Lanthionine synthetase C-like protein 2 (LanCL2) is important for adipogenic differentiation[S]

Adipogenic differentiation is a highly regulated process that is necessary for metabolic homeostasis and nutrient sensing. The expression of PPARγ and the subsequent activation of adipogenic genes is critical for the process. In this study, we identified lanthionine synthetase C-like protein 2 (LanCL2) as a positive regulator of adipogenesis in 3T3-L1 cells. Knockdown of LanCL2, but not LanCL1, inhibited adipogenic differentiation, and this effect was not mediated through cAMP or Akt signaling pathways. The expression of early adipogenic markers CCAAT enhancer binding protein β (C/EBPβ) and C/EBPδ remained intact in LanCL2 knockdown cells, but levels of late adipogenic markers PPARγ and C/EBPα were suppressed. The addition of the naturally occurring PPARγ activator 15-deoxy-Δ12,14-prostaglandin J2 or conditioned medium from differentiating cells did not restore differentiation, implying that LanCL2 may not be involved in the production of a secreted endogenous PPARγ ligand. Pulldown assays demonstrated a direct physical interaction between LanCL2 and PPARγ. Consistent with a regulatory role of LanCL2, luciferase reporter assays revealed that full transcriptional activation by PPARγ was dependent on LanCL2. Taken together, our study reveals a novel role of LanCL2 in adipogenesis, specifically involved in PPARγ-mediated transactivation of downstream adipogenic genes.

an endogenous ligand. Ligand binding induces a conformational change in the receptor, resulting in the dissociation of repressors and the differential recruitment of coactivators. This process is followed by the transcriptional activation of PPAR-target genes and marks the final stage of adipogenic differentiation, whereby the mRNA and protein levels for several enzymes involved in triglyceride metabolism, such as adipocyte protein 2 (aP2) and stearoyl-CoA desaturase, increase severalfold and result in the accumulation of fat (10)(11)(12).
Lanthionine synthetase C-like proteins (LanCLs) are eukaryotic homologs of bacterial LanC, a cyclase involved in the formation of lanthionine rings in lantibiotics (13). Humans have three LanCL proteins, LanCL1, LanCL2, and LanCL3, encoded on chromosomes 2, 7, and X, respectively. LanCL1 is involved in antioxidant responses in the brain-KO animals accumulate lipid, protein, and DNA damage with age and show evidence of mitochondrial dysfunction and apoptotic degeneration (14). LanCL2 has been shown to be involved in the abscisic acid (ABA) signaling pathway, acting downstream to propagate ABA-specific effects in immune and insulinoma cells (15). ABA was shown to increase PPAR reporter activity in RAW 264.7 macrophages, and knockdown (KD) studies showed LanCL2 expression to be necessary for this activity (16). Our previous results reported LanCL2 to be a novel regulator of Akt, specifically involved in facilitating optimal phosphorylation of Akt by mammalian target of rapamycin complex 2 (mTORC2), through direct physical interaction with both the kinase and the substrate (17). LanCL2 is highly expressed in the testis and the brain, with lesser, albeit ubiquitous expression in all other tissues examined (18). Very little is known about LanCL3. In this study, we found a novel role of LanCL2 in regulating the adipogenic differentiation process, acting at the stage of PPAR-mediated transactivation of adipogenic genes.

Plasmids
The following plasmids were obtained from Addgene (Cambridge, MA): PPAR response element (PPRE)-X3-TK-firefly luciferase (plasmid no. 1015) (19); pIS1, a derivative of pRL-TK containing a renilla luciferase reporter driven by the HSV TK promoter (plasmid no. 12179; gift from David Bartel); pcDNA FLAG-PPAR (plasmid no. 8895) (20,21); and pcDNA-enhanced green fluorescent protein (pcDNA eGFP; plasmid no. 13031; gift from Doug Golenbock). For LanCL2 overexpression studies, human LanCL2 cDNA was subcloned into p3×FLAG-CMV-14 vector (Sigma-Aldrich), with the 3×FLAG tag fused to the C terminus of LanCL2, as reported previously (17). For bacterial expression of LanCL2-His, human LanCL2 cDNA was subcloned into a pET-Duet-1 vector using NcoI and NotI cloning sites (Novagen, Madison, WI) in multiple cloning site I. A DNA sequence encoding a His 6 -tag was incorporated into the reverse primer of human LanCL2 cDNA. For bacterial expression of MBP-PPAR, mouse PPAR2 cDNA from FLAG-PPAR2 plasmid was subcloned into a pMAL-p2x vector (New England Biolabs, Beverly, MA) using BamHI and HindIII restriction sites. For localization experiments, human LanCL2 cDNA was cloned into the pcDNA-eGFP backbone using XhoI and XbaI cloning sites, with the eGFP tag fused to the C terminal end of LanCL2 generating hLanCL2-eGFP. A list of primers used for generating all the constructs is provided in supplemental Table S1.

