Ursodeoxycholic Acid Inhibits Liver X Receptor α-mediated Hepatic Lipogenesis via Induction of the Nuclear Corepressor SMILE*

Background: Small heterodimer partner interacting leucine zipper protein (SMILE) is a nuclear corepressor of the nuclear receptor family. Results: Ursodeoxycholic acid (UDCA) increases SMILE gene expression, which contributes to inhibiting liver X receptor α (LXRα)-mediated hepatic lipogenesis. Conclusion: UDCA-induced SMILE inhibits LXRα-mediated hepatic lipogenic gene expression. Significance: SMILE improves hepatic lipid metabolism. Small heterodimer partner interacting leucine zipper protein (SMILE) has been identified as a nuclear corepressor of the nuclear receptor (NRs) family. Here, we examined the role of SMILE in the regulation of nuclear receptor liver X receptor (LXRα)-mediated sterol regulatory element binding protein-1c (SREBP-1c) gene expression. We found that SMILE inhibited T0901317 (T7)-induced transcriptional activity of LXRα, which functions as a major regulator of lipid metabolism by inducing SREBP-1c, fatty acid synthase (FAS), and acetyl-CoA carboxylase (ACC) gene expression. Moreover, we demonstrated that SMILE physically interacts with LXRα and represses T7-induced LXRα transcriptional activity by competing with coactivator SRC-1. Adenoviral overexpression of SMILE (Ad-SMILE) attenuated fat accumulation and lipogenic gene induction in the liver of T7 administered or of high fat diet (HFD)-fed mice. Furthermore, we investigated the mechanism by which ursodeoxycholic acid (UDCA) inhibits LXRα-induced lipogenic gene expression. Interestingly, UDCA treatment significantly increased SMILE promoter activity and gene expression in an adenosine monophosphate-activated kinase-dependent manner. Furthermore, UDCA treatment repressed T7-induced SREBP-1c, FAS, and ACC protein levels, whereas knockdown of endogenous SMILE gene expression by adenovirus SMILE shRNA (Ad-shSMILE) significantly reversed UDCA-mediated repression of SREBP-1c, FAS, and ACC protein levels. Collectively, these results demonstrate that UDCA activates SMILE gene expression through adenosine monophosphate-activated kinase phosphorylation, which leads to repression of LXRα-mediated hepatic lipogenic enzyme gene expression.

Small heterodimer partner interacting leucine zipper protein (SMILE) has been identified as a nuclear corepressor of the nuclear receptor (NRs) family. Here, we examined the role of SMILE in the regulation of nuclear receptor liver X receptor (LXR␣)-mediated sterol regulatory element binding protein-1c (SREBP-1c) gene expression. We found that SMILE inhibited T0901317 (T7)-induced transcriptional activity of LXR␣, which functions as a major regulator of lipid metabolism by inducing SREBP-1c, fatty acid synthase (FAS), and acetyl-CoA carboxylase (ACC) gene expression. Moreover, we demonstrated that SMILE physically interacts with LXR␣ and represses T7-induced LXR␣ transcriptional activity by competing with coactivator SRC-1. Adenoviral overexpression of SMILE (Ad-SMILE) attenuated fat accumulation and lipogenic gene induction in the liver of T7 administered or of high fat diet (HFD)-fed mice. Furthermore, we investigated the mechanism by which ursodeoxycholic acid (UDCA) inhibits LXR␣-induced lipogenic gene expression. Interestingly, UDCA treatment significantly increased SMILE promoter activity and gene expression in an adenosine monophosphate-activated kinase-dependent manner. Furthermore, UDCA treatment repressed T7-induced SREBP-1c, FAS, and ACC protein levels, whereas knockdown of endogenous SMILE gene expression by adenovirus SMILE shRNA (Ad-shSMILE) significantly reversed UDCA-mediated repression of SREBP-1c, FAS, and ACC protein levels. Collectively, these results demonstrate that UDCA activates SMILE gene expression through adenosine monophosphate-activated kinase phosphorylation, which leads to repression of LXR␣-mediated hepatic lipogenic enzyme gene expression.
