Cluh plays a pivotal role during adipogenesis by regulating the activity of mitochondria

Cluh is a cytosolic protein that is known to specifically bind the mRNAs of nuclear-encoded mitochondrial proteins and play critical roles in mitochondrial biogenesis. Here, we report the role of Cluh in adipogenesis. Our study shows that mRNA expression of Cluh is stimulated during adipogenesis, and that cAMP/Creb signalling increases its transcription. Cluh depletion impaired proper adipocyte differentiation, with reductions seen in lipid droplets and adipogenic marker gene expression. Interestingly, the inductions of the brown adipocyte-specific genes, Ucp1, Cidea and Cox7a1, are severely blocked by Cluh depletion during brown adipogenesis. Mitochondrial respiration and the stability of mRNAs encoding mitochondrial proteins are reduced by Cluh depletion during brown adipogenesis. These results suggest that Cluh, which is induced during adipogenesis, promotes the post-transcriptional regulation of mitochondrial proteins and supports differentiation.


Results
Transcription of CLUH is activated during adipocyte differentiation. Since CLUH has been known to be a post-transcriptional regulator of mitochondrial protein, we hypothesized that it may play a crucial role in adipocytes, where the activity of mitochondria is increased [1][2][3][4] . To investigate the role of CLUH in adipocytes, we first examined its expression during adipogenesis. As in-vitro models for adipocyte differentiation, we selected murine cell lines that have been widely used in studying adipose tissue: 3T3-L1 cells represent a pre-adipose cell line for white adipocyte differentiation, and C3H10T1/2 cells represent a mesenchymal stem cell line for brown adipocyte differentiation. Differentiation protocol is illustrated in Supp. Fig. 1a. We validated our differentiation protocols for the two cell lines (Supp. Fig. 1a) by RT-PCR, western blotting, and Oil-Red O staining (Fig. 1a-c, Supp. Fig. 1b).
Cluh mRNA was found to increase during adipogenesis in C3H10T1/2 and 3T3-L1 cells (Fig. 1a,b, Supp. Fig. 2), beginning 2 days after adipogenic induction. This represents the clonal expansion stage, which is a relatively early stage of differentiation. The expression of Cluh during adipogenesis in murine pre-adipocytes was increased by approximately 7-10 fold, and the Cluh protein levels were also increased during adipogenesis in C3H10T1/2 and 3T3-L1 cells (Fig. 1c, Supp. Fig. 3).
To verify whether human CLUH also induced during adipogenesis, human mesenchymal stem cells (hMCSs) were differentiated into adipocytes and the expression of CLUH were analyzed. As shown in Fig. 1d, transcriptional level was enhanced up to approximately 11-fold. Since osteogenic and adipogenic lineages share common features, such as transcription pathways or increase of mitochondrial biogenesis 10 , we examined whether CLUH is also increased in hMSCs differentiating into osteocytes. CLUH was not found to increase in osteogenic differentiating cells, whereas the osteogenic marker, ALP, was noticeably induced (Fig. 1d).

Depletion of cluh impairs adipogenic differentiation.