Cell culture, differentiation, and nucleofection
The 3T3-L1 cells were maintained in DMEM with 10% FCS and 1× penicillin/streptomycin, at 37°C in the presence of 5% CO 2 . For adipogenic differentiation, 2 days after the cultures were confluent, cells were incubated with DMEM containing 10% FBS, 0.5 mM MIX, 1 M DEX, and 1 g/ml insulin (collectively called MDI). This time point was considered day 0. On day 2, the induction medium was replaced by DMEM with 10% FBS containing 1 g/ml insulin, and cells were subsequently fed every 2 days with the same medium. At the time points indicated, cells were either harvested for Western blot analysis or fixed in 10% formalin and stained with Oil Red O. For differentiation experiments with cAMP activators, the indicated concentrations of forskolin or 8-Br-cAMP were added to the MDI medium during the first 2 days of differentiation. For differentiation experiments with PPAR ligands, 3 M 15d-PGJ2 or 5 M troglitazone was added to the MDI medium during the first 2 days of differentiation. HEK293 or HEK293T cells were maintained in DMEM with 10% FBS and 1× penicillin/streptomycin. For LanCL2 overexpression in 3T3-L1 cells, 3×FLAG-tagged LanCL2 plasmid was nucleofected in these cells using a Nucleofector II/2b device (Lonza, Walkersville, MD), and nucleofected cells were subjected to selection with 700 µg/ml G418 for 2 days and subsequently plated for differentiation.

Oil Red O staining
3T3-L1 adipocytes were washed with Dulbecco's PBS (DPBS), fixed in 10% formalin for 30 min, and subsequently rinsed in deionized water. Cells were then incubated with 60% isopropanol for 5 min, followed by staining with 0.2% (wt/vol) Oil Red O solution in 60% isopropanol for 10 min. Cells were subsequently rinsed to remove nonspecific binding of Oil Red O and imaged. To quantify lipid content, Oil Red O was extracted with 100% isopropanol, and absorbance was measured at 500 nm. Oil Red O values from undifferentiated cells in parallel cultures served as blanks.

The 5-ethynyl-2'-deoxyuridine staining
Two-day confluent shScr and shLanCL2#1 cells were treated with the adipogenic cocktail. After 16 h, 5-ethynyl-2'-deoxyuridine (EdU, Berkshire, UK) was added to the medium at a final concentration of 10 M, for an additional 2 h. After the 2-h pulse, cells were rinsed twice with PBS and fixed in 10% formalin solution for 15 min. Cells were subsequently incubated with PBS buffer containing 40 µM FAM (fluorescein) azide 5-isomer (Lumiprobe, Hunt Valley, MD), 1 mM CuSO 4 , and 100 mM ascorbic acid for 30 min, protected from light. The cells were then washed with PBS containing 0.5% Triton X-100 and stained with 1 µg/ml 4',6-diamidino-2-phenylindole (DAPI, Thermo Scientific, Waltham, MA) for 30 min. Cells were finally washed with PBS, and images were taken using a Zeiss Axiovert 200M microscope. EdU-and DAPI-positive cells were counted using ImageJ software (National Institutes of Health).