Liver X receptor ␣ (LXR␣) 3 is a member of the nuclear receptor superfamily and heterodimerizes with retinoid X receptor (RXR). LXR␣ normally binds to the DR-4 motif known as the LXR response element (LXRE) in their target genes and is activated by specific cholesterol metabolites such as oxysterols or synthetic nonsteroidal LXR ligands (T0901317) (1)(2)(3). Two isoforms of LXR exist with different expression patterns. LXR␤ is expressed ubiquitously, whereas LXR␣ is mainly expressed in liver, adipose tissue, and macrophages known to play an important role in lipid metabolism (1,4,5). LXR␣ has emerged as an important regulator of gene expression involved in lipid and cholesterol metabolism (6,7). LXRs play an important role in fatty acid synthesis by directly or indirectly controlling lipogenic gene expression. LXR␣-mediated fatty acids synthesis is controlled by sterol regulatory element-binding protein-1c (SREBP-1c), which regulates gene expression involved in the lipogenic pathway, and acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS) in hepatocytes (8). Consistently, * This work was supported in part by National Creative Research Initiatives  LXR␣-deficient mice display markedly reduced hepatic fatty acid synthesis and triglycerides (9). LXR␣ activity is regulated by agonists, antagonists, SIRT1, NCoR, and SMRT (10,11). We previously demonstrated that orphan nuclear receptor DAX-1 inhibits LXR␣ transcriptional activity and improves hepatic lipogenesis (12). Small heterodimer partner interacting leucine zipper protein (SMILE) belongs to the basic leucine zipper family transcription factor. In contrast to other basic leucine zipper proteins, SMILE cannot bind to DNA as a homodimer, unlike other basic leucine zipper proteins (13). SMILE has also been proposed as a coactivator of activating transcription factor 4 (14). Moreover, we have revealed that SMILE functions as a nuclear co-repressor of the estrogen receptor (ER), glucocorticoid receptor (GR), constitutive androstane receptor (CAR), hepatocyte nuclear factor 4␣, and estrogen receptor-related receptor ␥ (ERR ␥) (15)(16)(17). Moreover, we reported that curcumin induces SMILE gene expression through a liver kinase B1/adenosine monophosphate-activated kinase (AMPK) pathway and represses endoplasmic reticulum (ER) stress-responsive gene transcription (18). A recent study demonstrated that SMILE activates tumor suppressor p53 and inhibits the function of BMP-6 by interacting with Smads (19,20). However, the detailed mechanism by which SMILE regulates expression of genes involved in hepatic lipid metabolism remains unknown.
Bile acids affect triglyceride (TG) homeostasis and have recently emerged as a key metabolic regulator of glucose and lipid metabolism (21). Ursodeoxycholic acid (UDCA) is a bile acid and is used to treat several liver diseases such as primary biliary cirrhosis and hepatitis C (22,23). UDCA induces hepatobiliary secretion and decreases retention of toxic hydrophobic bile acids, thereby rendering bile more hydrophilic and cytotoxic (23). Moreover, UDCA has beneficial effects on cholestatic disorders due to its anti-apoptotic, anti-fibrotic, and cytoprotective effects (24). UDCA treatment increases glutathione synthesis by activating the PI3K/Akt/Nrf2 pathway (25) and have beneficial effects on hepatic steatosis and insulin resistance (26,27). Other bile acids such as cholic acid (CA) and chenodeoxycholic acid (CDCA) prevent hepatic TG accumulation (28). However, the molecular mechanism of UDCA in hepatic lipid metabolism remains largely unknown.
In the present study, we show that SMILE negatively regulates LXR␣ transcriptional activity by directly interacting with LXR␣ and competes with the LXR␣ coactivator SRC-1. Moreover, we observed that SMILE overexpression inhibited LXR␣-mediated Srebp-1c gene expression and decreased LXR␣ agonist-induced hepatic TG level and lipid accumulation. Moreover, UDCA treatment attenuated LXR␣-mediated lipogenic gene expression, whereas SMILE knockdown released UDCA-mediated gene expression of Srebp-1c, Fas, and Acc. Overall, our results suggest that SMILE acts as a novel corepressor of LXR␣ and UDCA-induced SMILE gene expression, which leads to inhibition of LXR␣-mediated hepatic lipogenic gene expression.

EXPERIMENTAL PROCEDURES
Animals and Treatment-C57BL/6J male mice were obtained from Korea Research Institute of Bioscience and Biotechnology (Daejeon, South Korea). Age matched mice (weight, 25-28 g) were maintained on a standard diet under a 12-h light/dark cycle at 22 Ϯ 2°C for 2 weeks with free access to food and water in a pathogen-free facility. UDCA was administered daily orally at a dose of 2 mg, after which the mice had free access to food and water. All animal experiments were approved by the Institutional Animal Care and Use Committee of Chonnam National University.
Cell Culture and Transient Transfection Assay-HepG2 and 293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Cambrex Bio Science, Walkersville, MD) and antibiotics (Invitrogen). The cells were split in 24-well plates at densities of 2-8 ϫ 10 4 cells/well the day before transfection. Transient transfections were performed using the Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's instructions. Total DNA used in each transfection was adjusted to 1 g/well by adding the appropriate amount of empty vector, and CMV-␤-galactosidase plasmids were cotransfected as an internal control. Cells were harvested 40 -48 h after the transfection for luciferase and ␤-galactosidase assays. Luciferase activity was normalized to ␤-galactosidase activity.
