Dong-il Kim,1 Min-jung Park,1 Seul-ki Lim,2 Jae-il Park,3 Kyung-chul Yoon,4 Ho-jae Han,5

Jan-Åke Gustafsson,6,7 Jae-hyang Lim,8 and Soo-hyun Park1

PRMT3 Regulates HepaticLipogenesis Through DirectInteraction With LXRaDiabetes 2015;64:60–71 | DOI: 10.2337/db13-1394

Arginine methylation is responsible for diverse biologi-cal functions and is mediated by protein argininemethyltransferases (PRMTs). Nonalcoholic fatty liverdisease (NAFLD) is accompanied by excessive hepaticlipogenesis via liver X receptor a (LXRa). Thus we ex-amined the pathophysiological role of PRMTs in NAFLDand their relationship with LXRa. In this study, palmiticacid (PA) treatment increased PRMT3, which is corre-lated with the elevation of hepatic lipogenic proteins.The expression of lipogenic proteins was increased byPRMT3 overexpression, but decreased by PRMT3 si-lencing and use of the PRMT3 knockout (KO) mouseembryonic fibroblast cell line. PRMT3 also increasedthe transcriptional activity of LXRa by directly bindingwith LXRa in a methylation-independent manner. In ad-dition, PA treatment translocated PRMT3 to the nucleus.In animal models, a high-fat diet increased the LXRaand PRMT3 expressions and binding, which was notobserved in LXRa KO mice. Furthermore, increasedPRMT3 expression and its binding with LXRa were ob-served in NAFLD patients. Taken together, LXRa andPRMT3 expression was increased in cellular and mousemodels of NAFLD and human patients, and PRMT3translocated into the nucleus bound with LXRa as a tran-scriptional cofactor, which induced lipogenesis. In con-clusion, PRMT3 translocation by PA is coupled to thebinding of LXRa, which is responsible for the onset offatty liver.

Nonalcoholic fatty liver disease (NAFLD) is a worldwidemetabolic syndrome defined by an increased accumulationof fat, mainly triglycerides (TGs), within hepatocytes.NAFLD is closely related to diabetes because its patho-genesis is accompanied by hepatic insulin resistance andexcessive hepatic lipogenesis (1,2). Many studies havebeen conducted to verify the molecular mechanism oflipogenesis. However, the whole mechanism is highlycomplex and difficult to understand.

In the liver, liver X receptor a (LXRa) regulates SREBP1cand carbohydrate-responsive element-binding protein(ChREBP), which induce lipogenesis (3,4). LXRa that hasbeen activated by oxysterol, an endogenous ligand, and/orT0901317, an exogenous ligand, increases the expressionand transcriptional activity of SREBP1c and ChREBP andsubsequently increases the transcription of fatty acid syn-thase (FAS) and acetyl CoA carboxylase (ACC), which arelipogenic enzymes (5–7). The transcriptional activity ofLXRa is regulated by its interaction with its transcriptionalcofactor, which determines whether binding with LXRresponse element (LXRE) will occur (8,9). LXRa-deficientmice have shown markedly lower hepatic expression ofSREBP1c and its target genes (10). Moreover, T0901317-induced hepatic expression of ACC and FAS was attenu-ated in ChREBP-knockout (KO) mice (4). Many studieshave demonstrated the obvious role of LXRa in lipogen-esis through SREBP1c and/or ChREBP, but recent studies

1College of Veterinary Medicine, Chonnam National University, Gwangju, Republicof Korea2Metabolism and Functionality Research Group, R&D Division, World Institute ofKimchi, Gwangju, Republic of Korea3Korea Basic Science Institute, Gwangju Center at Chonnam National University,Gwangju, Republic of Korea4Department of Ophthalmology, Chonnam National University Medical School andHospital, Gwangju, Republic of Korea5Department of Verterinary Physiology, College of Veterinary Medicine, SeoulNational University, Seoul, Republic of Korea6Molecular Nutrition Unit, Department of Bioscience and Nutrition, KarolinskaInstitutet, Stockholm, Sweden

7Center for Nuclear Receptors and Cell Signaling, University of Houston, Houston, TX8Department of Microbiology, Ewha Womans University School of Medicine, Seoul,Republic of Korea

Corresponding author: Soo-hyun Park, parksh@chonnam.ac.kr.

Received 9 September 2013 and accepted 26 July 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1394/-/DC1.

© 2015 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

60 Diabetes Volume 64, January 2015

METABOLISM

also suggested the dispensable role of LXRa activationfor the SREBP1c- and ChREBP-dependent lipogenesis(11,12).

Protein arginine methyltransferases (PRMTs) arethought to be new biomarkers that regulate epigeneticevents and posttranslational modification (13). PRMTsmethylate histone and nonhistone proteins in arginineresidue. The nine mammalian forms of PRMTs can bedivided into three subtypes. Type I (PRMTS 1, 2, 3, 4,6, and 8) enzymes catalyze arginine residue to a monome-thylarginine intermediate, which is then converted toasymmetric dimethylarginine (ADMA).

In the past two decades, many studies have reportedthat increased ADMA production associated with nitricoxide synthase was related to the onset of heart failureand/or chronic kidney disease (14,15). However, becauseprotein methylation is one of the most abundant proteinmodifications and ADMA is catalyzed by type I PRMTs,current studies have focused on arginine methylation asa molecular signal implicated in various pathophysiologicaldiseases, including cancer and metabolic syndrome. Honget al. (16) recently reported the expression of PRMTs inmale rat tissues; they found that PRMT1, PRMT4, andPRMT5 were minimally expressed in the liver. However,several lines of evidence have shown the importance oftype I PRMTs in the liver, as they are responsible forhepatic glucose metabolism. Choi et al. (17) and Han et al.(18) reported that PRMT1 and PRMT6 regulate hepaticgluconeogenesis. Lee et al. (19) also demonstrated an in-creased ADMA plasma concentration and decreased hepaticPRMT1 expression in diabetic mice (db/db), suggestingthat PRMTs play an important role in the onset of diabetesin the liver. Krones-Herzig et al. (20) suggested that PRMT4is a critical component of cyclic adenosine monophosphate–dependent glucose metabolism in the liver.

Another important isoform of type I PRMT is PRMT3,which induces ADMA methylation. Its substrates are veryrestricted because of its predominant cytoplasmic location(21). The role of PRMT3 in dendritic spine maturationwas recently reported. However, the functional role ofPRMT3 in metabolic disorders has not been elucidated.Thus we examined the role of PRMT3 in hepatic lipogen-esis using a hepatocyte cell line, a high-fat diet animalmodel, and NAFLD patients.

