRNA at each point. We tthesedataintoa Finally, to test the human relevance of these observa- fi ’ rst-phase decay model to derive mRNAs half-life, tions, we examined their expression across a differentiation À Á time course of primary preadipocytes isolated from human Y 5ðY0 2 PlateauÞexp 2kdecay t 1 Plateau t fetal interscapular BAT and subcutaneous WAT (27). The expressionofYBX2andRBPMS2increasedthroughoutthe with Yt indicating the remaining percentage at a given time, human cell differentiation course with higher levels in BAT Y0 indicating the initial amount of RNA, t indicating time adipocytes. AKAP1 exhibited a significant induction from after transcription inhibition, and kdecay indicating the rate day 0 to day 7 and then decreased toward the end of constant. differentiation, but its level was still higher in brown adi- Statistical Analysis pocytes than in white adipocytes (Fig. 1K). Data are presented as the mean 6 SEM. Statistical signif- To investigate the function of these five RBPs in brown icance was assessed using the unpaired two-tailed Student t adipocyte differentiation, we depleted them by infecting test. Statistical significance for samples with more than two brown preadipocytes with retroviral shRNAs and then in- groups was determined by one-way ANOVA. The distribu- duced cells to differentiate for 5 days. Depletion of each of tion difference between different cumulative curves was these RBPs resulted in distinct phenotypes. Knocking down determined by the Kolmogorov-Smirnov test. P values of Ybx2 expression by ;90% (sh-3) severely blocked lipid ac- ,0.05 were considered to be significant. cumulation (Fig. 2A) and reduced the expression of pana- Accession Numbers dipogenic markers Fabp4 and Pparg2, indicating a block of The accession number for the RNA-seq data reported in this the panadipogenesis gene program, whereas inhibiting Ybx2 report is National Center for Biotechnology Information by ;70% (sh-1) affected BAT markers but did not affect B Omnibus (GEO): GEO SEries (GSE) 66686, panmarker expression and lipid accumulation (Fig. 2 ), sug- GSE29899, GSE86590, and GSE86338 gesting that the expression of BAT-selective is more sensitive to Ybx2 depletion. To determine the role of Ybx2 in cellular respiration, we inhibited its expression by ;70% RESULTS (sh-1) in brown adipocytes and used the Seahorse XFp Ex- Genome-Wide Identification of BAT-Enriched RBPs tracellular Flux Analyzer to measure the OCR. A significant To identify RBPs functionally important for BAT, we decrease of OCR for basal respiration and proton leakage profiled the gene expression of 413 RBPs annotated in was observed (Fig. 2E). the RBP database (18) in interscapular BAT, inguinal WAT Although knockdown of Akap1 slightly reduced lipids (iWAT), and epididymal WAT (eWAT), which led to identi- accumulation (Fig. 2A), it did not affect panadipogenic fication of 26 BAT-enriched RBPs. To further assess whether marker expression, but the BAT-selective markers were these RBPs are dynamically regulated during WAT browning downregulated (Fig. 2C,SupplementaryFig.1A). Inhibiting and brown adipogenesis, we examined their expression alter- Rbpms2 had a slight influence on lipid accumulation (Fig. 2A) nation during inguinal WAT browning induced by a b3- and panadipogenic marker expression (Fig. 2D)butstron- agonist (CL-316,243) and in primary brown preadipocytes ger effects on BAT-selective markers (Fig. 2D and Supple- versus mature adipocytes. By intersecting these gene sets, mentary Fig. 1A). OCR analysis consistently showed a 2990 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

Figure 1—Genome-wide identification of BAT-enriched RBPs. A–C : Gene expression of RBPs by RNA-seq in BAT, iWAT, and eWAT (A), during iWAT browning (B), and in primary brown preadipocyte and mature adipocytes (C). Heat maps show the row mean-centered abundance. D: Selection of gene expression from profiling studies A–C,plottedinVenndiagrams.E: Real-time PCR validation of gene expression for 5 RBPs across 15 mouse organs. Heat map shows the row mean-centered expression. FC, fold change. F and G: Gene expression of RBPs by real-time PCR in BAT (F)andiWAT(G) after mice (8 weeks old) were housed at 4°C for 7 days (n =6).H and I: Gene expression of RBPs by real-time PCR in BAT (H)andiWAT(I) after mice (8 weeks old) were housed at 30°C for 7 days. Mice housed at room temperature (RT) were used as the control group (n =5pergroup).J: Gene expression of RBPs during the differentiation of mouse primary brown (mBAT) and white (mWAT) adipocyte cultures (n =4).K: Gene expression of RBPs by real-time PCR during in vitro differentiation of stromal vascular fraction cells isolated from human fetal BAT (hBAT) and subcutaneous WAT (hWAT) (n = 4). The error bars are mean 6 SEM. *P , 0.05 by Student t test. diabetes.diabetesjournals.org Xu and Associates 2991