Western blotting
Cells were lysed using RIPA buffer (Thermo Scientific) in the presence of 1× protease inhibitor cocktail (Thermo Scientific) and centrifuged at 14,000 g for 5 min at 4°C. Protein concentrations were determined using Dc assay (Bio-Rad, Hercules, CA), and equal amounts of protein for each sample were mixed with Laemmli sample buffer (Bio-Rad) and boiled for 5 min at 95°C. Proteins were resolved using polyacrylamide gels (Bio-Rad) and subsequently transferred to nitrocellulose membranes (GE-Healthcare, Piscataway, PA). Membranes were then blocked for 1 h in 5% nonfat dry milk, followed by overnight incubation with the appropriate primary antibodies. Membranes were subsequently washed with TBS with 0.05% Tween 20 and incubated with secondary antibodies. Chemiluminescent signals were developed using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific), and signals were captured using MyECL Imager (Thermo Scientific). cAMP assay cAMP concentrations in 3T3-L1 cells were measured by using a Cyclic AMP XP assay kit from Cell Signaling Technology.
The 3T3-L1 cells plated in 12-well plates were treated with 1 M forskolin, in the presence of 0.5 mM MIX for 30 min at 37°C. Cells were subsequently lysed, and cAMP concentrations were measured following the manufacturer's recommendations. The absolute cAMP concentrations were calculated from a standard curve generated using cAMP standards provided in the kit.

Supernatant replacement assay
Two-day postconfluent shScr 3T3-L1 cells were treated with the adipogenic cocktail for 2 days, after which the conditioned medium (CM) was removed and added to scramble and LanCL2 KD cells in another plate. Fresh medium containing insulin was added to the first plate, and this medium was again added to the scramble and LanCL2 KD cells 2 days later. This protocol was continued until the end point of the experiment.

Real-time quantitative PCR
Total RNA was extracted from 3T3-L1 cells using the E.N.Z.A Total RNA kit (Omega Bio-tek, Norcross, GA). A total of 800 ng of RNA from each sample was reverse-transcribed into cDNA using Protoscript II reverse transcriptase (New England BioLabs) using oligo-dT primers. Quantitative PCR (qPCR) was performed on a Roche LightCycler 480 system using the iTaq Universal SYBR Green Supermix (Bio-Rad), following manufacturer's instructions. Gene expression was normalized to 36B4, which encodes for a ribosomal phosphoprotein and the expression of which remained unchanged during the course of differentiation (data not shown). The primers for PPAR and aP2 were purchased from Sigma-Aldrich (Kiqstart SYBR Green predesigned primers). The sequences for other primer pairs are as follows: C/EBP forward (F): 5′-TGTTGGGGATTTGAGTCTGTG-3′; C/EBP reverse (R): 5′-GGAAACCTGGCCTGTTGTAAG-3′; 36B4 F: 5′-GTCACTGT-GCCAGCTCAGAA-3′; 36B4 R: 5′-TCAATGGTGCCTCTGGA-GAT-3′. The efficiencies for each primer pair were calculated using serial dilutions of cDNA and were found to lie within a 5% range of the housekeeping gene (36B4, data not shown).

LanCL2-His and MBP-PPAR expression and purification
For expression of LanCL2-His, human LanCL2 cDNA was cloned into pETDuet-1 vector and expressed in Rosetta 2 Escherichia coli cells. Cells were induced with 0.2 mM isopropyl -D-1-thiogalactopyranoside (IPTG) at 18°C for 18 h and were subsequently lysed in buffer containing 20 mM Tris, pH 7.5, 0.5 M NaCl, 1 mM tris(2-carboxyethyl)phosphine, 5 mM imidazole, and 10% glycerol. LanCL2 protein was purified by using a HisTrap column (GE Healthcare) and desalted using a PD 10 column (GE Healthcare). Purified LanCL2 protein was used for pulldown assays. For expression of MBP-PPAR, E. coli BL21 (DE3) competent cells were transformed with pMAL-p2X vector encoding mouse PPAR and induced with 1 mM IPTG for 18 h at 18°C. PPAR was expressed as a fusion protein with MBP and was further purified using amylose resin (New England Biolabs).