RESEARCH DESIGN AND METHODS

Reporter Gene AssayCells were transiently transfected with indicated plasmidsin addition to b-galactosidase and luciferase plasmids.After 24 h, the culture medium was changed with mediumcontaining indicated concentrations of drugs and incu-bated for an additional 6 h. The cells were lysed withpassive lysis buffer (E194A; Promega), and then cellextracts were analyzed. Luminescence was analyzed withGloMax Microplate Reader. Final 1 mmol/L luciferin(E1602; Promega) was dissolved in luciferase assay buffer(1 mol/L KH2PO4, pH 7.8, 1 mol/L MgCl2, and 0.1 mol/L

ATP). b-Galactosidase plasmid was used as a control forthe normalization of transfection efficiency.

Chromatin ImmunoprecipitationChromatin immunoprecipitation assay was performed usingthe EZ-ChIP kit (Millipore, Billerica, MA) as instructed bythe manufacturer. Briefly, cultured cells were fixed with1% formaldehyde and harvested. Cell pellets were lysed inSDS lysis buffer, and sonication was performed to sharechromatin. Shared chromatins were precleared with proteinG agarose, then normal mouse IgG/LXRa antibody or nor-mal rabbit IgG/PRMT3 antibody was added. After 17 h,immune complex was incubated with protein G agarosefor 2 h. Precipitated chromatins were eluted, then protein/DNA crosslinking was reversed. Purified DNA was amplifiedby real-time quantitative PCR using the following primers;forward 59-GCCAGGACTTCTCTGCTTTGA-39 and reverse59-CTGAATGGGGTTGGGGTTACT-39 for2461/2326 regionof human SREBP1c promoter.

Human Liver TissueSurgical liver resections were performed from 10 patients(four men, six women; mean age, 60.4 years; range, 49–74years). Control nonfatty liver samples were collected fromfive patients undergoing hepatic resection for intrahepaticbile duct stone (n = 5). Fatty liver samples were collectedfrom fatty livers of five patients. All materials were usedunder the protocol approved by the Institutional ReviewBoard at Chonnam National University Hospital.

Proximity Ligation AssayParaffin sections were deparaffinized with HistoChoice(Amresco) and rehydrated with serial diluted ethanol.Sections were steamed in sodium citrate buffer (10 mmol/Lsodium citrate, 0.05% Tween 20, pH 6.0) for antigen retrievaland washed with PBS. Then proximity ligation assay (PLA)was performed using Duolink (Olink Biosciences, Uppsala,Sweden) according to the manufacturer’s instructions.Briefly, the sections were incubated with primary anti-bodies (anti-LXRa and anti-PRMT3) diluted to 1:100 inantibody diluents supplied for overnight at 4°C. TheDuolink PLA probes were incubated in preheated humiditychamber for 1 h at 37°C, followed by ligation, amplifica-tion. To reduce tissue autofluorescence, sections were in-cubated with copper sulfate (5mmol/L CuSO4 in 50mmol/Lammonium acetate buffer, pH 5.0) for 20 min (22),followed by mounting. The stained sections were observedusing Leica TCS SP5 AOBS laser scanning confocal micro-scope (Leica Microsystems, Heidelberg, Germany) using aLeica 633 (numerical aperture 1.4) oil objective located atKorea Basic Science Institute Gwangju Center.

Statistical AnalysisThe results were expressed as mean 6 SEM. Values aremean 6 SEM of three or four independent experiments.For two group comparisons, a Student t test was used, andfor multiple comparisons, one-way ANOVA by SPSS (SPSSInc., Chicago, IL), followed by the Tukey post hoc test, wasused. A value of P , 0.05 was considered significant.

diabetes.diabetesjournals.org Kim and Associates 61

RESULTS

Palmitic Acid Treatment Alters Arginine AsymmetricDimethylation Status and Increases PRMT3Expression and Hepatic LipogenesisAlthough the importance of protein arginine asymmetricdimethylation in various tissues is increasingly recog-nized, the functional role of hepatic arginine asymmetricdimethylation in hepatic lipogenesis is poorly understood.To confirm the changes in ADMAs induced by saturatedfatty acids, HepG2 cells were treated with 750 mmol/L pal-mitic acid (PA), followed by ASYM24 antibody, which recog-nizes ADMAs. Twenty-four hours after this PA treatment,many dynamic changes in protein arginine asymmetric

dimethylation were observed. Pretreatment with adenosine-29,39-dialdehyde (Adox), a global methylation inhibitor,blocked the arginine asymmetric dimethylation statusin the control and PA groups (Fig. 1A). This result sug-gests that type I PRMTs are associated with hepatocytefunction, as they induce arginine asymmetric dimeth-ylation. Thus we examined the effect of PA on type I PRMTexpression. PA treatment significantly increased PRMT3expression (Fig. 1B), and PA only slightly increasedPRMT1 expression and did not alter PRMT4 expression.PRMT1 was recently reported to have a functional role inhepatic gluconeogenesis (17). Based on this finding, weexamined the role of PRMT3 in hepatic lipogenesis. PA

Figure 1—PA changes the arginine asymmetric dimethylation status and increases PRMT3 expression and hepatic lipogenesis. A: Afterpretreatment with 30 nmol/L Adox for 30 min, HepG2 cells were treated with 750 mmol/L PA for 24 h. Cell extracts were subjected toWestern blot analysis with indicated antibodies and then normalized to the b-actin level. Data are mean 6 SEM of three independentexperiments. *P < 0.05 vs. control; †P < 0.05 vs. PA treatment. B: HepG2 cells were treated with 750 mmol/L PA for 12 and 24 h. Type IPRMTs were assayed by Western blot analysis with the indicated antibodies and then normalized to the b-actin level. Data are mean 6SEM of five independent experiments. *P < 0.05 vs. 0 h. C: HepG2 cells were treated with 750 mmol/L PA for the indicated time intervals.Cell extracts were subjected to Western blot analysis with the indicated antibodies and then normalized to the b-actin level. Data aremean 6 SEM of five independent experiments. *P < 0.05 vs. 0 h. D and E: After pretreatment with 30 nmol/L Adox for 30 min, HepG2 cellswere treated with 750 mmol/L PA for 24 h. D: Cultured cells were subjected to Oil Red O staining then counterstained with hematoxylin.Representative images were from at least three independent experiments. E: Cell extracts were subjected to Western blot analysis with theindicated antibodies and then normalized to the b-actin level. Data are mean6 SEM of five independent experiments. *P< 0.05 vs. control;†P < 0.05 vs. PA.

