Abstract

   graphic file with name ao0c04282_0007.jpg

   Whole-transcriptome analysis of α-mangostin-treated HepG2 cells
   revealed that genes relevant to lipid and cholesterol metabolic
   processes responded to α-mangostin treatment. α-Mangostin downregulated
   a series of cholesterol biosynthetic genes, including SQLE, HMGCR, and
   LSS, and controlled specific cholesterol trafficking-associated genes
   such as ABCA1, SOAT1, and PCSK9. In particular, the downregulation of
   SREBP2 expression highlighted SREBP2 as a key transcriptional factor
   controlling lipid or cholesterol metabolic processes. Gene network
   analysis of SREBP2 and responses of its target proteins demonstrated
   that the effect of α-mangostin on HepG2 cells was mediated by the
   downregulation of SREBP2 expression, which was further supported by the
   reduction of the amount of SREBP2–SCAP complex. In the presence of
   exogenous cholesterols, α-mangostin downregulated SREBP2 expression and
   suppressed PCSK9 synthesis, which might contribute to the increased
   cholesterol uptake in cells, in part explaining the
   cholesterol-lowering effect of α-mangostin.

Introduction

   Mangosteen (Garcinia mangostana L., Clusiaceae) is a plant which
   produces fruit that is dark purple or reddish in color with a soft and
   juicy white edible pulp and a slightly acidic sweet flavor and pleasant
   aroma and is widely used as a dietary supplement.^[34]1,[35]2
   Mangosteen fruit contains a variety of xanthone derivatives as
   secondary metabolites.^[36]3,[37]4 One of the major constituent
   xanthones in mangosteen is α-mangostin ([38]Figure [39]1
   A), which exhibits various pharmacological properties, including
   antiallergic, antiasthmatic, antifungal, anti-inflammatory, and
   antitumor, and cytotoxic activities.^[40]5−[41]9 Several recent studies
   have reported that α-mangostin also displays in vivo antiobesity and
   antiatherosclerotic effects in addition to reducing the plasma levels
   of low-density lipoprotein (LDL) cholesterol and total
   cholesterol.^[42]10−[43]12 Moreover, α-mangostin could suppress the
   development of atherosclerotic lesions and lipogenesis in gallbladder
   cells and transgenic mice.^[44]11,[45]13 However, the comprehensive
   transcriptional or post-transcriptional effect of α-mangostin relevant
   to hepatocellular cholesterol metabolism based on whole-transcriptomic
   analysis remains unexplored.

Figure 1.

   [46]Figure 1
   [47]Open in a new tab

   (A) Structure of α-mangostin. Comprehensive gene profiling of HepG2
   cells treated with α-mangostin for 24 h. (B) Protein–protein
   interaction network for the DEGs was analyzed with GeneMANIA (ver.
   3.4.1) performed with the Cytoscape plugin of the Cytoscape (ver.
   3.5.1, [48]http://www.cytoscape.org/) network visualization and
   analysis environment. “Physical interaction” was chosen with the
   default options, and the “2013 consolidated-pathway” option was
   selected to find associations among the DEGs. (C) Enrichment of KEGG
   pathways in the metabolic gene profiling analysis was determined using
   DAVID software. (D) Enrichment of GO pathways in the metabolic gene
   profiling analysis was determined using DAVID software. (E) R (ver.
   3.3.2) was used for all statistical analyses
   ([49]https://cran.r-project.org/). The “factoMineR”
   ([50]http://factominer.free.fr) and “rgl”
   ([51]https://r-forge.r-project.org/projects/rgl/) packages were used
   for PCA and visualization. (F) Hierarchical clustering of all metabolic
   genes regulated by α-mangostin in HepG2 cells. The genes are annotated
   to the term “lipid metabolic process” (GO:0006629) or its children in
   GO.