Figure 2—Ybx2 is an essential regulator of brown adipocyte differentiation in vitro. A: Primary brown preadipocytes were infected by retroviral shRNAs targeting RBPs, Ybx2, and Akap1, followed by induction of differentiation for 5 days. Oil Red O staining was used to assess lipid accumulation. B–D: Real-time PCR was used to measure the knockdown efficiency (left), panadipogenic marker expression (right), and BAT-selective marker expression (bottom) in cultured primary brown adipocytes (day 5) infected by retroviral shRNAs targeting Ybx2 (B), Akap1 (C), and Rbpms2 (D) (n =3).*P , 0.05 by one-way ANOVA. F: Representative metabolic flux curves from cultured brown adipocytes (day 5) infected by retroviral shRNA targeting Ybx2. Cells were sequentially treated with oligomycin, FCCP, and rotenone. OCRs are normalized by protein concentration (n = 5). *P , 0.05 by Student t test. F: Western blot was used to examine the protein levels of Ybx2 during primary brown and white adipocyte differentiation in culture. The error bars in the graphs are mean 6 SEM. Sh-1, short hairpin RNA-1; Sh-2, short hairpin RNA-2; Sh-3, short hairpin RNA-3. significant decrease of OCR attributed to proton leak in the Ybx2 Is an Essential Regulator of Brown Adipocyte Rbpms2- and Akap1-depleted cells (Supplementary Fig. 1B Differentiation In Vitro and C). In contrast, inhibiting Grsf1 and Larp4 did not Ybx2 harbors an ultraconserved cold shock RNA binding affect lipid accumulation (not shown) or marker expression domain (CSD). Proteins bearing CSDs, known as cold shock (Supplementary Fig. 2A–D). proteins, have been reported to regulate cellular adaptation 2992 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017 response, mainly at posttranscriptional levels, to cold stress real-time PCR and observed no change in panadipogenic in prokaryotes (34,35). BAT is a major organ for cold adap- markers (Supplementary Fig. 3E) but a detectable decrease tion in mammals; thus, the presence of CSD in Ybx2 sug- in Ucp1, Prdm16, and Dio2 (Supplementary Fig. 3F). RNA- gests that Ybx2 may play a role in BAT thermogenesis. We seq was performed to examine the global effects of Ybx2 KO validated the expression of Ybx2 at the protein level by on gene expression, but very few genes showed a significant Western blot in different adipose depots (Fig. 4A) and dur- difference (Supplementary Table 2), indicating that Ybx2 is ing brown and white adipogenesis (Fig. 2F). Consistent with dispensable for BAT to maintain its gene-expression pro- its mRNA expression pattern, Ybx2 protein level is higher in gram at room temperature. In iWAT, we did not observe BAT and induced during differentiation. To determine its significant change of BAT-selective markers or of HoxC10, a function in beige adipocytes, we knocked it down in prea- WAT marker (Supplementary Fig. 3G). The glucose toler- dipocytes isolated from inguinal WAT, followed by induc- ance test revealed a glucose intolerance (Supplementary tion of differentiation, and observed a clear reduction of Fig. 3H), and the insulin tolerance test detected a trend of BAT markers (Supplementary Fig. 2E–G). To ensure the insulin intolerance, but the difference was not statistically phenotypes of Ybx2 knockdown are not the result of off- significant (Supplementary Fig. 3H). Nevertheless, to what targeting effect, we further targeted different regions in its extent the impaired glucose tolerance can be accounted for mRNA using a different shRNA retroviral vector. Inhibiting by a smaller BAT or by systemic effects from other organs Ybx2 invariably impaired lipid accumulation and BAT needs to be investigated in the future. marker expression in both BAT and iWAT adipocyte cul- Because whole-body Ybx2 deficiency may have indirect tures (Supplementary Fig. 2H–L). effects on BAT phenotypes, to confirm whether Ybx2 KO We next tested whether Ybx2 is sufficient to promote may exert a cell-autonomous effect, we isolated brown pre- beige and brown adipogenesis by overexpressing Ybx2 in adipocytes from KO and WT mice for differentiation. Real- primary white and brown preadipocytes with retroviral time PCR revealed decreased expression of panadipogenic fi vector (Fig. 3A and D), followed by induction of differenti- markers and a more signi cant reduction of BAT-selective A–C ation. Ectopic expression of Ybx2 in white adipocytes en- markers in the KO cells (Supplementary Fig. 4 ), consis- hanced lipid accumulation assessed by boron-dipyrromethene tent with the shRNA knockdown phenotypes. RNA-seq was fi staining (Fig. 3B) and increased the expression of key BAT then performed to pro le the genome-wide