Pulldown assays
For LanCL2-His pulldown of PPAR in HEK293 cells, FLAG-PPAR was overexpressed for 48 h, and the cells were lysed in His pulldown buffer (20 mM Tris-Cl, pH 8.0, 150 mM NaCl, 25 mM NaF, 25 mM -glycerolphosphate, 0.1 mM NaVO 3 , 20 mM imidazole, and 0.3% Triton X-100) containing 1× protease inhibitor cocktail (Roche, Indianapolis, IN). Cell lysates were centrifuged at 14,100 g, and supernatants were incubated with 50 g of human LanCL2-His protein for 10 h at 4°C. For LanCL2-His pulldown of PPAR in 3T3-L1 cells, the cells were differentiated for 3 days, and lysates were subsequently collected and incubated with 150 g of recombinant human LanCL2-His protein for 10 h at 4°C. The cell lysates from both HEK293 and 3T3-L1 cells that had been incubated with LanCL2-His were subsequently treated with His-Pur cobalt resin (Thermo Scientific) for 2 h. The beads were washed three times with His pulldown buffer and boiled in 2× SDS buffer for 10 min. The samples were subsequently used for Western blotting. FLAG-PPAR from HEK293 cells and endogenous PPAR from differentiated 3T3-L1 cells were detected by anti-PPAR antibody. LanCL2-His was detected by anti-His-tag antibody.
For pulldown of PPAR using FLAG antibody in HEK293 cells, FLAG-PPAR was overexpressed in HEK293 cells for 48 h, and the cells were lysed in IP lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40, 1 mM EDTA, 5% glycerol, 1× protease inhibitor cocktail). Cell lysates were centrifuged at 14,100 g for 15 min, and the supernatants were incubated with anti-FLAG-M2 beads (Sigma-Aldrich) at 4°C for 12 h. The beads were washed three times with IP lysis buffer for 5 min each and boiled in 2× SDS buffer for 10 min. The samples were subsequently used for Western blotting. FLAG-PPAR was detected by anti-PPAR antibody, and LanCL2 was detected by homemade anti-LanCL2 antibody.
To determine whether PPAR and LanCL2 have a direct physical interaction, equimolar amounts (500 nM) of purified MBP or MBP-PPAR and LanCL2-His were incubated together at 4°C for 30 min with gentle shaking. LanCL2-His was subsequently pulled down by incubating with cobalt beads for 10 min. The beads were washed three times with His-pulldown buffer for 5 min each and boiled in 2× SDS for 10 min. The samples were subsequently used for Western blotting. The LanCL2-His protein was detected by anti-His tag antibody, and MBP or MBP-PPAR protein were detected by anti-MBP tag antibody.

Confocal and fluorescence microscopy
For LanCL2 localization experiments, 3T3-L1 or HEK293 cells were plated in 35 mm glass-bottom dishes (MatTek Corporation, Ashland, MA) and transfected with hLanCL2-eGFP construct using Polyfect reagent (Qiagen, Valencia, CA). Twenty-four hours later, nuclei were stained with 5 µg/ml Hoechst 33342 (Molecular Probes, Eugene, OR) for 15 min at 37°C, and cells were imaged using a Zeiss LSM 700 confocal microscope. Images were processed using Zenlite software (Carl Zeiss). For fluorescence imaging of 3T3-L1 cell counts, scramble or LanCL2 KD cells were plated in 35 mm glass-bottom dishes and allowed to grow for 2 days postconfluency. Cells were subsequently stained with 5 µg/ml Hoechst 33342 for 15 min at 37°C, washed with DPBS, and imaged using a Zeiss Axiovert 200M microscope. Stained nuclei were quantified using ImageJ.

Luciferase assay
HEK293 cells were transduced with lentiviral vectors expressing shRNAs against human LanCL2 or scramble control. Transduced cells were selected with 1 g/ml puromycin and subsequently transfected with PPRE-X3-TK-firefly luciferase plasmid, pIS1 renilla luciferase construct, and pcDNA-FLAG-PPAR plasmid using TransIT-LT1. Twenty-four hours later, 3 M 15d-PGJ2 or 5 M troglitazone was added. After a subsequent 24 h, cells were lysed, and luciferase assays were performed using the Dual Glo Luciferase Assay kit from Promega (Madison, WI). For overexpression studies, hLanCL2-3X-FLAG was cotransfected along with the other constructs in HEK293 cells without KD. The readings from firefly luciferase were normalized to that of renilla to account for differences in cell number and transfection efficiencies. To determine whether LanCL2 KD in HEK293 cells decreases the levels of exogenously expressed PPAR, lysates were collected from scramble or LanCL2 KD cells overexpressing FLAG-PPAR for 24 h, and lysates were subsequently processed for Western blotting.

Statistics
All statistical analyses were performed using GraphPad Prism 4 (GraphPad Software, Inc., San Diego, CA). For statistical analysis, Student's t-test and one-way ANOVA were performed, wherever applicable. Statistical significance was set to P < 0.05, and the data are presented as mean ± SE.