62 PRMT3, Hepatic Lipogenesis, and LXRa Diabetes Volume 64, January 2015

increased the level of PRMT3 mRNA (Supplementary Fig.1A), which is correlated with the levels of PRMT3 pro-teins. We also examined the time-dependent effect of PAon the expression of PRMT3, LXRa, and other lipogenicproteins. PA treatment increased PRMT3, LXRa, cleavedSREBP1c (cl-SREBP1c; which is an active form of SREBP1c[23]), FAS, and total ACC expression, but not peroxisomeproliferator-activated receptor g, ChREBP expression, andACC phosphorylation, in a time-dependent manner (Fig.1C and Supplementary Fig. 1B). These results suggest thatthe protein arginine methylation status is important inhepatic lipogenesis. Indeed, PA-induced lipid accumula-tion and expression of lipogenic proteins were blockedby treatment with Adox (Fig. 1D and E). However,ChREBP expression was not influenced by PA and Adoxtreatment (Supplementary Fig. 1C).

PRMT3 Regulates Hepatic Lipogenesis ParametersThrough LXRa SignalingTo reveal the functional role of PRMT3, transient trans-fection of PRMT3 was conducted in HEK293 cells. Over-expression of hemagglutinin (HA)-tagged PRMT3 increasedthe expression of lipogenic proteins, including LXRa, FAS,ACC, and cl-SREBP1c but did not alter the expressionof peroxisome proliferator-activated receptor g, ChREBP,p-ACC, and PEPCK (Fig. 2A and Supplementary Fig. 2A).These results suggest that PRMT3 increases lipogenicprotein expression. We examined this effect further usingPRMT3 KO mouse embryonic fibroblast (MEF) cells.We found markedly decreased expression of LXRa, cl-SREBP1c, FAS, and total ACC in KO cells comparedwith wild-type (WT) cells (Supplementary Fig. 2B). More-over, the PA-induced upregulation of lipogenic proteinsobserved in the WT MEF cells did not occur in thePRMT3 KO cells (Fig. 2B). Indeed, PRMT3 silencing bysmall interfering RNA (siRNA) transfection blocked PA-induced lipogenic proteins, including LXRa, cl-SREBP1c,FAS, and total ACC, in HepG2 cells (Fig. 2C). However,ChREBP and PEPCK expressions were not influenced byPRMT3 expression as well (Supplementary Fig. 2C and D).In addition, PA-induced lipid accumulation, fatty acid syn-thesis, and TG accumulation were blocked by PRMT3 si-lencing in HepG2 cells (Fig. 2D–F and Supplementary Fig.2E). LXRa promotes cholesterol efflux by increase of ATP-binding cassette (ABC) transporter isoforms, those areABCA1, ABCG1, ABCG5, and ABCG8 (24,25). However,PRMT3 silencing did not alter the expression of thesegenes and intracellular cholesterol level (SupplementaryFig. 2F–H). These results suggest that PRMT3 regulatesLXRa SREBP1c–dependent hepatic lipogenesis. To furtherconfirm this signaling axis, HA or HA-PRMT3 plasmid wastransfected to AML-12 cells, normal mouse hepatocyte,with scramble or LXRa siRNA. PRMT3-induced expressionof lipogenic proteins and lipid accumulation were abol-ished by LXRa silencing (Fig. 2G and Supplementary Fig.2I and J). These results suggest that LXRa is required forPRMT3-induced expression of lipogenic proteins.

PRMT3 Upregulates the Transcriptional Activity ofLXRaTo evaluate whether overexpression of PRMT3 increasesthe transcriptional activity of LXRa, we performed a re-porter gene assay using LXRE or native SREBP1c pro-moter (up to 0.9 Kb) containing luciferase plasmids inHepG2 cells. Treatment with T0901317, an exogenousLXRa ligand, increased the transcriptional activity ofLXRa (Fig. 3A and C). In addition, overexpression ofPRMT3 increased the transcriptional activity of LXRa.Treatment with T0901317 significantly increased the ac-tivity of LXRa in PRMT3-overexpressed cells comparedwith green fluorescent protein (GFP)-overexpressed cells.Moreover, PA treatment further increased the transcrip-tional activity of LXRa. Interestingly, Adox treatment didnot influence the transcriptional activity of LXRa inPRMT3-overexpressed cells (Fig. 3A and C). Similar resultwas observed in HEK293 cells (Supplementary Fig. 3A).These results suggest that PRMT3 overexpression in-creases LXRa activity, but not via PRMT3-mediated meth-ylation. For further confirmation, we conducted PRMT3siRNA transfection in HepG2 cells. As expected, silencingof PRMT3 attenuated the T0901317- and PA-inducedtranscriptional activity of LXRa (Fig. 3B and D). Theseresults suggest that PRMT3 is an intermediate moleculeto represent the effect of LXRa in hepatocytes.

To evaluate that PRMT3 could directly regulatetranscriptional activity of SREBP1c or ChREBP, HepG2cells were transfected with SREBP response element orcarbohydrate response element luciferase plasmids. How-ever, PRMT3 overexpression altered neither of theirtranscriptional activities (Supplementary Fig. 3B and C).

PRMT3 Binds with LXRa In Vitro and in Intact Cells andThen Acts as a Transcriptional CofactorTo verify that increased transcriptional activity of LXRa isinvolved in direct binding of PRMT3, we conducted a gluta-thione s-transferase (GST) pull-down assay. OverexpressedGST and GST-PRMT3 proteins in the BL21 strain of Escher-ichia coli were purified with glutathione sepharose beads.The purified proteins were then incubated with HEK293whole cell lysate, which is overexpressed with HA-taggedLXRa. LXRa strongly interacted with GST-PRMT3, but notwith GST alone (Fig. 4A). To elucidate whether the inter-action between PRMT3 and LXRa also occurs in intact cells,we cotransfected Flag-PRMT3 with HA or HA-LXRa inHEK293 cells. The whole cell lysate was then immunopre-cipitated with M2 antibody, which specifically recognizesFlag. Immunoprecipitation was reciprocally conductedwith HA antibody. As in the GST pull-down assay, weconfirmed that the interaction between PRMT3 andLXRa also occurred in intact cells (Fig. 4B).