To investigate the role of increased Cluh during adipogenesis, we performed siRNA-mediated knockdown of Cluh in C3H10T1/2 and 3T3-L1 cells, which were then differentiated into adipocytes and assessed for phenotypes related to adipogenesis (cell morphology, lipid droplets and adipocyte-specific gene expression). As shown in Fig. 2a, differentiated control cells took on the rounded features characteristic of differentiated adipocytes, whereas the Cluh-depleted cells, especially in the case of C3H10T1/2 cell line, maintained their fibroblast-like morphology. Moreover, number and size of lipid droplets (organelles in which lipids accumulate in differentiating adipocytes) was significantly reduced by Cluh depletion (Fig. 2a), as was the intensity of Oil Red O staining, which can be used to monitor lipid droplet accumulation Figure 1. Cluh is increased during adipogenesis. (a,b) Expression of Cluh was analyzed in C3H10T1/2 cells (a) and 3T3-L1 cells (b) at the indicated times after adipocyte induction, using RT-PCR and quantitative RT-PCR. Differentiation was confirmed by analyzing the adipogenesis markers, Ppar-γ, FABP4, and C/ebpα. Fulllength gels are presented in Supplementary Fig. 2a,b. (c) Increase of Cluh protein levels during adipogenesis was analyzed by western blotting. Differentiation was confirmed by detection of the Ppar-γ protein. A fulllength blot is presented in Supplementary Fig. 3. (d) Transcription of human CLUH during adipogenesis and osteogenesis was assessed using hMSCs. A dipogenesis or osteogenesis in hMSCs was confirmed by quantifying the transcription of PPAR-γ or ALP, respectively. In (a-d), all data are representative of three (c,d) or five (a,b) independent experiments. For quantitative RT-PCR, the Ct values of target mRNAs were normalized to those of Rpl13a for murine cells and B2M for human cells. Results were calculated using the ΔΔCt method and presented as the means ± SD of triplicate reactions. * indicates p < 0.05, ** indicates p < 0.01, and *** indicates p < 0.001. Figure 3. Transcription of Cluh is increased by cAMP/Creb, not by Ppar-γ. (a) C3H10T1/2 and 3T3-L1 cells were treated with rosiglitazone (5 µg/ml) without differentiation and the mRNA levels of Cluh and Fabp4 were analyzed with real time PCR. (b) Plasmids encoding C/ebpα, C/ebpβ, and Ppar-γ were transfected into exponentially growing 3T3-L1 and NIH-3T3 cells. After 24 h, the mRNA levels of Cluh and FABP4 were analyzed with real time PCR (upper panel). The expression of transfected plasmids was confirmed with western blotting using Flag antibody (lower panel). (c) NIH-3T3 cells were treated with forskolin for 6 h and the mRNA levels of Cluh were analyzed with real time PCR. (d,e) NIH-3T3 cells were transfected with plasmids encoding Creb (d), Prkaca, or its catalytic site mutants (e). After 24 h, the mRNA levels of Cluh were analyzed with real time PCR. Expression of transfected plasmids was confirmed with western blotting using Flag antibody (d; right panel, e; lower panel). (f,g) The activities of mCluh and hCLUH promoters were enhanced by Creb and Prkaca. Cells were co-transfected with pGL3-basic, pGL3-mCLUH promoter, or pGL3-hCLUH promoter plus Creb-(f) or Prkaca-(g) expressing plasmids or their backbone vectors. Firefly luciferase activity was normalized www.nature.com/scientificreports www.nature.com/scientificreports/ ( Fig. 2a,b). This apparent CLUH depletion-induced reduction in lipid droplets was quantitatively confirmed hereafter, by measuring the optical density of dissolved Oil Red O (Fig. 2b).
We next analyzed the transcription of adipose cell biomarkers, Ppar-γ and Fabp4, and found that they were reduced by Cluh depletion in cells undergoing adipogenesis (Fig. 2c). Interestingly, during the adipogenesis of C3H10T1/2 cells, expression levels of the brown adipocyte marker genes, Cidea and Cox7a1, were severely reduced by Cluh depletion (6~18 fold and 14 fold, respectively), whereas those of Fabp4 and Ppar-γ decreased by only about 2 fold.
Cluh is up-regulated by cAMP/Pka/Creb signaling, but not by Ppar-γ or C/EBP α/β. To investigate the machinery that controls the transcriptional induction of Cluh during adipogenesis, we examined the potential involvement of master regulators of adipogenic differentiation, including the transcription factors, Ppar-γ, C/ebpα and C/ebpβ, and cAMP signaling. We added cell line, NIH-3T3, to find a key adipogenic transcription factor that regulates Cluh directly, regardless of the conditions related to adipogenic differentiation.
First, we treated with Ppar-γ agonist, rosiglitazone to analyze the effect of Ppar-γ activation on the transcription of Cluh. Transcription levels of Cluh were not increased by the activation of Ppar-γ in 3T3-L1 and NIH-3T3 cells whereas Fabp4 was increased (Fig. 3a). To validate the effect of C/ebpα, C/ebpβ and Ppar-γ on the transcription of Cluh, we transfected cells with plasmids encoding the various regulators and confirmed their protein expression by western blot analysis ( Fig. 3b lower panel, Supp. Fig. 4a). Ppar-γ overexpression increased Fabp4 as reported ( Fig. 3b, upper panel) 11 , but did not affect the transcription of Cluh. Effect of Ppar-γ on the transcription of Fabp4 was greater in NIH-3T3 cells due to the higher transfection efficiency ( Fig. 3b lower panel). Although ectopic protein expression was higher than Ppar-γ, C/ebpα, and C/ebpβ also did not affect the transcription of Cluh and Fabp4 significantly (Fig. 3b).