Preparation of Recombinant Adenovirus-The recombinant adenovirus encoding human SMILE has been described previously (15). The shSMILE (AAGGCGTCGTCGTCTCTTAAA) constructs were constructed with a 21-mer double-stranded oligonucleotide containing ϩ1053 to 1074 of the SMILE cDNA sequence into the pBS/U6 vector. The cDNA encoding shSMILE was cloned into the pAdTrack-CMV vector. pAdTrack-CMV-shSMILE was recombined with adenoviral gene carrier vector by transformation in pretransformed adEasy-BJ21-competent cells. The shAMPK␣2 (ATCATCT-TATCATTGACAATCGGAGAA) were constructed using double-stranded oligonucleotides containing ϩ1071 to 1094 of the AMPK␣2 cDNA sequence into the pBS/U6 vector and then adenoviral vector systems were used as previously described (12).
Isolation and Culture of Primary Mouse Hepatocytes-Mouse primary hepatocytes were isolated from the livers of 7-week-old male C57BL6 mice. The mice were anesthetized with Zoletile and their livers were exposed surgically. The liver was first perfused with resuspension buffer and then perfused with collagenase solution. Subsequently, the liver was finely chopped in a Petri dish and then filtered through 85-m pore mesh. Hepatocytes were collected by centrifugation at 800 ϫ g for 2-5 min at 4°C. Hepatocyte viability was assessed by trypan blue exclu-sion assay and was consistently in excess of 85%. Hepatocytes were then seeded onto collagen type 1-coated 60-mm dishes as described previously (12).
Triglyceride and Cholesterol Measurements-Once the animals were sacrificed, blood was immediately collected, and serum levels of TG and cholesterol were measured using chemistry analyzer kits (Hitachi 7150; Tokyo, Japan).
Histomorphological Analysis-Liver tissues were processed in paraffin after fixation in 10% neutral-buffered formalin. Tissue sections (3-4 m thick) were deparaffininzed, rehydrated, and stained with hematoxylin and eosin (H&E) for histological examination. The histological evaluation was carried out using a BX51 light microscope (Olympus, Tokyo, Japan). Histological damage in H&E sections was examined under a microscope at ϫ200 magnification.
Oil Red-O Staining-Fresh liver tissues were embedded carefully in OCT in a plastic mold after freezing at Ϫ80°C. Tissue sections (8 -10 m thick) were incubated for 30 min at room temperature on a slide. The tissues were fixed with 10% formalin for 1 h, and then washed 60% isopropyl alcohol. Then 0.6% Oil Red-O working solution (w/v, 60% isopropyl alcohol and 40% water) was added to each slide for 30 min at room temperature, the solution was removed, and deionized water was added to wash the tissue. Fat droplets in the liver were stained red.
RNA Interference-SMILE knockdown was performed using the pSuper vector system (15). HepG2 cells were transfected with siRNA constructs using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The siRNA-transfected cells were subjected to the second transfection as indicated in the figure legends.
GST Pulldown Assay-Briefly, LXR␣ was labeled with [ 35 S]methionine using the TNT-coupled reticulocyte lysate system (Promega Corp., Madison, WI) according to the manufacturer's instructions. GST alone and GST-fused SMILE (GST-SMILE) proteins were prepared for in vitro GST pulldown assays as previously described (17). Briefly, 293T cells were transfected with 1 g of each of the indicated plasmids using the Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's instruction. The cells were collected 24 h after transfection and solubilized. The in vivo GST pulldown was performed as described previously (17).
Western Blot Analysis and Co-immunoprecipitation (Co-IP)-Western blot and co-IP analyses were performed as described previously (17). Briefly, HepG2 cells were transfected with the indicated expression vectors or small interfering RNA oligonucleotides and treated with T7. Cell lysates were prepared 48 h after transfection and separated on 10% SDS-PAGE gels. The proteins were transferred to a nitrocellulose membrane (Amersham Biosciences), and the membranes were probed with anti-HA, FLAG, SMILE, LXR␣, SREBP1, FAS, ACC, or ␤-actin antibodies and developed after incubation with a secondary antibody using an enhanced chemiluminescence kit (Amersham Biosciences) according to the manufacturer's instruction.
Statistical Analysis-Data are expressed as mean Ϯ S.E. The statistical analysis was performed using the two-tailed Student's t test or one-way analysis of variance and results were considered to be statistically significant at p Ͻ 0.05 or p Ͻ 0.001.