PRMT3 is an enzyme that can methylate the arginineresidue of a substrate. To verify whether direct bindingbetween LXRa and PRMT3 occurs because of LXRamethylation, we performed an in vitro methylation assay.Purified GST-PRMT3 was incubated with HA–ribosomal

diabetes.diabetesjournals.org Kim and Associates 63

protein S2 (rpS2) or HA-PRMT3 extract purified froman overexpressed HEK293 cell lysate, and S-adenosyl-L-[methyl-3H]methionine (methyl donor) was thenadded. Purified GST-PRMT3 methylated rpS2, a PRMT3substrate, but not LXRa (Fig. 4C). To further confirm, weconducted chromatin immunoprecipitation with LXRaand PRMT3 antibodies. PA treatment recruited LXRaand PRMT3 to the LXREs of SREBP1c promoter region(Fig. 4D). However, PA-induced recruitment of LXRa was

abolished by PRMT3 silencing (Fig. 4D). Thus we suggestthat PRMT3 directly binds to LXRa, then acts as a tran-scriptional cofactor without LXRa methylation.

PA Treatment Induces the Translocation of PRMT3 tothe NucleusPRMT3 location is reportedly restricted to the cytosolfraction in some cell lines (21). PRMT3 might bind withLXRa in the nucleus because LXRa is a nuclear receptor.

Figure 2—PRMT3 regulates hepatic lipogenesis through LXRa signaling. A: HEK293 cells were transfected with HA or HA-PRMT3 plasmids.Twenty-four hours later, cultured cells were harvested and cell extracts were subjected to Western blot analysis with the indicated antibodiesand then normalized to the b-actin level. Data are mean 6 SEM of four independent experiments. *P < 0.05 vs. HA. B: WT and KO PRMT3MEF cells were treated with 750 mmol/L PA for 12 and 24 h. Cell extracts were subjected to Western blot analysis with the indicatedantibodies. Representative immunoblots were from at least three independent experiments. C: HepG2 cells were transfected with scramblesiRNA or PRMT3 siRNA following the reverse transfection method. Twenty-four hours later, cells were treated with 750 mmol/L PA for 12 and24 h. The cell extracts were then subjected to Western blot analysis with the indicated antibodies and then normalized to the b-actin level.Data are mean6 SEM of five independent experiments. *P< 0.05 vs. scramble + 0 h; †P< 0.05 vs. same time treatment of scramble group;‡P < 0.05 vs. siPRMT3 + 0 h. D–F: HepG2 cells were transfected with scramble siRNA or PRMT3 siRNA following the reverse transfectionmethod. Twenty-four hours later, the cells were treated with 750 mmol/L PA for 24 h. D: Cultured cells were subjected to Oil Red O stainingthen counterstained with hematoxylin. E: Cultured cells were subjected to free fatty acid assay. Presented data are representative of threeindependent experiments, and experiments were performed in triplication. Data represent mean6 SEM. *P< 0.05 vs. scramble; †P< 0.05 vs.scramble + PA. F: Cultured cells were subjected to TG assay. Presented data are representative of three independent experiments, andexperiments were performed in triplication. Data represent mean 6 SEM. *P < 0.05 vs. scramble; †P < 0.05 vs. scramble + PA. G: AML-12cells were transfected with scramble siRNA or LXRa siRNA following the reverse transfection method. Twenty-four hours later, the cells weretransfected with HA or HA-PRMT3 plasmids. Twenty-four hours later, cultured cells were harvested and cell extracts were subjected toWestern blot analysis with the indicated antibodies. Representative immunoblots were from at least three independent experiments, andrelative quantification of these immunoblots are presented in Supplementary Fig. 2I. scr, scramble.

64 PRMT3, Hepatic Lipogenesis, and LXRa Diabetes Volume 64, January 2015

To identify the colocalization of PRMT3 and LXRa in thenucleus, we cotransfected GFP-PRMT3 and HA-LXRain HEK293 cells, then conducted immunofluorescencewith anti–tetramethylrhodamine-5-(and 6)-isothiocyanateantibody for HA staining. Stained cells were observedwith confocal microscopy. Overexpressed GFP-PRMT3was located in the cytoplasm (Fig. 5A). However, PAtreatment increased the nuclear location of PRMT3 andcolocalization with LXRa, despite the presence of somecytosolic PRMT3 (data not shown). The experiment usingan HA-tagged PRMT3 construct stained with anti–fluoresceinisothiocyanate antibody also validated the fact that

the nuclear expression of PRMT3 was increased by PAtreatment in HEK293 cells (Supplementary Fig. 4A).These results suggest that in spite of the domi-nant location of PRMT3 in the cytosol, PA treatmentcan increase the translocation of exogenous PRMT3 tothe nucleus.

To investigate whether nuclear translocation of endog-enous hepatic PRMT3 occurs in response to PA, weconducted immunofluorescence staining for the endoge-nous PRMT3 in HepG2 cells. Endogenous PRMT3 wasstained with anti–fluorescein isothiocyanate antibody.When treated with vehicle, PRMT3 was located in the

Figure 3—PRMT3 upregulates the transcriptional activity of LXRa. HepG2 cells were cotransfected with HA/HA-LXRa and HA-RxRa, GFP/GFP-PRMT3, LXRE-Luc (A) or HA/HA-PRMT3 and native SREBP1c promoter-Luc (C) and b-galactosidase plasmids. Twenty-four hourslater, after pretreatment with 30 nmol/L Adox or 750 mmol/L PA for 30 min, 10 mmol/L T0901317 was added for 6 h. Luciferase activity wasmeasured and normalized to b-galactosidase activity. Presented data are representative of three independent experiments, and experi-ments were performed in triplication. Data represent mean6 SEM. A: *P< 0.05 vs. lane 1; †P< 0.05 vs. lane 2; ‡P< 0.05 vs. lane 4; §P<0.05 vs. lane 3; || P< 0.05 vs. lane 5; ¶P < 0.05 vs. lane 6. C: *P < 0.05 vs. lane 1; †P < 0.05 vs. lane 2; ‡P< 0.05 vs. lane 3; §P< 0.05 vs.lane 4; || P < 0.05 vs. lane 5. HepG2 cells were transfected with scramble siRNA or PRMT3 siRNA following the reverse transfectionmethod. Twenty-four hours later, the cells were cotransfected with HA/HA-LXRa and HA-RxRa, LXRE-Luc (B) or native SREBP1cpromoter-Luc (D) and b-galactosidase plasmids. Twenty-four hours later, after pretreatment with 750 mmol/L PA for 30 min, 10 mmol/LT0901317 was added for 6 h. Luciferase activity was measured and normalized to b-galactosidase activity. Presented data are represen-tative of three independent experiments, and experiments were performed in triplication. Data represent mean6 SEM. B: *P< 0.05 vs. lane1; †P< 0.05 vs. lane 2; ‡P< 0.05 vs. lane 3; §P< 0.05 vs. same treatment of scramble group. D: *P< 0.05 vs. lane 1; †P< 0.05 vs. lane 2;‡P < 0.05 vs. lane 3; §P < 0.05 vs. same treatment of scramble group. E.V., empty vector; n.s., nonspecific; scr, scramble.

diabetes.diabetesjournals.org Kim and Associates 65

cytoplasm and nucleus (Fig. 5B and SupplementaryFig. 4B). However, PA treatment increased the nuclearaccumulation of PRMT3. Interestingly, T0901317 treat-ment recruited almost all cytosolic PRMT3 to the nucleus.These results show that endogenous hepatic PRMT3 islocated not only in the cytoplasm, but also in the nucleus,and that PA and the exogenous LXRa ligand increase thenuclear accumulation of PRMT3 in HepG2 cells.