To test whether cAMP signaling regulates the transcription of Cluh, we treated NIH-3T3 cells with forskolin for 6 hours and found that the transcription of Cluh to be three-fold higher after post-treatment (Fig. 3c). We also observed that overexpression of the cAMP-responsive transcription factor, Creb, increased the transcription of Cluh (Fig. 3d). Since the cAMP signaling pathway activates Creb through protein kinase A (Pka), we examined whether Pka could regulate Cluh transcription. Toward this end, we used plasmids encoding Prkaca (the catalytic subunit alpha of Pka) and its kinase-defective mutants H87G, T196A, and H87G/T196A 12  To determine whether Pka and/or Creb regulate the transcription of Cluh by modulating its promoter activity, we performed luciferase assays using the promoters of mCluh and hCLUH (Fig. 3f,g). The promoter activity of CLUH was 110-fold (in mouse) and 40-fold (in human) greater than that of the promoter-less mock vector, pGL3-Basic The activities of both human and mouse promoters were increased by the overexpression of Creb (3.6-and 3-fold, respectively; Fig. 3f) and Prkaca (3-and 2.8-fold, respectively; Fig. 3g). The catalytic-site mutant of Prkaca did not affect the activity of either promoter (Fig. 3g). We also confirmed the increase of Cluh protein by the overexpression of Creb and Prkaca in 3T3-L1 cells using western blotting (Supp. Fig. 4c).
Cluh is required for mitochondrial respiration during adipogenesis. We examined whether the function of Cluh during adipogenesis and in adipocytes was related to its known ability to regulate mitochondrial biogenesis 7,13 . We analyzed mitochondrial activity in Cluh-depleted and adipocyte differentiated (at day 4) C3H10T1/2 cells using a Seahorse XF analyzer to determine the indicator of mitochondrial respiration, Oxygen Consumption Rate (OCR) (Fig. 4a). Depletion of Cluh decreased ATP-linked respiration, maximum respiration rate and spare respiratory capacity (Fig. 4b). The maximal OCR was not increased relative to the basal OCR and the spare respiratory capacity was severely reduced compared to that of control cells. These results imply that the mitochondria in Cluh-depleted cells were already working at their maximal capacity to compensate for the defective mitochondrial respiration. We further analyzed mitochondrial dysfunction by measuring ROS, and found it to be increased upon Cluh depletion during adipogenesis in both 3T3-L1 and C3H10T1/2 cells (Fig. 4c).
Cluh regulates mRNA stability of mitochondrial proteins during adipogenesis. Since Cluh is known to control the stability of mRNA and translation of many mitochondrial proteins 13 , we investigated whether it could control the nuclear-encoded mitochondrial genes, Acat1, Atp5a1, Hadha, Pcca, and Pdha1, during adipogenesis. These genes had been previously selected as representative nuclear-encoded mitochondrial genes from among hundreds of screened candidate transcripts that may bind to CLUH 7 , and the ability of CLUH to stabilize their mRNAs has been verified 13 . RNA levels of all tested genes were found to have increased during adipogenesis in C3H10T1/2 cells, and these increases were inhibited by depletion of Cluh (Fig. 5a). Cluh depletion had greater effect on the transcript levels during adipogenesis compared to those observed prior to differentiation (day 0).
To address whether the reduction of mRNA levels in Cluh-depleted cells was related to post-transcriptional regulation of RNA stability, we measured mRNA decay in siRNA-transfected cells that had differentiated into to that of Renilla luciferase and fold induction was plotted with respect to the normalized luciferase activity of pGL3-Basic vector-transfected cells. In (a-g), results presented are representative of three independent experiments, all performed in non-differentiated conditions. In (a,b), the average Ct value was calculated from triplicate quantitative RT-PCRs; in (c-g), experiments were performed in triplicate from the cell-seeding stage. * indicates p < 0.05, ** indicates p < 0.01, and *** indicates p < 0.001. www.nature.com/scientificreports www.nature.com/scientificreports/ adipocytes (day 6) by treatment of actinomycin D. As shown in Fig. 5b, the mRNAs of all tested genes were destabilized by depletion of Cluh in adipose-differentiated C3H10T1/2 cells (Fig. 5b, shown in log scale in Supp. Fig. 5).