effect of Ybx2 KO, markers such as Ucp1 and Pgc1a (Fig. 3C). Overexpression and gene set enrichment analysis revealed that the pathways of Ybx2 in primary brown adipocyte culture also enhanced of adipogenesis, fatty acid oxidation, oxidative phosphory- fi lipid accumulation (Fig. 3E)andBATmarkerexpression lation, and cellular respiration were signi cantly downregu- D (Fig. 3F) in the early phase of differentiation (day 3), which lated (Supplementary Fig. 4 ).Thus,Ybx2shouldhave was accompanied by a higher basal OCR and proton leakage cell-autonomous effects on brown adipocyte differentiation OCR (Fig. 3G). Western blot showed elevated protein levels in vitro, but such an effect was much ameliorated in vivo. of Ucp1 and two fatty acid oxidization regulators, Mcad and Cpt1a, at day 3 of differentiation (Fig. 3H). After 6 days of Ybx2 Is Required for Cold-Induced BAT Activation differentiation, the expression of BAT markers in control To determine the role of Ybx2 in BAT activation, we cells caught up with that in the Ybx2-overexpressing cells, exposed WT and KO animals to 4°C for 6 h. The WT BAT probably because the abundance of endogenous Ybx2 at this mass upon cold activation became smaller than that at stage is sufficient to support full induction of the BAT- room temperature, but the KO BAT mass did not decrease after cold exposure (Fig. 4E). Hematoxylin and eosin stain- selective gene program. Taken together, these observations ing consistently revealed that lipids in WT BAT but not the indicate that Ybx2 can promote brown adipogenesis in KO BAT were largely depleted (Fig. 4F and G), indicating white adipocyte culture and accelerate brown adipogenesis that the KO BAT failed to combust lipids upon cold activa- in brown adipocyte culture. tion. To directly assess the effects of Ybx2 KO’sfunction, Ybx2 Is Needed for Full BAT Development In Vivo we measured the OCRs for cold-activated WT and KO BAT To determine the function of Ybx2 in BAT in vivo, we with Oroboros respirometry. We observed a decreased OCR imported Ybx2 KO mice. The KO animals were infertile (36) in KO BAT before but not after FCCP treatment, which but viable and born at expected Mendelian ratios. We con- suggested that loss of Ybx2 did not change the maximal firmed their lack of Ybx2 by Western blot (Fig. 4A). The KO OCR capacity but reduced the cold-provoked mitochondria animals did not exhibit significant alteration in their body activity (Supplementary Fig. 5A). Consistently the core body weight (Fig. 4B), fat, and lean mass (Supplementary Fig. 3A temperature of KO mice dropped faster than that of WT and B). The iWAT and eWAT of KO animals also did not animals at cold temperature (Fig. 4H). Although the BAT change significantly in size (Supplementary Fig. 3C and D), defect is a certain culprit of the cold intolerance, we cannot whereas their iBATs were moderately but significantly preclude the possibility that the effect of Ybx2 KO on other smaller (Fig. 4B), coincident with slightly smaller lipid drop- organs can also contribute to this phenotype. lets under microscopy (Fig. 4C and D). To study the effect of We examined the lipolysis rates in the WT and KO BAT Ybx2 KO at the molecular level, we quantified the expres- and did not observe any significant change (Supplementary sion of panadipogenic and BAT-selective marker genes by Fig. 5B), indicating that the larger BAT mass in the KO BAT diabetes.diabetesjournals.org Xu and Associates 2993

Figure 3—Ybx2 can promote BAT-selective gene expression in white and brown adipocyte cultures. A:Westernblotwasusedtoconfirm the overexpression of Ybx2 in primary white adipocyte culture. B: Representative photomicrograph of boron-dipyrromethene staining for lipids in primary white adipocytes infected by Ybx2-expressing or empty vector. C: Real-time PCR was used to examine marker gene expression during the time course of white adipocyte cultures expressing Ybx2 or vector (n =4).D–F:SameasinA–C, but in primary brown adipocyte culture (n = 4). The error bars are mean 6 SEM. *P , 0.05. G: Representative metabolic flux curves from cultured brown adipocytes (day 3) infected by retroviral-overexpressing Ybx2. Cells were sequentially treated with oligomycin, FCCP, and rotenone. OCRs are normalized by protein concentration (n =5).H: Western blot was used to detect the protein levels of Ucp1 and two fatty acid oxidation components, Cpt1a and Mcad, at day 3. 2994 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