   RNA-sequencing (RNA-seq) has been widely applied to understand and
   predict the transcriptional activity of chemical
   compounds^[52]14,[53]15 since transcriptome profiling of individual
   compounds can provide a comprehensive picture of transcriptional
   responses. Accordingly, deciphering the transcriptome can facilitate
   the discovery of new biological pathways or processes that had not been
   previously identified.^[54]16 As controlling the cellular cholesterol
   metabolism involves complex biological processes,^[55]17,[56]18 we
   employed whole-transcriptome sequencing of α-mangostin-treated and
   -untreated HepG2 cells to identify the transcriptional factors and
   relevant genes responsible for the cholesterol metabolism-regulating
   effects.

   Whole-transcriptome sequencing techniques such as RNA-seq can reveal
   all of the genes activated by α-mangostin treatment and determine their
   transcriptional expression levels. In the current study, we aimed to
   identify candidate genes from RNA-seq data using differentially
   expressed gene (DEG) analysis, gene ontology (GO) analysis, and the
   related protein–protein interaction network. Candidate genes with a
   high probability of relevance were further investigated to verify their
   correlations with cellular cholesterol metabolism in the context of
   α-mangostin stimulation.

Results

α-Mangostin Inhibits Lipid Metabolic Pathways

   To assess the transcriptional regulation of α-mangostin on HepG2 cells,
   an RNA-seq approach was adopted, in which the mRNAs were collected from
   HepG2 cells treated with α-mangostin at concentrations of 10 and 20 μM.
   A total of 68 DEGs with at least a 2-fold change in expression level
   were identified by comparing the group of cells treated with 10 μM
   α-mangostin with the mRNA levels of the untreated control group
   ([57]Figure [58]2
   B, [59]Table S1). Genes involved in the cholesterol biosynthesis
   process (16/33), sterol biosynthesis process (16/34), sterol metabolic
   process (17/80), and cholesterol metabolic process (16/69) were
   significantly affected by α-mangostin treatment ([60]Tables S2 and S3).
   In particular, genes categorized into GO terms of “lipid metabolic
   process” (GO:0006629) and “cholesterol metabolic process” (GO:0008203)
   were enriched among the identified DEGs
   ([61]http://amigo1.geneontology.org/; [62]Table S2). The predicted
   functional correlations of these genes were then visualized in a
   protein–protein interaction network ([63]Figure [64]1A).^[65]19
   Enrichment of the top five Kyoto Encyclopedia of Genes and Genomes
   (KEGG) pathways and GO biological processes in the metabolic gene
   profiling analysis were determined using DAVID software. The results
   highlighted steroid biosynthesis and metabolic pathways from the KEGG
   pathways and sterol biosynthesis, cholesterol biosynthesis, and sterol
   metabolic process in GO biological processes ([66]Figure [67]1C,D;
   Benjamini–Hochberg-adjusted P-values are shown for each indicated bar).

Figure 2.

   [68]Figure 2
   [69]Open in a new tab

   Genes regulated in HepG2 cells by α-mangostin. (A) Steroid biosynthesis
   (hsa00100) for the DEGs was analyzed with KEGG pathway performed with
   the Cytoscape plugin of the Cytoscape (ver. 3.5.1,
   [70]http://www.cytoscape.org/) visualization and analysis environment.
   (B) α-Mangostin (10 μM) profiles for cholesterol biosynthesis process
   (GO:000695): 32 genes for which the mRNA sequencing was >2.0 and the
   flag was the p value (representing a higher expression) (C) α-Mangostin
   (10 μM) profiles for cholesterol metabolic process (GO:0008203): 19
   genes for which the mRNA sequencing was >2.0 and the flag was the p
   value (representing a higher expression). (D) Cholesterol biosynthesis
   process and associated changes assessed by western blot analysis. (E)
   Validation of the metabolic gene profiles for the example genes by mRNA
   sequencing in HepG2 cells treated with α-mangostin in a dose-dependent
   manner. Data represent the mean ± standard deviation of triplicate
   samples.

   Principal component analysis (PCA) was first used to identify outliers
   between nontreated cells and those treated with different
   concentrations of α-mangostin. The PCA plot of the RNA-seq data showed
   clear segregation and clustering of α-mangostin-treated and nontreated
   groups ([71]Figure [72]1
   E). Heat mapping confirmed the dose-dependent regulation of the gene
   expression levels altered by treatment with α-mangostin ([73]Figure
   [74]1F).