Discussion
Mitochondria play a crucial metabolic energy-consuming role in adipose tissues. To reduce obesity and improve insulin sensitivity, adipose and other tissues must enhance their mitochondrial oxidation rates, and hence increase energy expenditure. Several previous had reports shown that the mitochondrial function of adipocytes is reduced in metabolic diseases including diabetes, hepatosteatosis, obesity, and cardiovascular complications 14,15 . Mitochondrial functions such as mitochondrial oxidative capacity, mitochondrial membrane potential, and citrate synthase activity were found to be lower in the adipocytes of obese patients compared to that in control group 16,17 . These findings suggest that proteins related to mitochondrial function in adipocytes could be potential targets for treating adipose-related metabolic diseases.
In order to explore the genes that play key roles in adipogenesis, we screened for adipogenesis-induced genes. Amongst the identified genes, we focused on CLUH, since it was known to specifically function as a post-transcriptional regulator in mitochondria. Here, we report that Cluh is up-regulated in both white and brown adipose cells, and plays a crucial role during adipocyte differentiation. Cluh depletion was found to impair lipid droplet formation, proper adipocyte differentiation, capacity of mitochondrial respiration, and stability of genes encoding mitochondrial proteins. We observed lipid droplets as a measure of adipogenic differentiation. Reduction of lipid droplets in Cluh-depleted adipogenic cell differentiation reflects improper differentiation. A recent study has reported that lipid droplets are increased in liver tissue of Cluh-deficient mice due to the metabolic dysfunction of mitochondria, resulting in inefficient β oxidation 13 . However, the different results may have arisen from the same cause of mitochondrial defect. This defect may lead to defective adipogenesis, thereby reducing lipid droplets, or cause inefficient β oxidation, thereby inducing lipid droplets. The effect of Cluh depletion on lipid droplet in other tissues might be different from our results.
Although this is the first report of the contribution of Cluh in adipogenesis, it also confirms that Cluh does not alter the stability of brown adipocyte-specific genes, such as Ucp1, Cidea, and Cox7a1. Instead, its direct targets were Acat1, Atp5a1, Hadha, Pcca, and Pdha1, which function in relation to central pathways for mitochondrial respiration, namely fatty acid β-oxidation (FAO) and oxidative phosphorylation (OXPHOS). Mitochondria are particularly active in adipocytes, where they help to maintain the normal metabolic functions such as FAO and OXPHOS. We found these genes were increased during adipogenesis, and mRNAs were stabilized by Cluh especially in adipocytes. Therefore, Cluh may contribute to the maintenance of metabolic homeostasis by supporting the appropriate protein expression pattern and subsequent function of mitochondria, thereby facilitating proper adipogenesis.
To define the mechanism by which Cluh is transcriptionally activated during adipogenesis, we analyzed whether adipogenesis-related core transcription factors regulate the transcription of Cluh. Our results indicated that transcription of Cluh was activated by Pka/Creb signaling, but not by Ppar-γ or C/EBPα/β. Pka/Creb is an upstream regulator of adipogenesis that is known to control C/EBPβ to activate adipocyte-related genes, such as C/EBPα, KLF5, PPAR-γ and AP2 18,19 . cAMP signalling has been reported to play a critical role in brown fat by stimulating the levels of key thermogenic factors, such as UCP1 20,21 . In this study, we uncover a novel role of cAMP signaling, as an upstream regulator of Cluh, during adipogenesis. However, the increased levels of Cluh by cAMP/CREB signaling were relatively lower than expected. Changes in transcription machinery or in epigenetics during adipogenesis may cooperatively regulate the transcription of Cluh.