Figure 4—Ybx2 is needed for cold-induced BAT activation. A: Western blot was used to detect Ybx2 expression in eWAT, BAT, and iWAT from WT and KO mice. B: Body weight and BAT organ weight of WT (n =6)andKO(n = 7) male mice at 8–9weeksold.C: Representative photomicrograph of hematoxylin and eosin staining of WT and KO BAT. D: Distribution of the diameters of lipid droplets from panel C measured by ImageJ software. E: Body weight and BAT weight of WT and KO animals (8–9weeksold,n = 5) after 6 h at 4°C exposure. F:Representative picture and photomicrograph with hematoxylin and eosin staining of BAT from WT and KO mice after cold exposure. G: Distribution of the diameters of lipid droplets from panel F. H: Body temperature was measured by rectal probe at the indicated times at 4°C (n = 5). Error bars are the mean 6 SEM. *P , 0.05. is unlikely caused by any change in lipolysis. In addition, to expression in beige adipocytes, we performed real-time test whether loss of Ybx2 affects insulin signaling in BAT, PCR for iWATs and found a significant downregulation we performed Western blot to detect p-AKT in BAT and of Ucp1 but not other detected markers (Supplementary foundthatcoldchallengecouldenhancetheinsulinsensi- Fig. 5D). tivity in WT but not in KO BAT (Supplementary Fig. 5C). To To examine the effect of Ybx2 on BAT activation at the test whether Ybx2 KO may affect BAT-selective gene molecular level, we conducted RNA-seq of BAT isolated diabetes.diabetesjournals.org Xu and Associates 2995 from WT and KO mice at both room temperature and after Ybx2 versus the IgG sample and plotted the cumulative cold activation. One of the most striking observations was distributions for mitochondrion-related genes (206 genes) that the cold-induced thermogenic program in BAT was as well as for all genes detectable in the RIP-seq assay (4,095 severely hindered in KO animals. Ucp1, Dio2, Pgc1a,and genes). The cumulative curve of mitochondrion significantly Elvol3 were among the most significantly depleted genes in shifts toward the right (Fig. 6D), confirming that the targets KO upon cold exposure (Fig. 5A and Supplementary Fig. 6A), of Ybx2 are enriched for mitochondrial functions. which we validated by real-time PCR and Western blot (Fig. Next, we asked whether Ybx2 exerts a functional effect 5C and D). Consistent with the individual markers, pathways on its mRNA targets. We used RNA-seq data to calculate the analysis revealed that one of the most enriched pathways fold change of each mRNA between KO and WT BAT and associated with the downregulated genes is oxidation reduc- plotted the cumulative distributions for Ybx2 target and tion (Fig. 5B). nontarget mRNAs. At room temperature, distributions of To integrate the gene expression profiles at room targets and nontargets did not show a significant difference temperature and after cold exposure, we calculated the (Fig. 6E), but upon cold activation, targets of Ybx2 were fold change of each gene after cold exposure in WT and KO markedly repressed in KO BAT (Fig. 6F). These data support BAT (Supplementary Fig. 6B) and looked for enriched path- a role of Ybx2 in stabilizing its target mRNAs, which was ways among the most differentially regulated genes. The suggested by earlier work in oocytes (43). mitochondrion and fatty acid metabolic process pathways Ybx2 Targets and Stabilizes Pgc1a mRNA were among the top downregulated pathways (Supplemen- Among the top Ybx2 targets was the Pgc1a mRNA that is tary Fig. 6C). Importantly, the BAT mass and gene expres- significantlydecreasedinKOBAT.WeusedPgc1a as an sion changes in KO animals were not sex-dependent and example to illustrate how Ybx2 recognizes and affects its were also observed among female animals (Supplementary targets. To confirm binding between Pgc1a mRNA and Fig. 7A–D). Thus, although BAT can still form in the ab- Ybx2 in vivo, we performed the RIP-PCR in tissue lysate sence of Ybx2, its thermogenic response to cold tempera- from KO and WT BAT to detect Pgc1a mRNA precipitated ture is impaired. by Ybx2. A clear reduction of Pgc1a signal was detected in As a part of BAT adaptation to cold exposure, glucose RIP from KO BAT, but such a reduction was not observed uptake, lipogenesis, and combustion of long-chain fatty for Fabp4 mRNA, which bears a short 39 untranslated re- acids are increased in coordination with stimulation of gion (UTR; 180 bp) and is used as a control (Fig. 7A). To b-oxidation and thermogenesis (37,38). In Ybx2 KO BAT, dissect Ybx2-binding sites within Pgc1a mRNA, we per- besides thermogenic genes, those involved in glucose uptake formed an RNA pull-down assay using four in vitro tran- (Glut4), lipogenesis (Scd1, Fasn, Dgat1, Dgat2, Acaca), and scribed sequential RNA fragments from Pgc1a 39UTR and long-chain fatty acid generation (Elvol3, Elvol6) were also E E found that a 1,101 nucleotide RNA fragment (fragment 3) reduced (Fig. 5 and Supplementary Fig. 7 ), which was B B can readily retrieve Ybx2 protein (Fig. 7 ). We then gener- further supported by