KD of LanCL2 impairs 3T3-L1 differentiation
The 3T3-L1 preadipocytes were induced to differentiate by the addition of MDI. Accumulation of lipid droplets was observed by day 3, with complete differentiation evident by day 8. We first probed the expression of LanCL2 during the differentiation process and noted a constant level of expression throughout (Fig. 1A). To explore any potential function of LanCL2 in adipogenic differentiation, we knocked it down using lentiviral transduction of two different shRNAs directed against mouse LanCL2. An shRNA with a scrambled hairpin sequence was used as a negative control for all experiments. Efficient KD of LanCL2 was confirmed by Western blotting (Fig. 1B). LanCL2 KD strongly decreased differentiation of 3T3-L1 cells, as determined by Oil Red O staining (Fig. 1C) and spectrophotometric quantification of triglyceride accumulation (Fig. 1D). KD of LanCL1, on the other hand, did not affect differentiation of 3T3-L1 cells, implying that the inhibitory effect on adipocyte differentiation is LanCL2-specific (Fig. 1E-G). Overexpression of LanCL2 in 3T3-L1 cells (Fig. 1H) did not increase the differentiation potential of these cells further (Fig. 1I, J), suggesting that endogenous levels of LanCL2 may be sufficient for mediating full differentiation of these cells to adipocytes.
We observed LanCL2 KD cells to have slightly reduced proliferation rates in comparison to use of the scrambled hairpin. Hoechst 33342 staining of scramble and LanCL2 KD cells 2 days postconfluency showed slightly reduced cell count for LanCL2 KD cells (supplemental Fig. S1A, B). A slower proliferation rate in LanCL2 KD cells indicated that LanCL2 KD cells may spend reduced time in contact inhibition. Since contact inhibition during confluency is critical for adipogenic differentiation of 3T3-L1 cells (22,23), we investigated the possibility that the decreased differentiation potential of LanCL2 KD cells resulted from reduced time spent in contact inhibition. LanCL2 KD cells were plated at a higher density such that they reached confluency before the scramble control cells (supplemental Fig.  S1C). However, the higher cell count did not rescue differentiation in LanCL2 KD cells, ruling out a role of decreased proliferation rate or reduced contact inhibition in mediating the inhibited differentiation phenotype (supplemental Fig. S1D). We next checked whether LanCL2 KD cells could be defective in clonal expansion, a necessary step for differentiation mediated by the addition of MDI. EdU staining in scramble and LanCL2 KD cells showed similar proliferation rates after the addition of MDI (supplemental Fig. S2), suggesting that there is no defect in LanCL2 KD cells in clonal expansion.

LanCL2 KD does not affect cAMP in 3T3-L1 cells
Recent reports have implicated LanCL2 to act downstream of the hormone ABA. The functional aspects of ABA are mediated by increasing the cellular concentrations of cAMP (15). An increase in cAMP concentration is essential during early stages of 3T3-L1 differentiation, acting through Epac (24). We therefore sought to determine whether the inhibition of differentiation in LanCL2 KD cells is the result of decreased cAMP signaling. The increase in cAMP concentration in 3T3-L1 cells is mediated by MIX (25), included in the adipogenic cocktail. Our analysis of cAMP concentrations revealed no difference in scramble and LanCL2 KD cells after the addition of MDI ( Fig. 2A). To further rule out a role of cAMP signaling in the adipogenic function of LanCL2, we included forskolin in the differentiation cocktail during the first 2 days of differentiation. A known cAMP inducer (26), forskolin resulted in a robust increase in cellular cAMP concentrations in both scramble and LanCL2 KD cells (Fig. 2B). However, addition of forskolin did not rescue differentiation in LanCL2 KD cells (Fig. 2C, D). Similar observations were made in cells treated with 8-Br-cAMP (Fig. 2E, F), a cell-permeable analog of cAMP that is resistant to degradation and is known to elicit similar effects as cAMP (27). Therefore, LanCL2 does not appear to function upstream of cAMP in adipogenic differentiation.