α-Mangostin Regulates Genes Involved in the Cholesterol Metabolism Pathway

   To select and investigate specific genes associated with cholesterol
   metabolism, the DEGs identified between nontreated (normal) and
   α-mangostin-treated HepG2 cells were further analyzed ([75]Figure [76]2
   A–C). Treatment with α-mangostin led to changes in the signaling
   pathways related to steroid biosynthesis (hsa00100), as indicated by
   the KEGG pathway enrichment analysis ([77]Figure [78]2A). De novo
   cholesterol production shares pathways with those used in steroid
   biosynthesis. Among these genes, α-mangostin treatment clearly
   downregulated the expressions of FDFT1, SQLE, LSS, CYP51A1, MSMO1,
   HSD17B7, and DHCR7 ([79]Figure [80]2A), which are closely associated
   with de novo cholesterol synthesis. Furthermore, these DEGs fell into
   two categories: 16 genes dysregulated in the α-mangostin-treated cells
   are involved in cholesterol metabolic process (GO:0008203) and 19 genes
   differentially expressed in the α-mangostin-treated cells are involved
   in the cholesterol biosynthesis process (GO:0006695). Pathway analysis
   revealed that SREBP2, LSS, SQLE, HMGCR, 3-hydroxy-3-methylglutaryl-CoA
   synthase 1 (HMGCS1), isopentenyl-diphosphate delta isomerase 1 (IDI1),
   cytochrome P450 family 51 subfamily A member 1 (CYP51A1), and farnesyl
   diphosphate synthase (FDPS) participate in the lipid metabolism
   ([81]Figure [82]2B,C). These results were consistent with those of
   protein expression level analysis determined using immunoblotting,
   which validated the accuracy of the RNA-seq data.

   PCSK9, SQLE, HMGCR, and LSS are enzyme-encoding metabolic genes with
   significantly downregulated expression in HepG2 cells following
   α-mangostin treatment ([83]Figures [84]2
   D and [85]S2A). We selected 11 genes from the 2 clusters that had
   differential signals >6.0 in mRNA levels. The q value, representing
   higher expression, was validated by reverse transcription-quantitative
   polymerase chain reaction (RT-qPCR; [86]Figure [87]2E). Transcriptional
   expressions of ABCA1, ACSL6, DHCR7, FDFT1, FDPS, HMGCR, IDI1, PCSK9,
   SOAT1, SQLE, SREBP1, and SREBP2, which are related to cholesterol
   synthesis, uptake, and efflux, were modulated by α-mangostin treatment.
   This suggested that SREBF2 is associated with cholesterol homeostasis.

α-Mangostin Preferentially Downregulates SREBP2 over SREBP1 Expression

   To examine the effect and specificity of α-mangostin on SREBP1 and
   SREBP2 expressions, western blotting of HepG2 cells was performed,
   which indicated that the SREBP2 expression was apparently more
   downregulated than the SREBP1 expression ([88]Figures [89]3
   A and [90]S2B). Confocal microscopic observations showed that both
   SREBP1 and SREBP2 expressions were suppressed by α-mangostin treatment
   in the HepG2 cells ([91]Figure [92]3B). Similar to the pattern observed
   in western blotting, the mRNA expression of SREBP1 and SREBP2 was
   observed in HepG2 and Huh7 cells treated by α-mangostin ([93]Figure
   [94]3C,D). SREBP1 mRNA expression was not significantly suppressed by
   α-mangostin treatment compared with the vehicle-treated groups for both
   HepG2 and Huh7 cells, whereas SREBP2 mRNA expression was significantly
   downregulated in both HepG2 and Huh7 cells.

Figure 3.