SMILE Decreases LXR␣-mediated Lipogenic Enzyme Gene
Expression-To investigate the role of SMILE in T7-induced lipogenic gene expression, we infected adenovirus expressing SMILE (Ad-SMILE) in T7-treated mouse primary hepatocytes and HepG2 cells. We found that adenoviral overexpression of SMILE significantly decreased T7-induced protein levels of LXR␣ target genes Srebp-1c, Fas, and Acc (Fig. 1, A and B). In addition, adenoviral overexpression of SMILE significantly decreased mRNA levels of SREBP-1c, FAS, and ACC in both primary mouse hepatocytes and HepG2 cells (Fig. 1, C and D). Taken together, these results indicate that SMILE inhibits LXR␣-mediated lipogenic enzyme gene expression.
SMILE Competes with SRC-1 to Decrease LXR␣ Transcriptional Activity-To confirm whether SMILE inhibits LXR␣ transcriptional activity, 293T and HepG2 cells were co-transfected with SMILE, the LXR␣ expression vector, and the LXR␣specific reporter in the presence or absence of the LXR␣ agonist T7. Overexpression of SMILE strongly inhibited the activity of the LXR␣-mediated reporter containing the LXR binding site (LXRE-luc) (Fig. 2, A and B). It has been reported that SRC-1 acts as a co-activator of LXR␣ to activate LXR␣ transcriptional activity (29). To define the functional mechanism of LXR␣ repression by SMILE, we employed competition experiments using transient transfection of HepG2 cells. We observed that SMILE-mediated repression of LXRE-luc and Srebp-1c promoter activity were significantly released by SRC-1 co-transfection (Fig. 2, C and D). These results suggest that SMILE competes with SRC-1 to decrease LXR␣-mediated transcriptional activity. Next, we performed co-IP assays using LXR␣, SRC-1, and FLAG-SMILE antibodies to confirm whether SMILE inhibits the protein interaction between LXR␣ and SRC-1. Overexpression of SMILE decreased the endogenous interaction of SRC-1 with LXR␣. In contrast, overexpressed SMILE strongly interacted with LXR␣ (Fig. 2E). Moreover, we performed ChIP assay to examine whether SMILE also inhibited recruitment of LXR␣ on the Srebp-1c promoter. We found that SMILE abolished recruitment of LXR␣ on the Srebp-1c promoter (Fig. 2F). On the basis of inhibition of LXR␣ DNA binding by SMILE, we hypothesized that the effect of SMILE on LXR␣ DNA binding was mediated by alterations in the interaction between LXR␣ and RXR. To investigate this possibility, we performed co-IP assays to test if SMILE affects LXR␣/RXR␣ dimerization. The effect of SMILE on heterodimerization of LXR␣/RXR␣ was determined by co-immunoprecipitation in vitro. The interaction between LXR␣ and RXR␣ was suppressed by SMILE expression in a dose-dependent manner (supplemental Fig. S1). These results suggest that SMILE impairs the interaction between LXR␣ and RXR␣, accounting for SMILE-mediated inhibition of LXR␣ DNA binding, at least in part. Taken together, these results indicate that SMILE physically competes with SRC-1 to repress FIGURE 1. SMILE decreases T7-mediated lipogenic gene expression. A and B, mouse primary hepatocytes and HepG2 cells were infected with Ad-SMILE or Ad-GFP and then exposed to T7. Whole cell extracts were isolated and analyzed using Western blot analysis with the indicated antibodies. Protein levels were normalized to ␣-tubulin level. Each data point represents mean Ϯ S.E. C and D, quantitative PCR analysis was performed using total RNA from Ad-SMILE or Ad-GFP-infected mouse primary hepatocytes or HepG2 cells after T7 treatment. The Srebp-1c, Fas, Acc, SMILE, and ␤-actin genes were amplified using specific primers for Srebp-1c, Fas, Acc, SMILE, and ␤-actin. mRNA levels were normalized to ␤-actin expression. Results are representative of three independently performed experiments.
LXR␣ transcriptional activity and leads to the blocking of LXR␣ DNA binding on the target gene promoter.