To further confirm these results, HepG2 cells weretreated with PA, and the nuclear fraction was thenextracted. Similar to the immunofluorescence results, PAtreatment simultaneously increased nuclear PRMT3,LXRa, and nuclear SREBP1c expression (Fig. 5C).

Expression and Binding of LXRa and PRMT3 AreIncreased in the Livers of Mice Fed a High-Fat DietTo validate that this event also takes place in mousemodels, we fed mice a 60% high-fat diet for 12 weeks. Asexpected, high-fat-fed mice exhibited increased hepaticPRMT3 and LXRa expression, as well as lipogenic proteinexpression, compared with mice on a normal chow diet(Fig. 6A). Furthermore, PRMT3 and LXRa expression in-creased in the nuclear fraction isolated from the livers ofhigh-fat-fed mice (Fig. 6B). To examine whether bindingbetween LXRa and PRMT3 manifests in mouse models,we conducted immunoprecipitation with PRMT3 antibodyfrom mouse liver lysates. Increased binding between

Figure 4—PRMT3 binds with LXRa and acts as a transcriptional cofactor. A: HEK293 cells were transfected with HA-LXRa plasmid.Twenty-four hours later, total proteins were extracted and incubated with purified GST/GST-PRMT3–fused glutathione sepharose beadsovernight. In vitro binding was assessed with LXRa antibody. Purified GST/GST-PRMT3 proteins were confirmed by Coomassie bluestaining, and transfection efficacy was assessed with HA antibody from HEK293 lysate. B: HEK293 cells were cotransfected with Flag-PRMT3 and HA/HA-LXRa plasmids. Twenty-four hours later, total proteins were extracted, and reciprocal coimmunoprecipitation wasconducted with Flag or HA antibodies and immunoblotted with the indicated antibodies. Expression of Flag, HA, and PRMT3 from 5% inputwere analyzed by immunoblotting. C: HEK293 cells were transfected with HA-rpS2 or HA-LXRa plasmids. Twenty-four hours later, totalproteins were extracted and in vitro methylation assay was performed following the method introduced in RESEARCH DESIGN AND METHODS.D: HepG2 cells were transfected with scramble siRNA or PRMT3 siRNA following the reverse transfection method. Twenty-four hours later,the cells were treated with 750 mmol/L PA. Eight hours later, the cells were treated with final 1% of formaldehyde then chromatinimmunoprecipitation assay was performed with LXRa or PRMT3 antibodies following the method introduced in RESEARCH DESIGN AND METHODS.*P < 0.05 vs. scramble; †P < 0.05 vs. scramble + PA. IP, immunoprecipitation; WB, Western blotting.

66 PRMT3, Hepatic Lipogenesis, and LXRa Diabetes Volume 64, January 2015

LXRa and PRMT3 was observed in high-fat-fed micecompared with mice on a normal chow diet (Fig. 6C).However, this binding was not observed in high-fat-fedLXRa-deficiency mice (Fig. 6C), and LXRa deficiency didnot influence high-fat diet–induced PRMT3 expression(Fig. 6C; input). These results suggest that a high-fatdiet increases lipogenesis via increased PRMT3 andLXRa expression and binding in mouse models.

PRMT3 Expression and Binding of PRMT3 and LXRaAre Increased in the Livers of NAFLD PatientsTo confirm that increased PRMT3 expression and bindingof PRMT3 and LXRa are observed in human liver, we usedthe liver tissues from nonfatty liver patients and NAFLD

patients. As expected, dramatically increased PRMT3 ex-pression was observed in liver from NAFLD patients com-pared with liver from nonfatty liver patients (Fig. 7A). Inaddition, lots of PRMT3-positive cells were stained in thenucleus (Fig. 7A; red arrow). Binding between PRMT3 andLXRa, determined by PLA, was also increased in NAFLDpatients (Fig. 7B). These results suggest that increasedhepatic PRMT3 expression and binding with LXRa pro-mote the progression of NAFLD in humans.

DISCUSSION

NAFLD, an outstanding example of the complex patho-physiology of metabolic syndrome, is considered to be

Figure 5—PA treatment induces the translocation of PRMT3 to the nucleus. A: HEK293 cells were cotransfected with GFP-PRMT3 and HA-LXRa plasmids. Twenty-four hours later, the cells were treated with 750 mmol/L PA for 12 h then fixed, immunostained with HA antibody,and evaluated. Green, GFP-PRMT3; red, HA-LXRa; blue, DAPI. B: HepG2 cells were treated with 750 mmol/L PA or 10 mmol/L T0901317 for12 h then fixed, immunostained with PRMT3 antibody, and evaluated. Green, endogenous PRMT3; blue, DAPI. C: HepG2 cells were treatedwith 750 mmol/L PA for 12 and 24 h. Nuclear proteins were extracted then subjected to Western blot analysis with the indicated antibodiesand finally normalized to the lamin B level. Data are mean 6 SEM of three independent experiments. *P < 0.05 vs. 0 h. DIC, differentialinterference contrast; Pal, palmitic acid.