pathway analysis (Fig. 5 and Supple- ated eight small RNA fragments (200–300 bp) from this mentary Fig. 6C). Thus, Ybx2 is a regulator orchestrating segment for a second round of pull-down assays and found glucose metabolism, lipid metabolism, and thermogenesis thatfragments3.1,3.3,3.5,and3.7canretrieveYbx2(Fig.7C). during BAT activation. Intersecting segments 3.1 vs. 3.5 and 3.3 vs. 3.7 locate two Ybx2 Stabilizes mRNA Targets Encoding Proteins regions that harbor Ybx2 binding sites. Further truncation Enriched for Mitochondrial Functions of these two fragments abolished their interactions with To identify the mRNA targets of Ybx2, we performed RNA Ybx2 (data not shown), suggesting that a secondary or ter- immunoprecipitation, followed by RNA-seq (RIP-seq), by tiary nucleotide structure may be necessary for Ybx2 bind- using an antibody against Ybx2 in both brown and white ing. To test whether this identified RNA fragment can adipocyte culture (Supplementary Fig. 8A). First, we con- define the Ybx2 binding site in , we blasted the firmed the successful Ybx2 precipitation by Western blot human Pgc1a 39UTR and mouse Pgc1a 39UTR and identi- (Fig. 6A). We then selected candidates with at least eight- fied a ;1-kbsegmentthatwas.90% homologous to the fold enrichment in the Ybx2 IP sample compared with the fragment3inFig.7B (Fig. 7D). We cloned this fragment for IgG control, which revealed 800 and 1,822 potential mRNA pull-down assay. As expected, this fragment could retrieve targets in BAT and WAT, respectively. Targets in BAT and Ybx2 in BAT lysate (Fig. 7D), indicating that the Ybx2- WAT significantly overlapped, leading to identification of Pgc1a interaction is conserved. 414 common targets (Fig. 6B). As expected, Ybx2 can target To test whether the interactions between Ybx2 and Pgc1a is many mRNAs encoding proteins involved in posttranscrip- enhanced at cold exposure, we performed RIP-PCR in BAT tional RNA processing, a general feature of RBPs (39–42). before and after cold exposure and found that Ybx2 could Functional terms enriched among Ybx2 targets include ribo- retrieve more Pgc1a mRNAs upon cold exposure (Fig. 7E). some, ribonucleoprotein complex, and (Fig. 6C). To further study whether thisapparentincreaseisat- Interestingly, these targets were also enriched for mito- tributable to an enhanced binding affinity or an elevated chondria term (Fig. 6C). To further confirm this observa- Pgc1a mRNA abundance upon cold exposure, we inhibi- tion, we calculated the relative abundance of each mRNA in ted transcription with actinomycin D in cultured brown 2996 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

Figure 5—The effect of Ybx2 KO on cold-induced gene expression in BAT. A: Heat map of the gene expression in WT and KO BAT after cold exposure for 6 h. Heat map showed the row mean-centered abundance. FPKM, fragments per kilobase of exon per million fragments mapped. B: Five top nonredundant GO terms enriched among mRNAs that showed significantly low (top) or high (bottom) expression (P , 0.05, Cuffdiff) in KO vs. WT BAT. C: Real-time PCR to confirm gene expression of BAT-selective genes in WT and KO BAT (n =5).D: Western blot to confirm gene expression of BAT markers in WT and KO BAT. E: Real-time PCR to confirm the expression of genes involved in lipogenesis and glucose uptake (n = 5). Error bars are the mean 6 SEM. *P , 0.05. diabetes.diabetesjournals.org Xu and Associates 2997

Figure 6—Ybx2 stabilizes mRNA targets encoding proteins enriched for mitochondria functions. A: Western blot to confirm immunoprecipitation (IP) of Ybx2 protein by Ybx2 antibody; 10% IP cell lysate was used as the input. B: Targets of Ybx2 were selected based on their enrichment in the Ybx2 IP vs. IgG control. Venn diagram shows the overlapping of candidates from brown and white adipocytes. C:Bubblechartshowsthe GO terms enriched in the common targets. The x-axis indicates P values, and the y-axis indicates the enrichment score. The bubble size indicates the number of targets in that GO category. D: Relative abundance of each mRNA was calculated in anti-Ybx2 vs. IgG RIP-seq. The cumulative fraction of mRNAs involved in mitochondrion and all other detectable genes were plotted. The Kolmogorov-Smirnov test was performed to determine the distribution difference. E: Relative expression of each gene in KO vs. WT BAT at room temperature based on RNA-seq data. The cumulative fraction curves were plotted for 414 common target mRNAs and other genes detectable in the RIP-seq assays. F: The cumulative fraction curves were plotted for common target mRNAs and other genes after cold exposure. The Kolmogorov-Smirnov test was performed to determine the statistical significance of the difference in the distributions.