SMILE Directly Interacts with LXR␣-To determine whether the repression of LXR␣ transcriptional activity by SMILE is mediated through a direct physical interaction, we performed co-IP assays using LXR␣and SMILE-specific antibodies. Co-IP assays demonstrated that SMILE directly interacted with LXR␣ in HepG2 cells (Fig. 3A). Next, to confirm whether the interaction between the two proteins is direct or not, we performed in vitro GST pulldown assays. GST alone and GSTfused SMILE were bacterially expressed and incubated with in vitro translated 35 S-labeled LXR␣ protein. We found that in vitro translated 35 S-labeled LXR␣ protein directly interacted with bacterially expressed SMILE, but not with GST alone (Fig.  3B). To further confirm the interaction between SMILE and LXR␣ in vivo, we performed in vivo GST pulldown assays upon transfecting cells with GST-alone (pEBG) or GST-SMILE (pEBG-SMILE) with HA-LXR␣. After GST purification, HA-LXR␣ was detected in the coprecipitates only when co-expressed with GST-SMILE but not with GST alone in the presence or absence of the LXR␣ agonist T7 (Fig. 3C). Previous reports demonstrated that DAX-1 directly interacts with the ligand binding domain of LXR␣ to compete with SRC-1 (12). To identify the LXR␣ interacting domain with SMILE, we per-formed in vivo GST pulldown experiments using a series of GST-tagged LXR␣ mutants (Fig. 3D). We found that SMILE interacted with the LXR␣ C-terminal, which contains the ligand binding domain and the activation function-2 domain (Fig. 3E). In addition, all LXR␣ and SMILE GST fusion proteins used in the assays were expressed at comparable levels (Fig. 3E, middle and lower panels). Because the LXXLL motif plays an important role in the regulation of nuclear receptor (30), transfection analyses were carried out using LXXLL mutant constructs to examine whether SMILE-dependent inhibition of LXR␣ activity is mediated by the LXXLL motif of SMILE. The transient transfection study showed that third (m3) or fourth (m4) single LXXLL motif mutants of SMILE failed to inhibit LXR␣ transactivation to the level comparable with wild type SMILE. Therefore, these results suggest that the LXXLL motifs are essential for SMILE to inhibit LXR␣ activity (Fig. 3F). Taken together, these results demonstrate that SMILE directly interacts with LXR␣, and the LXXLL motif of SMILE is required to inhibit LXR␣ activity.
UDCA Induces SMILE Gene Expression-A previous report demonstrated that the bile acid CDCA inhibits T7-induced gene expression of ACC␣ and other lipogenic enzyme (31) and also reduces hypertriglyceridemia (28). Another natural bile acid, UDCA, also decreases TG level and improves cholestasis . E and F, HepG2 cells were co-transfected with pcDNA3-Flag-SMILE, and the cells were treated with vehicle (DMSO) or T7 (10 M). After 6 h, protein extracts were co-immunoprecipitated using LXR␣ or IgG antibody and Western blotted with the indicated antibody (E). Chromatin immunoprecipitation analysis was performed using LXR␣ antibody. PCR amplification of immunoprecipitated (IP) chromatin fragments was conducted using primer pairs specific for proximal, regulatory region, and a distal, nonregulatory region of the Srebp-1c gene promoter (F). (24,32). To test if UDCA improved hepatic lipid metabolism is caused by SMILE, we assessed levels of SMILE mRNA and protein following UDC treatment. UDCA treatment increased SMILE protein levels in a time-dependent manner both in mouse primary hepatocytes and HepG2 cells (Fig. 4, A and B). UDCA was administered to mice to further determine the effects of UDCA on SMILE gene expression in vivo. UDCA treatment (80 mg/kg/day) for 1 week significantly increased SMILE protein levels (Fig. 4C). In addition, we also observed that UDCA strongly increased the SMILE protein level in a dose-dependent manner (Fig. 4D). We next asked whether UDCA administration elicits hepatotoxicity. H&E staining revealed similar liver morphology between normal and UDCAtreated mice. Moreover, UDCA treatment did not alter AST (aspartate aminotransferase) and ALT (alanine aminotransferase) levels compared with control (supplemental Fig. S2, A-F). Previous reports demonstrated that CDCA suppress hepatic lipogenesis by inducing SHP gene expression (31). Next, we com- FIGURE 3. Interaction between SMILE and LXR␣. A, endogenous interaction between LXR␣ and SMILE. Protein extracts from HepG2 cells were subject to co-immunoprecipitated (IP) with LXR␣ or SMILE and the interaction between LXR␣ and SMILE was determined by Western blotting using the SMILE antibody (left panel) or LXR␣ antibody (right panel). LXR␣ and SMILE expression (lower two panels) from the 10% lysate was analyzed by Western blotting (WB) with the indicated antibodies. B, in vitro GST pulldown assay. 35 (m1, m2, m3, m4, and m5). Luciferase activity was measured 48 h after transfection.
pared the effects of different bile acids on SMILE expression in HepG2 cells. Treatment with UDCA time dependently increased the SMILE mRNA level but not SHP mRNA (Fig. 4E). In contrast, CDCA significantly increased the SHP mRNA level. However, no effect on SMILE gene expression was observed following CDCA treatment (Fig. 4F). Taken together, these results show that UDCA has a positive role in SMILE gene expression and suggest that UDCA-mediated inhibition of LXR␣-mediated hepatic lipogenesis may be elicited through the induction of SMILE gene expression.