Figure 6—Expression and binding of LXRa and PRMT3 are increased in the livers of mice fed a high-fat diet.A: C57BL/6J mice were challengedwith a normal or high-fat diet for 12 weeks. Total proteins were extracted from the liver, and tissue extracts were subjected to Western blotanalysis with the indicated antibodies. Representative immunoblots were from at least three independent experiments. B: Nuclear proteins wereextracted from the liver, subjected to Western blot analysis with the indicated antibodies, and then normalized to the lamin B level. Data aremean6 SEM of three independent experiments. *P< 0.05 vs. normal diet. C: LXRa WT and KO mice were challenged with a normal or high-fatdiet for 12 weeks. Endogenous PRMT3 in the tissue extracts was immunoprecipitated with PRMT3 antibody and immunoblotted with LXRaantibody. Expression of PRMT3 and b-actin from 5% input were analyzed by immunoblotting. Representative immunoblots were from at leastthree independent experiments. HFD, high-fat diet; IP, immunoprecipitation; ND, normal diet; WB, Western blotting.

diabetes.diabetesjournals.org Kim and Associates 67

among the most common liver diseases worldwide.However, to our knowledge, hepatic ADMA status andtype I PRMTs in hepatic lipid metabolism have not yetbeen assessed. In this study, we first showed that PAtreatment alters ADMA status and type I PRMTs inhepatocytes. Interestingly, Adox treatment markedly re-duced the PA-induced ADMA status and lipogenesis,including lipid accumulation and related lipogenic mole-cules, indicating that type I PRMTs usually play importantroles in hepatic lipid metabolism. Our study revealed thatPA also slightly increased the expression of PRMT1, whichis consistent with a previous report on hepatic glucosemetabolism (17). However, in this study, we focused onthe as-yet unidentified role of PRMT3 in lipid metabolismin hepatocytes because PA markedly increased the expres-sion of PRMT3 in type I PRMTs. Based on these results,we hypothesized that asymmetric dimethylation byPRMT3 would be involved in high-fat–induced hepatic

lipogenesis. This hypothesis was supported by severalfindings: 1) PA activated lipogenesis and PRMT3 expres-sion in hepatocytes, 2) overexpression of HA-PRMT3induced lipogenesis, 3) PA-induced lipogenesis wasnot observed in the PRMT3 KO MEF cell line, and4) PRMT3 siRNA reduced PA-induced lipid accumulationin hepatocytes.

LXRs are known regulators of fatty acid metabolism invarious cells, including hepatocytes, adipocytes, andpancreatic b-cells (26–28). Recent studies have revealedthat high-fat diets increase the expression of lipogenicproteins via LXRa in the liver (26,29,30). In the currentstudy, we also confirmed that PA treatment increases notonly LXRa expression, but also lipid accumulation inHepG2 cells, which is consistent with the finding ofKuang et al. (31) that PA stimulates LXR activities.LXRa is considered to be a master of lipogenesis in theliver because of its functional role of SREBP1c and

Figure 7—PRMT3 expression and binding of PRMT3 and LXRa are increased in the livers of NAFLD patients. A and B: Human liver tissuesfrom nonfatty liver patients and NAFLD patients were prepared as introduced in RESEARCH DESIGN AND METHODS. A: Immunohistochemistryusing paraffin sections was performed with PRMT3 antibody. Representative images were from at least three independent experiments.(yellow arrows, PRMT3-positive cells; red arrows, PRMT3-positive cells stained in nucleus) B: PLA using paraffin sections was performedwith anti-rabbit PRMT3 antibody and anti-mouse LXRa antibody (green fluorescence, binding of PRMT3 and LXRa). Representative imageswere from at least three independent experiments. C: Summarized scheme. DIC, differential interference contrast; HFD, high-fat diet.

68 PRMT3, Hepatic Lipogenesis, and LXRa Diabetes Volume 64, January 2015

ChREBP regulation (32). However, the current studyrevealed that PRMT3 regulates only the expression oflipid metabolism molecules, such as LXRa and SREBP1c,but not ChREBP or PEPCK. This result is supported bya report showing that glucose is required for ChREBPfunctional activity and that LXRs are not necessary forthe induction of glucose-regulated genes in the liver, sug-gesting that ChREBP and LXRs are independent (12).LXRa also regulates cholesterol efflux through its receptorexpression (25). However, in this study, expression ofABC transporter isoforms and intracellular cholesterollevel were not altered by PRMT3 silencing. These resultssuggest that effect of PRMT3 on LXRa activity has veryselective targets such as SREBP1c but not ChREBP or ABCtransporters. Until now, no report has shown that PRMTis an upstream regulator of LXRa activation in the liver.The current study showed that PRMT3 overexpressionincreases SREBP1c native promoter activity and that si-lence of PRMT3 decreases the binding of LXRa to theSREBP1c promoter region. Based on these results, wesuggest that PA treatment induces hepatic lipogenesisvia the PRMT3-LXRa-SREBP1c axis.

Striking observations in this study were that PRMT3bound with LXRa in vitro and in vivo and increasedLXRE-dependent transcriptional activity. However, Adoxtreatment did not prevent the transcriptional activity ofLXRa. Moreover, LXRa was not methylated by PRMT3.These results suggest that the PRMT3-induced transcrip-tional activity of LXRa is methylation independent. Al-though LXRa has a few RxR motifs (arginine–x aminoacid–arginine) that often act as the methylation sites ofPRMT substrates (33,34), PRMT1 and PRMT4 did notmethylate LXRa in the current study (data not shown),suggesting that type I PRMTs do not directly methylateLXRa. Furthermore, we observed that PRMT3 binds withLXRa in the livers of mice on a high-fat diet and NAFLDpatients. Based on our results, we speculate that methyl-ation is an important modification of hepatic lipogenesis.However, upregulation of LXRa transcriptional activity byPRMT3 is not associated with the enzymatic activity ofPRMT3. LXRa activation induced by T0901317 has beenknown to increase the transcription of its target genes viatrimethylation of H3K4 (35,36). In this study, as PA-induced lipogenesis was blocked by Adox, it may be specu-lated that reduced trimethylation of H3K4 by Adox islinked to the diminished lipogenesis. Further studiesshould be performed to reveal this speculation.

PRMT3 has limited substrates because its location isrestricted (21). rpS2, which balances between ribosomalsubunits, is a well-known substrate of PRMT3. Most stud-ies of PRMT3 have revealed that it is located exclusivelywithin the cytoplasm (37–39). Consistent with thesereports, we observed an overwhelming cytoplasmic por-tion of GFP-PRMT3. However, we observed that PA treat-ment caused some translocation of PRMT3 to the nucleus,while a large portion remained in the cytoplasm. LikeGFP, HA-PRMT3 was observed in the nucleus after PA

treatment, ruling out the possible problem of GFP tag-ging. Moreover, we found that endogenous PRMT3 wasrecruited to the nucleus by treatment of HepG2 cells withT0901317. These results suggest that the coupling ofPRMT3 and LXRa is very important for lipogenesis inthe liver. To our knowledge, evidence of the location ofPRMT3 in the nucleus is restricted to a report by Tavanezet al. (40), which stated that PRMT3 accumulates in intra-nuclear inclusions of muscle in patients with oculopha-ryngeal muscular dystrophy. These authors did notevaluate the translocation of PRMT3. The current studyprovides novel evidence that PRMT3, which is known tobe restricted to the cytosolic fraction, translocates to thenucleus in the hyperlipidemic condition. Furthermore, theresult of chromatin immunoprecipitation assay, per-formed with PRMT3 antibody, is another strong evidencethat PRMT3 exists in the nucleus. PRMT proteins werefound to be implicated in epigenetic events by methylat-ing arginine residue of histone tail (41). However, histonearginine methylation by PRMT3 has not been discoveredyet. A recent study reported that LXRa target gene ex-pression is regulated by histone modification (36). Ourpresent findings along with previous reports may suggestthat LXRa targets genes selectively being introduced bynuclear PRMT3-mediated histone modification. Furtherstudies on the possible mechanisms by which PRMT3selectively increases LXRa target genes by modifying his-tone will bring new insight into the roles of PRMT3.