adipocytes and then performed RIP-PCR to detect the Ybx2- the mRNA-stabilizing effect from Ybx2, we constructed two Pgc1a mRNA interaction in the presence or absence of reporter plasmids: one containing an ;2kb Pgc1a 39UTR norepinephrine treatment. Interestingly, in the actinomycin after the Renilla luciferase (WT) and another containing a D treatment cells, the Pgc1a mRNA retrieved by Ybx2 an- truncated 39UTR without the Ybx2-binding fragment (mu- tibody was similar before and after norepinephrine treat- tant). We measured the decay rates of the Renilla luciferase ment (Supplementary Fig. 8B). Therefore, BAT activation mRNA in 293 cells in the presence and absence of a vector likely increases the Ybx2-retrived Pgc1a mRNA by stimu- expressingYbx2.IntheabsenceofYbx2,bothreporter lating Pgc1a mRNA expression but not by changing their constructs manifested similar decay rates (Supplementary binding affinity. Fig. 8C);inthepresenceofYbx2,themRNAdecayrateof To examine the influence of Ybx2 knockout on Pgc1a the WT reporter is approximately twofold slower than that mRNA stability, we used actinomycin D to stop mRNA of the mutant reporter (;8.4 h vs. ;4.1 h), indicating our transcription in WT and KO brown adipocyte culture and identified Ybx2-binding sites (Fig. 7B and C) in the Pgc1a measured the decay rates for Pgc1a and Fabp4 mRNA. The 39UTR are required for Ybx2’s mRNA stabilization function. half-life of Pgc1a mRNAdecreasedfrom2.39to1.29hin We further examined the functional interactions between the absence of Ybx2 (Fig. 7F), supporting a role of Ybx2 in Ybx2 and Pgc1a by overexpressing a full-length open read- stabilizing Pgc1a mRNA. To investigate whether the above- ing frame Pgc1a in the Ybx2-inhibited brown adipocytes. As identified Ybx2-binding sites in Pgc1a 39UTR can mediate described above (Fig. 2B), knockdown of Ybx2 reduced BAT 2998 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

Figure 7—Ybx2 binds and stabilizes Pgc1a mRNA. A: RIP assay with anti-Ybx2 in BAT lysate from WT and KO animals to examine the amount of Pgc1a mRNA in the immunoprecipitation (IP) samples. Fabp4 was used as a control; 5% tissue lysate in the IP reaction was used as the input (n =3).B and C: RNA pull-down assay was conducted to determine which RNA segments from Pgc1a 39UTR can bind Ybx2. Segments in 39UTR, as shown in the diagram, were cloned for in vitro transcription to generate RNA fragments, which were used for RNA pull-down assay in BAT lysate, followed by Western blot to determine presence of Ybx2 in each pull-down reaction. An adenylate-uridylate–enriched ;100 nucleotide (nt) fragment from androgen receptor (AR) was used as a negative control. D: RNA pull-down assay was conducted using a ;1-kb fragment from human Pgc1a 39UTR that is homologous to the fragment 3 in panel C.CDS,codingDNAsequence.E:RIP-PCRwas conducted to determine Pgc1a mRNA retrieved by anti-Ybx2 in BAT from room temperature (RT) and cold-exposed animals (n =3).F:Primary brown preadipocytes were isolated from WT and KO BAT for culture and then induced to differentiate for 5 days (left). Actinomycin D was added to stop transcription, and RNAs were harvested at the indicated times (x-axis) after transcription inhibition. Real-time PCR was used to determine remaining RNA level compared with the starting time. The trajectory of Pgc1a mRNA was fitintoafirst-order decay curve to derive the RNA half- life (WT T1/2 =2.39h;KOT1/2 = 1.29 h). Fabp4 mRNA was used as a control (n =6).G: We used retroviral constructs to knock down Ybx2 and overexpress Pgc1a in primary brown preadipocytes, followed by induction of differentiation. BAT-selective markers were examined by real-time PCR at day 6 (n = 4). The error bars are mean 6 SEM. *P , 0.05. diabetes.diabetesjournals.org Xu and Associates 2999 marker expression, but Pgc1a overexpression could signifi- Funding. This work was supported by Singapore National Research Foundation cantly rescue the phenotype (Fig. 7G). Therefore, although (NRF) fellowship (NRFF) (NRF-2011NRF-NRFF 001-025) to L.S. This research is also stabilizing Pgc1a alone is unlikely to account for all the supported by the Singapore NRF under its Cooperative Basic Research Grant (CBRG) ’ phenotypes of Ybx2 KO, Ybx2 may be a key target of grant administrated by the Singapore Ministry of Health s National Medical Research Council (NMRC) (NMRC/CBRG/0070/2014 and NMRC/CBRG/0101/2016). Ybx2, and the function of Ybx2 to some extent relies on Duality of Interest. No potential conflicts of interest relevant to this article a Pgc1 expression. were reported. Author Contributions. D.X., S.X., A.M.M.K., Y.C.L., S.Y.C., J.R.A.-D., and DISCUSSION D.T.C.S. performed experiments. D.X. and L.S. designed experiments and wrote the Early studies have suggested a role of Ybx2 in global mRNA manuscript.P.C.andM.K.