UDCA-induced SMILE Gene Expression Depends on AMPK Signaling-We have reported that AMPK increases SMILE gene expression (18), which prompted us to evaluate the effect of UDCA on AMPK signaling. First, we analyzed AMPK phosphorylation and subsequent changes in SMILE gene expression in primary mouse primary hepatocytes following UDCA treatment. Consistent with our hypothesis, UDCA time dependently increased AMPK phosphorylation and SMILE gene expression in mouse primary hepatocytes (Fig. 5A). Next, to investigate whether AMPK signaling is involved in UDCA-induced SMILE gene expression, we employed compound C, AMPK inhibitor, and adenovirus expressing short hairpin RNA for AMPK␣2 (Ad-shAMPK␣2). UDCA-induced SMILE gene expression was significantly attenuated by compound C treatment, which is consistent with abolished phosphorylation of AMPK (Fig. 5B). Consistent with the effect of compound C, knockdown of AMPK␣2, a liver-enriched isoform of AMPK␣, significantly decreased induction of SMILE protein and mRNA level by UDCA treatment (Fig. 5, C and D). Moreover, we found that UDCA also activated SMILE promoter activity, whereas compound C treatment dose dependently inhibited UDCA-induced SMILE promoter activity (Fig. 5E). Consistent with the compound C effect, dominant-negative AMPK (dn-AMPK) overexpression strongly inhibited UDCA-induced SMILE promoter activity (Fig. 5F). To further investigate the role of AMPK, we tested if AMPK has any direct role on the stability of SMILE proteins. We overexpressed FLAG-tagged SMILE and infected AMPK-CA (constitutively active form) or AMPK-DN (dominant-negative form) virus to confirm that AMPK affects SMILE protein stabilization. However, overexpressed SMILE protein level was not significantly changed by adenoviral overexpression of AMPK, indicating that AMPK plays a role on SMILE expression at transcriptional level (supplemental Fig.  S3). These results suggest that the AMPK pathway is mainly involved in the UDCA-induced expression of SMILE at the transcriptional level.
UDCA Decreases LXR␣-mediated Lipogenic Gene Expression by Inducing SMILE-To test if UDCA also inhibits lipogenic gene expression and whether the ability of UDCA to inhibit lipogenic gene expression is associated with SMILE, we performed an adenovirus-mediated knockdown experiment with adenovirus expressing short hairpin RNA for SMILE (Ad- for various periods and then the cells were harvested for Western blot analysis using SMILE antibody. Protein levels were normalized to those of ␣-tubulin. C, 80 mg/kg of UDCA was administered daily to mice for 1 week. Then, the mice were sacrificed and the isolated liver was processed for Western blot analysis with SMILE antibody. D, the indicated dose of UDCA (0 -200 mg/kg) was administered to mice for 4 days. Mice were sacrificed and Western blot analysis was performed using SMILE antibody from the mouse liver tissue. Each data point represents mean Ϯ S.E. shSMILE). UDCA strongly inhibited the T7-mediated increase in SREBP-1c, FAS, and ACC protein levels, which were significantly reversed by knockdown of endogenous SMILE expression in mouse primary hepatocytes and HepG2 cells (Fig. 6, A  and B). However, knockdown of basal SMILE expression did not display significant changes in lipogenic gene expression. Therefore, uninduced levels of SMILE may not be involved in the regulation of the basal hepatic lipogenic gene expression (supplemental Fig. S4). Taken together, these results indicate that SMILE is involved in UDCA-mediated repression of lipogenic gene expression. Next, we performed knockdown of endogenous SMILE using a SMILE small interfering RNA expressing plasmid to investigate whether UDCA inhibits LXR␣ target gene promoter activity by inducing SMILE. As expected, knockdown of endogenous SMILE significantly abolished UDCA-mediated inhibition of T7-induced Srebp-1c promoter activity (Fig. 6C). Furthermore, we performed ChIP assay experiments to define the molecular mechanism of UDCA-induced SMILE on the regulation of LXR␣. UDCA treatment significantly decreased recruitment of LXR␣ to the Srebp-1c promoter. However, SMILE knockdown significantly released UDCA-mediated suppression of LXR␣ recruitment at the Srebp-1c promoter (Fig. 6D). Overall, these results suggest that UDCA inhibits LXR␣-mediated lipogenic gene expression by inducing SMILE gene expression, which decreases LXR␣ recruitment on the target gene promoter.