Since mitochondria support the differentiation and homeostatic function of adipocytes, their dysfunction has serious consequences 1,2 . Excess white adipose tissue (WAT) undergoes various remodeling process, including hypertrophy, hyperplasia and/or inflammation. Sustained obesity and inflammation lead to abnormality of WAT, causing mitochondrial dysfunction and consequent metabolic diseases 22 . Many studies have focused on the remodeling of WAT into brown-like adipose tissue [23][24][25] . Here, we report that the transcription of Cox7a1, Cidea, and Ucp1, which are brown-fat-specific genes required for thermogenesis and mitochondrial biogenesis, were significantly reduced by the depletion of Cluh from C3H10T1/2 cells while that of common adipogenesis markers were decreased less than 2 fold. Moreover, Cluh depletion decreased lipid droplets more seriously in C3H10T1/2 cells than in 3T3-L1 cells. Thus, Cluh depletion appears to severely impair proper brown adipogenesis. Mitochondrial function is more critical in BAT than in WAT. For example, UCP1, which is exclusively expressed in BAT and known to play a key role in brown adipogenesis, is present in the mitochondrial inner membrane, its activity being greatly affected by mitochondria itself 26 . The abnormal BAT differentiation upon Cluh depletion could be caused by dysfunction of mitochondria. Together, our results suggest that Cluh is more critical in brown adipocytes than in white adipocytes, and affects the characteristics of differentiating adipocytes. Further studies would be required to fully elucidate the role of Cluh in modulating adipocytes.
In summary, we herein show that Cluh plays a key role in adipogenesis by post-transcriptionally regulating mitochondrial proteins in adipocytes. Our findings suggest that CLUH could be a novel target for modulating metabolic energy consumption, controlling adipocyte differentiation, and/or changing the characteristics of adipocytes.

Oil Red O staining and quantification of differentiation. Differentiated adipocytes on 35-mm dishes
were fixed with 10% formalin for 10 min, followed by washing with 60% isopropanol for 5 min. The fixed cells were dried, stained with Oil Red O solution (Sigma-Aldrich) for 1 h, and washed four times with water. For quantification, the retained Oil Red O was dissolved in 5 ml (per dish) isopropanol and its absorbance measured at 510 nm in a plate reader. siRNA, plasmids, and transfection. siRNA duplexes were synthesized by Genolution (Korea). The sequences were as follows: Cluh siRNA#1: 5′-GCU UCA AUC CUG ACA UCU U-3′; Cluh siRNA#2: 5′-GGG CAU CAU UGG CAA UGA U-3′; and negative control: 5′-AAU UCU CCG AAC GUG UCA CGU UU-3′. Cells were transfected with siRNA at a final concentration of 50 nM using RNAiMAX (Invitrogen, Carlsbad, CA) in Opti-MEM media (Invitrogen). During adipogenesis, siRNAs were transfected at day −1.
All plasmids used in this study were either purchased from Addgene, or directly cloned, or obtained as gifts. pcDNA flag mPPAR gamma was a gift from Bruce Spiegelman (Addgene plasmid # 8895) 27 . pcDNA-mC/EBPα and pcDNA-mC/EBPβ LAP were gifts from Jed Friedman (Addgene plasmids#66978 and #66979, respectively) 28 . pCF CREB was a gift from Marc Montiminy (Addgene plasmid #22968) 29 . For cloning of Pka-related constructs, cDNA synthesized from IEC-6 cells was used as the template for amplification of Prkaca, which was cloned into p3xFlag vectors. Its catalytic-site mutants (H87G, T196A, and H86G/T196A) 12 were constructed using a site-directed mutagenesis kit (iNtRON Biotechnology, Korea) according to the manufacturer's protocol. Cells were transiently transfected with 3 µg of plasmids using TurboFect (Fermentas Life Sciences, Ontario, Canada) in 60-mm dishes. After 24 h, the cells were harvested for analysis using quantitative RT-PCR or western blotting.