-S.L.discussed the experiment design and critically stabilization. Schultz and colleagues (44) knocked down reviewed the manuscript. L.S. is the guarantor of this work and, as such, had full Ybx2 in oocytes by expressing a transgenic Ybx2 hairpin access to all the data in the study and takes responsibility for the integrity of the data double-stranded RNA. They observed a 60% reduction of and the accuracy of the data analysis. Ybx2 protein and a 75–80% reduction of poly-(A) mRNAs. In another study, they generated a knockout strain and References observed severe defects in spermatogenesis and oocyte 1. Olshansky SJ, Passaro DJ, Hershow RC, et al. A potential decline in life ex- – development (36) accompanied by an ;25% decrease of pectancy in the United States in the 21st century. N Engl J Med 2005;352:1138 1145 mRNAs in the mutant oocytes (43). Exogenous mRNAs in- 2. Kajimura S, Spiegelman BM, Seale P. Brown and beige fat: physiological roles jected into mutant oocytes were lower than that in WT beyond heat generation. Cell Metab 2015;22:546–559 cells, consistent with a decreased mRNA stability in the 3. Lee P, Swarbrick MM, Ho KK. Brown adipose tissue in adult humans: a absence of Ybx2 (43). This is consistent with our conclu- metabolic renaissance. Endocr Rev 2013;34:413–438 sion, which demonstrated a role of Ybx2 in enhancing 4. Tseng YH, Cypess AM, Kahn CR. Cellular bioenergetics as a target for obesity mRNA stability in a more systemic manner (Fig. 6F). therapy. Nat Rev Drug Discov 2010;9:465–482 As an RBP, Ybx2 may affect multiple RNA processing 5. Wu J, Cohen P, Spiegelman BM. Adaptive thermogenesis in adipocytes: is steps, including but not limited to RNA stability. Given its beige the new brown? Genes Dev 2013;27:234–250 cytosol localization (Supplementary Fig. 9C), the ability of 6. Harms M, Seale P. Brown and beige fat: development, function and therapeutic Ybx2 to influence translational control for certain mRNAs is potential. Nat Med 2013;19:1252–1263 not surprising. This may explain why the changes in mRNA 7. Rosen ED, Spiegelman BM. What we talk about when we talk about fat. Cell 2014;156:20–44 and protein levels for some genes, to some extent, may fl 8. Qiang L, Wang L, Kon N, et al. Brown remodeling of white adipose tissue by display discordance. The potential in uence of Ybx2 on SirT1-dependent deacetylation of Pparg. Cell 2012;150:620–632 translation will be further studied in the future. 9. Kong X, Banks A, Liu T, et al. IRF4 is a key thermogenic transcriptional partner Skeletal muscle is known to contribute to nonshivering of PGC-1a. Cell 2014;158:69–83 thermogenesis (NST) mainly through sarcolipin-mediated ATP 10. Rajakumari S, Wu J, Ishibashi J, et al. EBF2 determines and maintains brown hydrolysis by SERCA (45,46). To investigate whether Ybx2 adipocyte identity. Cell Metab 2013;17:562–574 KO can alter this pathway, we examined the NST genes in 11. Seale P, Kajimura S, Yang W, et al. Transcriptional control of brown fat de- muscles. Despite an increase of sarcolipin expression ob- termination by PRDM16. Cell Metab 2007;6:38–54 served in KO muscle, Ybx2 KO did not affect Serca1-3 12. Kajimura S, Seale P, Tomaru T, et al. Regulation of the brown and white fat expression (Supplementary Fig. 9A) and OCRs (Supplemen- gene programs through a PRDM16/CtBP transcriptional complex. Genes Dev 2008; – tary Fig. 9B) directly measured by Oroboros respirometry. 22:1397 1409 13.KajimuraS,SealeP,KubotaK,etal.Initiationofmyoblasttobrownfat Therefore, although whether the increase of sarcolipin ex- switch by a PRDM16-C/EBP-beta transcriptional complex. Nature 2009;460:1154– pression in muscle is the result of a tissue autonomous 1158 effect or a cross-organ response to the compromised KO BAT 14. Harms MJ, Ishibashi J, Wang W, et al. Prdm16 is required for the mainte- is unclear, the NST function of muscle was not altered. nance of brown adipocyte identity and function in adult mice. Cell Metab 2014;19: In sum, we profiled the expression of .400 RBPs across 593–604 different fat depots, during adipogenesis and WAT browning, 15. Villarroya F, Vidal-Puig A. Beyond the sympathetic tone: the new brown fat and identified Ybx2, a CSD-containing protein that orchestrates activators. Cell Metab 2013;17:638–643 BAT activation. CSD-containing proteins are among the most 16. Liu T, Kong D, Shah BP, et al. Fasting activation of AgRP neurons requires phylogenetically conserved families and are known for their NMDA receptors and involves spinogenesis and increased excitatory tone. Neuron role in cold adaptation in prokaryotes (34,35). Because BAT 2012;73:511–522 activationisapartofcoldadaptationinmammals,we 17. Gerstberger S, Hafner M, Tuschl T. A census of human RNA-binding proteins. Nat Rev Genet 2014;15:829–845 speculated that CSD proteins, exemplified by Ybx2, may 18. Cook KB, Kazan H, Zuberi K, Morris Q, Hughes TR. RBPDB: a database of RNA- be evolutionarily conserved to mediate cold adaptation at the binding specificities. Nucleic Acids Res 2011;39:D301–D308 whole organismal level via roles in BAT activation. 19. Pihlajamäki J, Lerin C, Itkonen P, et al. Expression of the splicing factor gene SFRS10 is reduced in human obesity and contributes to enhanced lipogenesis. Cell Metab 2011;14:208–218 Acknowledgments. The authors thank Dr. Paula Stein, University of 20. Brosch M, von Schönfels W, Ahrens M, et al. SFRS10–a splicing factor Pennsylvania, for the KO mice as a generous gift and Dr. Manvendra Singh, gene reduced in human obesity? Cell Metab 2012;15:265–266; author reply Duke-NUS Medical School, for the coordination of mice transportation. 267–269 3000 RBP Affects RNA Stability During BAT Activation Diabetes Volume 66, December 2017