SMILE Improves T7-induced Hepatic TG Level and Lipid Accumulation in Mice-To assess the functional effects of SMILE on T7-induced hepatic lipogenesis, we analyzed hepatic fat accumulation and lipogenic gene expression in normal and T7-treated mice. A 1-week T7 treatment caused severe fat accumulation in the liver of mice, as indicated by increases in the intensity of Oil Red-O staining and hepatic TG levels. As expected, adenovirus injection of SMILE significantly improved hepatic lipid accumulation (Fig. 7A) and TG levels (Fig. 7B) compared with those in adenovirus GFP-injected mice. However, hepatic cholesterol accumulation and hepatic cholesterol levels remained unchanged (Fig. 7, C and D). In parallel with the improved hepatic fat accumulation and TG levels by SMILE, the increased Srebp-1c, Fas, and Acc gene expression were significantly decreased by SMILE (Fig. 7E). Next, we assessed the inhibitory effect of SMILE on high fat diet (HFD)-induced hepatic lipogenic gene expression in mice. Consistent with changes in T7-induced lipogenic gene expression by SMILE, adenoviral overexpression of SMILE significantly improved HFD-induced hepatic lipogenic gene expression (Fig. 7F). Taken together, these results indicate that activating SMILE considerably improved T7-or dietinduced hepatic fat accumulation by regulating of lipogenic enzyme gene expression.

DISCUSSION
We demonstrated previously that SMILE acts as a corepressor of nuclear receptors (16). In this study, we found that SMILE decreased LXR␣-mediated lipogenic target gene expression such as SREBP-1c, FAS, and ACC in hepatocytes. Moreover, we demonstrated that SMILE improved hepatic lipid accumulation, TG levels, and inhibited hepatic lipogenic gene expression using an in vivo model of T7-treated and HFD-fed mice. SMILE-mediated LXR␣ target gene transcription was regulated by competition with SRC-1 and the LXR␣ DNA binding block. Moreover, we showed FIGURE 7. SMILE improves T7-induced hepatic lipid accumulation. A, male 7-week-old C57BL6 mice were provided with a standard rodent diet. T0901317 (LXR agonist, 50 mg/kg body weight) or vehicle (1% methylcellulose and 1% Tween 80) were administered by oral gavage each day for 1 week. An aliquot of 0.5 ϫ 10 9 plaque-forming units of GFP or SMILE adenovirus were delivered by tail vein injection on day 4 of oral gavage. Three days after injection, the mice were sacrificed and Oil Red-O staining was performed on the liver samples. B, liver triglyceride levels were analyzed from the mouse liver tissue as in A. C and D, liver cholesterol staining (C) and hepatic cholesterol levels (D) were analyzed as in A and B. E, Srebp-1c, Fas, and Acc mRNA levels in mouse liver were analyzed by real time quantitative RT-PCR. F, real time-quantitative RT-PCR analysis of hepatic Srebp-1c, Fas, and Acc in adenovirus GFP or SMILE injected mice that were fed chow or HFD for 12 weeks. All data were normalized to those of ␤-actin and ribosomal L32 expression. Data in B and D-F are represented as mean Ϯ S.D.
that inhibiting lipogenic gene expression through UDCA-induced AMPK signaling depended on inducing SMILE gene expression.
LXR␣ is crucial for hepatic lipogenic gene expression and fatty acids synthesis in hepatocytes (9). In this study, we determined that SMILE is a corepressor of LXR␣. In the transient transfection assay, SMILE overexpression significantly repressed LXR␣ transcriptional activity. Therefore, the repression of LXR␣ transcriptional activity by SMILE suggests that SMILE could improve LXR␣-mediated fatty liver disease by regulating LXR␣ transcriptional activity. This hypothesis is corroborated by the observation that SMILE improved T7-induced hepatic lipid accumulation and TG levels. Moreover, SMILE inhibited hepatic lipogenic gene expression in HFD-fed mice. Overall, these findings strongly suggest that SMILE is a potent modulator of hepatic lipogenesis by regulating LXR␣. Moreover, a previous report demonstrated that SMILE represses GR, CAR, and hepatocyte nuclear factor 4␣ (16). GR, CAR, and hepatocyte nuclear factor 4␣ are important for hepatic regulation and processing of glucose, lipids, drugs, and bile acids (33)(34)(35). Based on previous reports, we suggest that SMILE may be involved in diverse liver metabolic activities. SMILE knock-out and a transgenic animal model would be useful to better understand the role of SMILE in liver metabolism. Similar to SMILE-mediated repression of LXR, DAX-1 and SHP also decrease LXR␣ transcriptional activity (12,36,37). Consistent with previous reports, we found that SMILE physically interacted with LXR␣ and significantly inhibited recruitment of LXR␣ on the Srebp-1c promoter. Moreover, a domain mapping analysis using an in vivo GST pulldown assay showed that the LBD/AF2 domain of LXR␣ was essential for interaction with SMILE (Fig. 3E). In addition, it has been demonstrated that SRC-1 activates LXR␣ transcriptional activity (29). Here, we demonstrated that SMILE competes with the coactivator SRC-1. Moreover, we also found that SMILE inhibits LXR␣ binding on the Srebp-1c promoter in part by blocking the interaction between LXR␣ and RXR␣ (supplemental Fig. S1). These data suggest that SMILE regulates LXR␣ activity via multiple inhibitory mechanisms.