LXRa agonists, such as T0901317 and GW3965, havebeen considered to be potent therapeutic agents againstatherosclerosis, obesity, and even cancer (42,43). How-ever, because of their side effects, including increased se-rum TG levels secondary to hepatic lipogenesis (44), theiruse has been limited. Many studies have been conductedto determine how to overcome these side effects (45). Inthis study, we found that PRMT3 regulates LXRa activity.We suggest that the targeting of PRMT3 could be a newtherapeutic approach to reduce the side effects of LXRa.

In conclusion, as described in Fig. 7C, we showed that1) PA treatment increased PRMT3 expression and hepaticlipogenesis, 2) PRMT3 increased LXRa activity and itstarget gene expression by direct binding in a methylation-independent manner, 3) PA treatment increased the trans-location of PRMT3 from the cytosol to the nucleus, and 4)PRMT3 and LXRa expression and binding increased inhigh-fat-fed mice and NAFLD patients. Our findings pro-vide novel evidence that PRMT3 regulates hepatic lipogen-esis via interaction with LXRa. Targeting of PRMT3 isa potential approach to the treatment of NAFLD and pre-vention of the side effects of the LXRa agonist.

Acknowledgments. The authors thank Mark T. Bedford (The Universityof Texas MD Anderson Cancer Center, Smithville, TX) for permitting to use PRMTconstructs and PRMT3 KO MEF cell lines. HA-PRMT constructs were kindlyprovided by Akiyoshi Fukamizu (Life Science Center of Tsukuba Advanced Re-search Alliance, University of Tsukuba, Japan).

diabetes.diabetesjournals.org Kim and Associates 69

Funding. This work was supported financially by a basic science researchgrant from the National Research Foundation of Korea (2010-0023627). Theanimal experiment in this study was partially supported by the Animal MedicalInstitute of Chonnam National University.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. D.-i.K. contributed to experimental design, per-formed the experiments, researched data, wrote and reviewed the manuscript,and contributed to the revision of the manuscript. M.-j.P. performed the animalexperiments, researched data, and contributed to discussion and revision of themanuscript. S.-k.L. researched data and contributed to discussion. J.-i.P. per-formed the confocal microscope experiments and contributed to discussion. K.-c.Y.,H.-j.H., and J.-Å.G. researched data and contributed to the critical review andrevision of the manuscript. J.-h.L. contributed to experimental design and con-tributed to the critical review and revision of the manuscript. S.-h.P. researcheddata and contributed to experimental design and critical review and revision ofthe manuscript. S.-h.P. is the guarantor of this work and, as such, had full accessto all the data in the study and takes responsibility for the integrity of the data andthe accuracy of the data analysis.

References1. Samuel VT, Liu ZX, Qu X, et al. Mechanism of hepatic insulin resistancein non-alcoholic fatty liver disease. J Biol Chem 2004;279:32345–323532. Diraison F, Moulin P, Beylot M. Contribution of hepatic de novo lipogenesisand reesterification of plasma non esterified fatty acids to plasma triglyceridesynthesis during non-alcoholic fatty liver disease. Diabetes Metab 2003;29:478–4853. Repa JJ, Liang G, Ou J, et al. Regulation of mouse sterol regulatoryelement-binding protein-1c gene (SREBP-1c) by oxysterol receptors, LXRalphaand LXRbeta. Genes Dev 2000;14:2819–28304. Cha JY, Repa JJ. The liver X receptor (LXR) and hepatic lipogenesis. Thecarbohydrate-response element-binding protein is a target gene of LXR. J BiolChem 2007;282:743–7515. Grefhorst A, Elzinga BM, Voshol PJ, et al. Stimulation of lipogenesis bypharmacological activation of the liver X receptor leads to production of large,triglyceride-rich very low density lipoprotein particles. J Biol Chem 2002;277:34182–341906. Talukdar S, Hillgartner FB. The mechanism mediating the activation ofacetyl-coenzyme A carboxylase-alpha gene transcription by the liver X receptoragonist T0-901317. J Lipid Res 2006;47:2451–24617. Joseph SB, Laffitte BA, Patel PH, et al. Direct and indirect mechanisms forregulation of fatty acid synthase gene expression by liver X receptors. J BiolChem 2002;277:11019–110258. Willy PJ, Umesono K, Ong ES, Evans RM, Heyman RA, Mangelsdorf DJ. LXR,a nuclear receptor that defines a distinct retinoid response pathway. Genes Dev1995;9:1033–10459. Hu X, Li S, Wu J, Xia C, Lala DS. Liver X receptors interact with corepressorsto regulate gene expression. Mol Endocrinol 2003;17:1019–102610. Peet DJ, Turley SD, Ma W, et al. Cholesterol and bile acid metabolism areimpaired in mice lacking the nuclear oxysterol receptor LXR alpha. Cell 1998;93:693–70411. Yecies JL, Zhang HH, Menon S, et al. Akt stimulates hepatic SREBP1c andlipogenesis through parallel mTORC1-dependent and independent pathways. CellMetab 2011;14:21–3212. Denechaud PD, Bossard P, Lobaccaro JM, et al. ChREBP, but not LXRs, isrequired for the induction of glucose-regulated genes in mouse liver. J Clin Invest2008;118:956–96413. Bedford MT, Clarke SG. Protein arginine methylation in mammals: who,what, and why. Mol Cell 2009;33:1–1314. Usui M, Matsuoka H, Miyazaki H, Ueda S, Okuda S, Imaizumi T. Increasedendogenous nitric oxide synthase inhibitor in patients with congestive heartfailure. Life Sci 1998;62:2425–2430