21. Huot ME, Vogel G, Zabarauskas A, et al. The Sam68 STAR RNA-binding protein 34. Obokata J, Ohme M, Hayashida N. Nucleotide sequence of a cDNA clone regulates mTOR alternative splicing during adipogenesis. Mol Cell 2012;46:187–199 encoding a putative glycine-rich protein of 19.7 kDa in Nicotiana sylvestris. Plant Mol 22. Chou CF, Lin YY, Wang HK, et al. KSRP ablation enhances brown fat gene Biol 1991;17:953–955 program in white adipose tissue through reduced miR-150 expression. Diabetes 35. Graumann PL, Marahiel MA. A superfamily of proteins that contain the cold- 2014;63:2949–2961 shock domain. Trends Biochem Sci 1998;23:286–290 23. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide association study of 36. Yang J, Medvedev S, Yu J, et al. Absence of the DNA-/RNA-binding protein type 2 diabetes in Finns detects multiple susceptibility variants. Science 2007;316: MSY2 results in male and female infertility. Proc Natl Acad Sci U S A 2005;102: 1341–1345 5755–5760 24. Dai N, Zhao L, Wrighting D, et al. IGF2BP2/IMP2-deficient mice resist obesity 37. Townsend KL, Tseng YH. Brown fat fuel utilization and thermogenesis. Trends through enhanced translation of Ucp1 mRNA and other mRNAs encoding mito- Endocrinol Metab 2014;25:168–177 chondrial proteins. Cell Metab 2015;21:609–621 38. Yu XX, Lewin DA, Forrest W, Adams SH. Cold elicits the simultaneous induction 25. Wang J, Rajbhandari P, Damianov A, et al. RNA-binding protein PSPC1 pro- of fatty acid synthesis and beta-oxidation in murine brown adipose tissue: prediction motes the differentiation-dependent nuclear export of adipocyte RNAs. J Clin Invest from differential gene expression and confirmation in vivo. FASEB J 2002;16:155–168 2017;127:987–1004 39. Pullmann R Jr, Kim HH, Abdelmohsen K, et al. Analysis of turnover and 26. Kim HJ, Cho H, Alexander R, et al. MicroRNAs are required for the translation regulatory RNA-binding protein expression through binding to cognate feature maintenance and differentiation of brown adipocytes. Diabetes 2014;63: mRNAs.MolCellBiol2007;27:6265–6278 4045–4056 40. Keene JD. Minireview: global regulation and dynamics of ribonucleic acid. 27. Seiler SE, Xu D, Ho JP, et al. Characterization of a primary brown adipocyte Endocrinology 2010;151:1391–1397 culture system derived from human fetal interscapular fat. Adipocyte 2015;4:303– 41. Diaz-Muñoz MD, Bell SE, Fairfax K, et al. The RNA-binding protein HuR is 310 essential for the B cell antibody response. Nat Immunol 2015;16:415–425 28. Hattangadi SM, Burke KA, Lodish HF. Homeodomain-interacting protein kinase 42. Mukherjee N, Corcoran DL, Nusbaum JD, et al. Integrative regulatory mapping 2 plays an important role in normal terminal erythroid differentiation. Blood 2010; indicates that the RNA-binding protein HuR couples pre-mRNA processing and mRNA 115:4853–4861 stability. Mol Cell 2011;43:327–339 29. Sun L, Xie H, Mori MA, et al. Mir193b-365 is essential for brown fat differ- 43. Medvedev S, Pan H, Schultz RM. Absence of MSY2 in mouse oocytes perturbs entiation. Nat Cell Biol 2011;13:958–965 oocyte growth and maturation, RNA stability, and the transcriptome. Biol Reprod 30. Bai Q, Bai Z, Sun L. Detection of RNA-binding proteins by in vitro RNA pull- 2011;85:575–583 down in adipocyte culture. J Vis Exp 2016;e54207 44. Yu J, Deng M, Medvedev S, Yang J, Hecht NB, Schultz RM. Transgenic RNAi- 31. Huang W, Sherman BT, Lempicki RA. Systematic and integrative analysis of mediated reduction of MSY2 in mouse oocytes results in reduced fertility. Dev Biol large gene lists using DAVID bioinformatics resources. Nat Protoc 2009;4:44–57 2004;268:195–206 32. Huang W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths 45. Bal NC, Maurya SK, Sopariwala DH, et al. Sarcolipin is a newly identified toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res regulator of muscle-based thermogenesis in mammals. Nat Med 2012;18:1575– 2009;37:1–13 1579 33. Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a 46. Rowland LA, Bal NC, Periasamy M. The role of skeletal-muscle-based ther- knowledge-based approach for interpreting genome-wide expression profiles. Proc mogenic mechanisms in vertebrate endothermy. Biol Rev Camb Philos Soc 2015;90: Natl Acad Sci U S A 2005;102:15545–15550 1279–1297