UDCA has been used as a therapeutic agent for fatty liver disease (27,38) and cholestatic disorder diseases (24,39). UDCA also has beneficial effects on liver regeneration in rats with non-alcoholic fatty liver disease (38). However, the effect of UDCA on lipid metabolism has remained largely unknown.
Here, we showed that UDCA significantly activated SMILE gene expression depends on AMPK activation. In contrast, UDCA strongly decreased SREBP-1c, FAS, and ACC gene expression by inducing the SMILE gene (Fig. 6, A and B). These results demonstrate that SMILE is a critical transcriptional regulator of genes involved in UDCA-mediated regulation of lipid metabolism. Moreover, UDCA improves hepatic ER stress and insulin sensitivity. Notably, UDCA, a side chain-shortened homologue of UDCA, improves fatty liver and atherosclerosis (21). Consistent with the efficacy of UDCA in ER stress, our previous study showed that SMILE plays critical roles in regulating ER stress (18). Similar to the UDCA effect on lipogenic gene expression, a previous report demonstrated that cholic acid decreases hepatic expression of SREBP-1c, FAS, and ACC by inducing SHP gene expression (28). In addition, CDCA also suppresses T7-induced lipogenic gene expression (31). In the present study, we observed that UDCA particularly activated SMILE expression, whereas CDCA had no effect on SMILE gene expression (Fig. 4, E and F). UDCA is also reported to be a FXR ligand that activates SHP gene expression. However, UDCA-mediated FXR activation is much less than frequent CDCD (41). Consistent with these observations, SHP gene expression was unchanged following UDCA treatment in our study. Therefore, UDCA may improve hepatic lipid metabolism through a SHP-independent mechanism. Moreover, the use of CDCA and cholic acid is limited in humans because they can cause hepatotoxicity and increase low density lipoprotein cholesterol (42). UDCA decreases hepatocyte sensitivity to hydrophobic bile acid-induced oxidative stress in the fatty liver (43). Therefore, these observations suggest that UDCA-induced SMILE expression has a beneficial effect on the regulation of hepatic lipogenic gene expression without side effects as with CDAC and CA.
Previous reports have demonstrated that UDCA activates p38 MAPK, extracellular signal-regulated protein kinase (ERK), and PI3K pathways (25,40). In addition, CDCA also activates the p38 MAPK, c-Jun N-terminal kinase, and ERK pathways but not the PI3K pathway (31). However, the molecular signaling to repress lipogenic gene expression by UDCA has remained unclear. In this study, UDCA strongly increased AMPK phosphorylation, and UDCA-mediated SMILE gene expression was blocked by repression of AMPK signaling. However, the downstream effectors of AMPK signaling to induce SMILE gene expression remain unknown. We investigated the underlying mechanism of UDCA, particularly focusing on gene expression involved in lipid metabolism in hepatocytes. However, whether the AMPK pathway plays a major role in UDCAinduced gene expression of SMILE and UDCA-mediated improvement of fatty liver needs to be examined in an animal study. Future studies will reveal the extent to which UDCAinduced SMILE mediates various liver metabolism disorders.
In summary, we found that SMILE acts as a novel corepressor of LXR␣ by competing with coactivator SRC-1 to inhibit hepatic lipogenic gene expression. Moreover, UDCA also inhibited lipogenic gene expression depending on SMILE gene expression. Based on these findings, we suggest that the UDCA-mediated AMPK signaling pathway induces SMILE gene expression. Moreover, SMILE decreased LXR␣ activity via competition with SRC-1 and subsequently decreased LXR␣-mediated lipogenic enzyme gene expression in the liver (Fig. 8E). Thus, activating SMILE gene expression represents a potential therapeutic approach to improve hepatic lipid metabolism.