Reverse transcription pCR and real-time pCR. mRNA expression was analyzed using RT-PCR and real-time PCR. Total RNA was extracted using the RNeasy Mini kit (GeneAll, Korea) and reverse-transcribed into cDNA using an iScript cDNA synthesis kit (Bio-Rad, Hercules, CA).The resulting cDNA was amplified using either a Maxime PCR PreMix kit (iNtRON Biotechnology) for conventional real-time PCR or the SYBR green method (iQ SYBR Green super mix, Bio-Rad) for real-time quantitative PCR. The primer sequences are listed in Table 1. For real-time PCR, triplicate reactions were performed for each analysis, and the cycle threshold (Ct) values of both test and reference genes (mouse sample, Rpl13a; human sample, B2M) were calculated using the ΔΔCt equation (2 −ΔΔCt method). www.nature.com/scientificreports www.nature.com/scientificreports/ Western blotting. Western blotting was performed using the following primary antibodies: anti-CLUH (A301-764A; Bethyl Laboratories, Montgomery, TX), anti-PPAR-γ (SC-7273; Santa Cruz Biotechnology, Santa Cruz, CA), anti-FABP4 (#2120; Cell Signaling Technology, Beverly, MA), anti-Flag (F3165, Sigma-Aldrich) and anti-GAPDH (SC-25778, Santa Cruz Biotechnology). Immunoreactive proteins were detected using enhanced chemiluminescence (ECL) reagents (Santa Cruz Biotechnology). promoter assay. A 1.4-kb genomic fragment encompassing bases −1414 to −1 from the translation start site (+1) of the mouse Cluh gene and a 1.4-kb genomic fragment encompassing bases −1391 to −36 from the translation start site (+1) of the human CLUH gene were cloned into the pGL3-Basic vector (Promega, WI). For the luciferase assay, NIH-3T3 cells on six-well plates were co-transfected with mock vector (pCDNA3 or p3XFlag) or plasmids encoding Creb or Prkaca (1 µg) plus the pRL Renilla luciferase control reporter vector (30 ng) and a promoter plasmid (1 µg). After 30 h, the cells were lysed with 0.2 ml passive lysis buffer (Promega). Luciferase activity was measured using a dual luciferase assay kit (Promega) according to the manufacturer's protocol. Subsequently, the Renilla luciferase activity was measured. The activity of Cluh promoter-driven firefly luciferase gene in pGL3 cells was normalized to that of the corresponding Renilla luciferase to correct for between-sample differences in transfection efficiency. Each normalized promoter activity was plotted in terms of its relative fold-change compared to the normalized luciferase activity of cells transfected with the pGL3-basic vector and control vector (pCDNA3 or p3xFlag).
Mitochondrial stress assay. The oxygen consumption rate (OCR) of siRNA-transfected differentiating cells was measured using a Seahorse XF24 analyzer and an XF assay kit (Seahorse Bioscience, North Billerica, MA). siRNA-transfected differentiating C3H10T1/2 cells (at day 3) were seeded at 3 × 10 4 cells/well on an XF24 plate at 24 h before the assay. All samples were prepared in triplicate. On the day of the assay, the medium was changed to DMEM containing 584 mg/ml L-glutamine, without serum, glucose or bicarbonate, and the plates were incubated for 1 h in a non-CO 2 incubator at 37 °C. The reagents for the assay were prepared using an XF Cell Mito Stress Test kit (Seahorse Bioscience, MA) according to the manufacturer's protocol, and injection was performed according to a standard assay protocol (Port A: oligomycin; Port B: FCCP; and Port C: rotenone/ antimycin A). Results were generated automatically using the Seahorse XF Mito Stress Test Reporter Generator.
Reactive oxygen species measurement. Production of intracellular reactive oxygen species (ROS) was analyzed using flow cytometry. Cells were transfected with the appropriate siRNA, differentiated to adipocytes for 3 days, incubated with DCFDA (10 μM) for 30 min, and harvested using trypsin. Intracellular ROS levels were analyzed using the FL1 channel of a FACSCalibur flow cytometer (BD Biosciences). mRNA decay measurement. C3H10T1/2 cells were examined for adipocyte differentiation. The cells were transfected with siRNAs on day −1, differentiated into adipocytes or left undifferentiated on day 0, and treated with 2 μg/ml actinomycin D on day 6. The cells were harvested at 1, 2, 4 and 8 h after drug treatment. statistical analyses. All data were analyzed as the mean ± SD. Statistical differences between two means were assessed using Student's t test (unpaired, two-tailed). A p value of >0.05 was considered to be statistically non-significant (n.s.).