15. Chan NN, Chan JC. Asymmetric dimethylarginine (ADMA): a potential linkbetween endothelial dysfunction and cardiovascular diseases in insulin re-sistance syndrome? Diabetologia 2002;45:1609–161616. Hong E, Lim Y, Lee E, Oh M, Kwon D. Tissue-specific and age-dependentexpression of protein arginine methyltransferases (PRMTs) in male rat tissues.Biogerontology 2012;13:329–33617. Choi D, Oh KJ, Han HS, et al. Protein arginine methyltransferase 1 regulateshepatic glucose production in a FoxO1-dependent manner. Hepatology 2012;56:1546–155618. Han HS, Jung CY, Yoon YS, et al. Arginine methylation of CRTC2 is critical inthe transcriptional control of hepatic glucose metabolism. Sci Signal 2014;7:ra1919. Lee JH, Park GH, Lee YK, Park JH. Changes in the arginine methylation oforgan proteins during the development of diabetes mellitus. Diabetes Res ClinPract 2011;94:111–11820. Krones-Herzig A, Mesaros A, Metzger D, et al. Signal-dependent control ofgluconeogenic key enzyme genes through coactivator-associated argininemethyltransferase 1. J Biol Chem 2006;281:3025–302921. Tang J, Gary JD, Clarke S, Herschman HR. PRMT 3, a type I protein arginineN-methyltransferase that differs from PRMT1 in its oligomerization, subcellularlocalization, substrate specificity, and regulation. J Biol Chem 1998;273:16935–1694522. Potter KA, Simon JS, Velagapudi B, Capadona JR. Reduction of auto-fluorescence at the microelectrode-cortical tissue interface improves antibodydetection. J Neurosci Methods 2012;203:96–10523. Zhao X, Feng D, Wang Q, et al. Regulation of lipogenesis by cyclin-dependent kinase 8-mediated control of SREBP-1. J Clin Invest 2012;122:2417–242724. Zhao C, Dahlman-Wright K. Liver X receptor in cholesterol metabolism.J Endocrinol 2010;204:233–24025. Zhang Y, Breevoort SR, Angdisen J, et al. Liver LXRa expression is crucialfor whole body cholesterol homeostasis and reverse cholesterol transport inmice. J Clin Invest 2012;122:1688–169926. Jwa H, Choi Y, Park UH, Um SJ, Yoon SK, Park T. Piperine, an LXRa an-tagonist, protects against hepatic steatosis and improves insulin signaling in micefed a high-fat diet. Biochem Pharmacol 2012;84:1501–151027. Liang H, Zhong Y, Zhou S, Li QQ. Palmitic acid-induced apoptosis in pan-creatic b-cells is increased by liver X receptor agonist and attenuated by eico-sapentaenoate. In Vivo 2011;25:711–71828. Stenson BM, Rydén M, Steffensen KR, et al. Activation of liver X receptorregulates substrate oxidation in white adipocytes. Endocrinology 2009;150:4104–411329. Han CY, Ki SH, Kim YW, et al. Ajoene, a stable garlic by-product, inhibitshigh fat diet-induced hepatic steatosis and oxidative injury through LKB1-dependent AMPK activation. Antioxid Redox Signal 2011;14:187–20230. Uemura T, Goto T, Kang MS, et al. Diosgenin, the main aglycon of fenu-greek, inhibits LXRa activity in HepG2 cells and decreases plasma and hepatictriglycerides in obese diabetic mice. J Nutr 2011;141:17–2331. Kuang YL, Paulson KE, Lichtenstein AH, Lamon-Fava S. Regulation of theexpression of key genes involved in HDL metabolism by unsaturated fatty acids.Br J Nutr 2012;108:1351–135932. Baranowski M. Biological role of liver X receptors. J Physiol Pharmacol2008;59(Suppl. 7):31–5533. Smith JJ, Rücknagel KP, Schierhorn A, et al. Unusual sites of argininemethylation in Poly(A)-binding protein II and in vitro methylation by protein ar-ginine methyltransferases PRMT1 and PRMT3. J Biol Chem 1999;274:13229–1323434. Wada K, Inoue K, Hagiwara M. Identification of methylated proteins byprotein arginine N-methyltransferase 1, PRMT1, with a new expression cloningstrategy. Biochim Biophys Acta 2002;1591:1–1035. Lee S, Lee J, Lee SK, Lee JW. Activating signal cointegrator-2 is an es-sential adaptor to recruit histone H3 lysine 4 methyltransferases MLL3 and MLL4to the liver X receptors. Mol Endocrinol 2008;22:1312–1319

70 PRMT3, Hepatic Lipogenesis, and LXRa Diabetes Volume 64, January 2015

36. Yu H, Wu J, Yang M, et al. Involvement of liver X receptor alpha in histonemodifications across the target fatty acid synthase gene. Lipids 2012;47:249–25737. Swiercz R, Person MD, Bedford MT. Ribosomal protein S2 is a substrate formammalian PRMT3 (protein arginine methyltransferase 3). Biochem J 2005;386:85–9138. Swiercz R, Cheng D, Kim D, Bedford MT. Ribosomal protein rpS2 is hy-pomethylated in PRMT3-deficient mice. J Biol Chem 2007;282:16917–1692339. Handrkova H, Petrak J, Halada P, Pospisilova D, Cmejla R. Tyrosine 87 isvital for the activity of human protein arginine methyltransferase 3 (PRMT3).Biochim Biophys Acta 2011;1814:277–28240. Tavanez JP, Bengoechea R, Berciano MT, Lafarga M, Carmo-Fonseca M,Enguita FJ. Hsp70 chaperones and type I PRMTs are sequestered at intranuclearinclusions caused by polyalanine expansions in PABPN1. PLoS ONE 2009;4:e6418

41. Di Lorenzo A, Bedford MT. Histone arginine methylation. FEBS Lett 2011;585:2024–203142. Joseph SB, McKilligin E, Pei L, et al. Synthetic LXR ligand inhibits thedevelopment of atherosclerosis in mice. Proc Natl Acad Sci U S A 2002;99:7604–760943. Chuu CP, Hiipakka RA, Kokontis JM, Fukuchi J, Chen RY, Liao S. Inhibitionof tumor growth and progression of LNCaP prostate cancer cells in athymic miceby androgen and liver X receptor agonist. Cancer Res 2006;66:6482–648644. Gao M, Bu L, Ma Y, Liu D. Concurrent activation of liver X receptor andperoxisome proliferator-activated receptor alpha exacerbates hepatic steatosis inhigh fat diet-induced obese mice. PLoS ONE 2013;8:e6564145. Kratzer A, Buchebner M, Pfeifer T, et al. Synthetic LXR agonist attenuatesplaque formation in apoE-/- mice without inducing liver steatosis and hyper-triglyceridemia. J Lipid Res 2009;50:312–326

diabetes.diabetesjournals.org Kim and Associates 71