Akt phosphorylation is not affected by LanCL2 KD
We have previously shown that LanCL2 positively regulates Akt phosphorylation in HepG2 liver cells (17). Akt activation involves the phosphorylation of Ser473 in the C-terminal hydrophobic motif and of Thr308 in the activation loop of the protein (28,29). Because activation of Akt is required for the transcriptional activity of PPAR (30), we probed whether impaired Akt signaling might be responsible for reduced differentiation in LanCL2 KD cells. Immunoblotting for phospho-Ser473 and phospho-Thr308 during differentiation did not show any difference between scramble and LanCL2 KD cells (supplemental Fig. S3), implying that this pathway is not responsible for the impaired adipogenesis observed here.

LanCL2 KD cells have decreased expression of late adipogenic markers
To identify the step at which adipogenesis is inhibited in LanCL2 KD cells, we monitored the expression of early and late adipogenic markers. Expression analysis of C/EBP, C/EBP, C/EBP, PPAR, and aP2 from day 0 through 6 of differentiation showed very low expression of these proteins on day 0 (supplemental Fig. S4), as expected (31,32). We therefore focused on the expression of these proteins during differentiation; i.e., from day 1 onward. Levels of early adipogenic markers C/EBP and C/EBP did not differ in scramble and LanCL2 KD cells at any time during the differentiation process (Fig. 3A). However, the expression of late adipogenic markers PPAR, C/EBP, and aP2 on days 4 and 6 was decreased in cells with LanCL2 KD. RT-qPCR analysis further revealed that the mRNA levels encoding PPAR, C/EBP, and aP2 were significantly reduced in LanCL2 KD cells on days 3 and 6 of differentiation (Fig. 3B).

15-deoxy- 12,14 -Prostaglandin J2, or CM from control cells, cannot rescue differentiation in LanCL2 KD cells
Because sustained expression of PPAR during adipogenesis requires the presence and binding of a ligand, we probed whether LanCL2 KD cells might have a defect at this stage. An endogenous ligand for PPAR has been proposed to be produced and secreted by actively differentiating cells (33), and the addition of CM from differentiat-ing 3T3-L1 cells can augment the transcriptional activity of PPAR, resulting in increased differentiation (19). If LanCL2 KD cells are defective in the generation of an endogenous ligand, the presence of such a ligand in the CM of normally differentiating cells could restore differentiation in the KD cells. We added CM from differentiating scramble cells to scramble or LanCL2 KD cells, every 2 days, until the endpoint of the experiment. This resulted in normal differentiation in CM-treated scramble cells, but no rescue of differentiation was observed in LanCL2 KD cells (Fig. 4A). Likewise, addition of CM from LanCL2 KD cells induced differentiation in scramble cells, but not in LanCL2 KD cells (Fig. 4A). These results suggest that LanCL2 may not regulate differentiation through the production of a secreted PPAR ligand.
We then explored the possibility that LanCL2 KD cells could be defective in signal propagation after the binding of a PPAR ligand. We tested whether addition of 15-deoxy- 12, 14 -prostaglandin J2 (15d-PGJ2), an endogenously occurring but relatively weak PPAR agonist (34), to the differentiation medium could rescue differentiation. As shown in Fig. 4B, C, 15d-PGJ2 could not restore differentiation in LanCL2 KD cells. Western blot and RT-qPCR analysis further showed that the protein and mRNA expression levels of late adipogenic genes encoding PPAR, C/EBP, fatty acid synthase, and aP2 remained low in LanCL2 KD cells grown in the presence of 15d-PGJ2 ( Fig. 4D  and supplemental Fig. S5).
Taken together, these results demonstrate that in the absence of LanCL2, endogenously occurring PPAR activators such as 15d-PGJ2, or ligand(s) potentially secreted into the medium of differentiating cells, could not induce differentiation. These observations suggest that LanCL2 is not involved in the production of a secreted ligand and that LanCL2 KD may interfere with the transactivation of adipogenic genes by PPAR.

LanCL2 interacts with PPAR
If LanCL2 is involved in the transactivation activity of PPAR, the simplest mechanism would be via a direct or indirect physical interaction. PPAR has been widely reported to reside in the nucleus, consistent with its role as a transcriptional activator (35). Immunofluorescence imaging of transiently expressed hLanCL2-eGFP revealed nuclear localization of LanCL2 in both 3T3-L1 and HEK293 cells ( Fig. 5A and supplemental Fig. S6). Because PPAR and LanCL2 localized to the same cellular compartment, we further explored the possibility of a physical interaction. We performed pulldown assays with lysates from day 3 differentiated 3T3-L1 cells, because the expression of endogenous PPAR is relatively high at this point (Fig. 3A), whereas lipid accumulation remains relatively low. We found that at this stage, LanCL2-His could pull down endogenous PPAR2, the isoform that is important for adipogenesis (8,36) (Fig. 5B and supplemental Fig. S7). To determine whether this interaction also occurred in nonadipogenic cells, we repeated the pulldown in HEK293 human embryonic kidney cells. Purified LanCL2-His pulled down transiently expressed PPAR in HEK293 cells experiments. One-way-ANOVA was used to compare each data point to scramble control. * P < 0.05; ** P < 0.01. (Fig. 5C). Furthermore, overexpressed FLAG-PPAR coimmunoprecipitated with endogenous LanCL2 in these cells (Fig. 5D), and in vitro pulldown assays using purified PPAR and LanCL2 proteins demonstrated a direct physical interaction between the two proteins (Fig. 5E).

LanCL2 is important for PPAR activity during differentiation
Given that LanCL2 physically interacts with PPAR, we explored the possibility that LanCL2 regulates PPARmediated transcriptional activation of target genes. To this One-way-ANOVA was used to compare each data point to day 0 undifferentiated controls. * P < 0.05; ** P < 0.01; *** P < 0.001. end, we utilized a luciferase reporter driven by PPRE-luc (37). As expected, expression of recombinant PPAR in HEK293 cells, which express a very low level of endogenous PPAR (Fig. 6A, left inset), stimulated the activity of this reporter (Fig. 6A). The reporter activity was further enhanced by the addition of 15d-PGJ2 or troglitazone, a strong synthetic activator of PPAR (34,38,39). Importantly, when LanCL2 was knocked down by lentivirusmediated delivery of shRNAs, the reporter activity was significantly reduced under all conditions (Fig. 6A). This reduction was not caused by decreased PPAR levels upon LanCL2 KD (supplemental Fig. S8). Overexpression of LanCL2 in HEK293 cells did not further increase PPRE-luc activity (Fig. 6B), indicating that the endogenous level of LanCL2 in these cells might be sufficient to induce full activation of PPRE. Consistent with troglitazone providing the strongest recovery of PPAR activity upon LanCL2 KD in these reporter assays, troglitazone could also partially restore adipogenesis in 3T3-L1 cells in which LanCL2 was knocked down (supplemental Fig. S9). Taken together with the observed physical interaction between LanCL2 and PPAR, these results strongly suggest an important role of LanCL2 in mediating the full transactivation potential of PPAR. We tried to assess whether LanCL2 KD would exert an effect on PPAR after completion of differentiation, but we were unable to knock down LanCL2 in already differentiated 3T3-L1 cells.

DISCUSSION
Our study establishes LanCL2 as a positive regulator of adipocyte differentiation, as KD of LanCL2 in 3T3-L1 cells strongly reduced triglyceride accumulation. The effect was specific for LanCL2, as KD of LanCL1 did not inhibit differentiation. We ruled out a role of cAMP or Akt signal transduction pathways in mediating the impaired adipogenic phenotype in LanCL2 KD cells. The arrest in differentiation did not involve early transcription factors C/EBP and C/EBP. Instead, the blockage in adipogenesis seems to be at the stage of PPAR-mediated transactivation of downstream adipogenic genes. Furthermore, we showed that LanCL2 physically interacts with PPAR in different cell types, is located in the same intracellular compartment as PPAR, and is required for the transactivation function of PPAR. We note that a previous study on ABA action in immune cells also reported decreased PPAR activity in reporter assays upon LanCL2 KD (16).
PPAR-mediated transcriptional activation of adipogenic genes is an elaborate process: Ligand binding induces a conformational change in the nuclear receptor, resulting in the dissociation of transcriptional corepressors, such as NCor/SMRT, and the recruitment of several coactivators, such as CREB binding protein and p300 (40)(41)(42). Such coactivators facilitate gene transcription by remodeling the chromatin to an open configuration, through their intrinsic histone acetyltransferase (HAT) activity (43,44). A second group of coactivators that forms a large multiprotein complex in itself, without any intrinsic enzymatic activity, is named the TRAP/DRIP/ARC/Mediator complex (44). Comprising 15-20 proteins, members of this complex serve as molecular bridges between coactivator proteins with HAT activity and the basal transcriptional machinery, such as the RNA polymerase II preinitiation complex (43). For instance, TRAP220 is a pivotal member of the TRAP/Mediator complex; it physically interacts with PPAR, and its specific deletion abrogates PPAR-mediated transcriptional activity (45). Because no enzymatic activity of LanCL2 has been reported to date, it is possible that LanCL2 belongs to this growing class of docking/scaffolding proteins, necessary for PPAR-mediated transactivation. Indeed, our studies show LanCL2 to directly interact with PPAR in vitro, as well as pull down PPAR on day 3 of differentiation-i.e., at a time when the transcriptional activation of downstream adipogenic genes is underway. In addition, a scaffolding function of LanCL2 has been suggested before: in orienting Akt for optimal phosphorylation by the kinase complex mTORC2 (17). It should, however, be noted that the expression of LanCL2 is ubiquitous (18,46), and it is possible that LanCL2 is involved in a general pathway of transcriptional activation, as opposed to specific roles in the adipose tissue. In support of this notion, a role of LanCL2 in regulating gene transcription in a human uterine carcinoma cell line has been proposed by Park and James (47), where it is involved in regulating the expression of multidrug resistance protein-1.
Overexpression of LanCL2 in 3T3-L1 preadipocytes did not enhance their differentiation. These results are in line with our observations from luciferase assays that did not show further enhancement of PPAR transactivation potential when LanCL2 was overexpressed. A similar observation has been made with adipogenic activators SRC-2 and SRC-3, which are critical components of the basal adipogenic machinery (48). Their single or double KD decreased adipogenesis in 3T3-L1 cells, but their overexpression showed little to no effect on the enhancement of adipogenesis (48). These findings suggest that the equilibrium levels of such proteins are sufficient to induce adipogenesis. Consistent with this hypothesis, we observed that the levels of LanCL2 remain constant throughout differentiation.
Both C/EBP and C/EBP are known to regulate the expression of PPAR and C/EBP at the early stages of adipogenesis (49). Although our results do not completely rule out a role of LanCL2 at this stage, our luciferase assays in HEK293 cells directly assessed the role of LanCL2 on PPAR-mediated transactivation in the absence of C/EBP and C/EBP (50). The results show a clear effect of LanCL2 KD on PPAR signaling, and hence we believe it is unlikely that LanCL2 affects C/EBP and C/EBP activity.
Our differentiation experiments in LanCL2 KD cells suggested that an endogenously occurring, natural ligand of PPAR, 15d-PGJ2, or CM from normally differentiating control cells was unable to induce lipid accumulation in the absence of LanCL2, whereas a strong, synthetic PPAR agonist such as troglitazone can partially rescue differentiation. Indeed, the thiazolidinedione class of ligands (such as troglitazone and rosiglitazone) consists of robust activators of PPAR with an affinity of <2 nM, and they are very strong inducers of adipogenesis in cell culture models (19). On the other hand, all endogenously occurring putative ligands, such as 15d-PGJ2, 9-hydroxyoctadecadienoic acid (9-HODE), 13-HODE, and linoleic acid, bind PPAR with a much lower affinity (K d 2-50 µM) and are considered to be relatively weak, albeit natural, activators of PPAR (34,(51)(52)(53). Binding of a strong (or full) PPAR agonist induces a conformational change different than that of a weak (or partial) agonist (54). Partial agonists are known to associate with different structural domains inside the PPAR ligand binding domain, resulting in the recruitment of a different set of coactivators and subsequent differences in the expression of target genes (54). Hence, one possible explanation of our observations is that LanCL2 is important for mediating the activation of PPAR by an endogenous ligand, but a strong synthetic ligand like troglitazone can partially overcome this need by inducing a potent but nonphysiological mode of transcriptional activation.
In summary, our study identifies LanCL2 as a novel regulator of adipogenesis in 3T3-L1 cells, involved in PPARmediated activation of downstream adipogenic genes. We ruled out that this effect is caused by decreasing cAMP levels or Akt phosphorylation, and our data demonstrate that LanCL is not involved in the production of a secreted ligand. Instead, we show that LanCL2 is engaged in a physical interaction with PPAR, which we propose to be important for its full transactivation potential.