   [95]Figure 3
   [96]Open in a new tab

   Effect of α-mangostin on SREBP inhibition in the HepG2 human
   hepatocellular liver carcinoma cell line. (A) Protein expressions of
   SREBP1 SREBP2, and β-actin were assayed by western blotting in cells
   treated with α-mangostin for 24 h. (B) Regulation of SREBPs by
   α-mangostin. Cells were cultured for 24 h, fixed, permeabilized, and
   incubated with anti- SREBP1 and SREBP2 antibody followed by an Alexa
   594 conjugated anti-rabbit IgG (red). The nuclei of the corresponding
   cells were visualized by DAPI staining (blue) (magnification: ×60,
   scale bars: 1 μm). (C) mRNA expression of SREBP1 was assayed by western
   blotting in HepG2 and Hur7 cells treated with α-mangostin for 24 h. (D)
   mRNA expression of SREBP2 was assayed by western blotting in HepG2 and
   Hur7 cells treated with α-mangostin for 24 h (E) SREBP2 molecular
   network obtained with Cytoscape v. 3.5.1 using STRING database.

   To investigate the plausible proteins with functions that interact with
   SREBPs, we used the STRING protein–protein interaction network database
   ([97]www.string-db.org). The protein–protein interactions were
   predicted using version 9.1 of the STRING database with a combined
   score >0.9.^[98]20 Cytoscape software version 3.5 was used to visualize
   the protein–protein interaction network ([99]www.cytoscape.org). A
   STRING interactive network was used to identify proteins that can
   interact with SREBPs. As shown in [100]Figure [101]3
   E, SREBP2 has predominant connections to the proteins associated with
   cholesterol biological process in α-mangostin-treated hepatocytes.

α-Mangostin Reduces the Interaction between SCAP and SREBPs

   Immunoprecipitation assays were used to determine whether α-mangostin
   could affect complex formation between SREBP cleavage activating
   protein (SCAP) and SREBP2. Cell lysates were subjected to western
   blotting using anti-SREBP2 or anti-SCAP antibodies. As shown in
   [102]Figure [103]4
   , α-mangostin decreased the interaction between SCAP and SREBP2,
   suggesting that α-mangostin suppressed the formation of SCAP–SREBP2
   complexes in the endoplasmic reticulum and Golgi.

Figure 4.

   [104]Figure 4
   [105]Open in a new tab

   Effect of α-mangostin on SREBP/SCAP interaction in the HepG2 human
   hepatocellular liver carcinoma cell line. Binding of SCAP to SREBP1 and
   SREBP2. Immunoprecipitation with the SCAP antibody (IP: SCAP) was
   performed on the extracts of HepG2 cells with or without the indicated
   concentrations of α-mangostin (10 μM) for 24 h. The resulting
   immunocomplexes were immunoblotted with SCAP, SREBP1, and SREBP2
   antibody.

α-Mangostin Increases Cholesterol Uptake and Selectively Downregulates SREBP2
in the Presence of Exogenous Sterols

   To determine if α-mangostin could increase cholesterol uptake, we
   monitored intracellular cholesterol by staining the cells with filipin
   ([106]Figure [107]5
   A). Since cholesterol synthesis was inhibited and cholesterol
   metabolism was elevated, we assessed the cellular regulation of
   cholesterol in HepG2 cells. Cells were incubated in media supplemented
   with cholesterol (10 μg/mL) and then exposed to α-mangostin or dimethyl
   sulfoxide (0.1%) as a vehicle control. As shown in [108]Figure [109]5A,
   α-mangostin treatment increased cholesterol uptake, and quantification
   of the fluorescence showed significantly increased signal levels of
   filipin in the HepG2 cells ([110]Figure [111]5B). To determine if
   α-mangostin could improve LDL-cholesterol uptake, we monitored
   intracellular LDL-cholesterol by staining the cells with
   1,1′-dioctadecyl-3,3,3,3′-tetramethylindocarbocyanineperchlorate
   (DiI)-labeled LDL. HepG2 cells were incubated with DiI-LDL for 4 h in
   the absence (control) or presence of α-mangostin (10 μM) and then
   examined by microscopy. As shown in [112]Figure [113]5C, α-mangostin
   treatment increased DiI-labeled LDL uptake. Quantification of the
   fluorescence showed significantly increased signal levels ([114]Figure
   [115]5D).

Figure 5.

   [116]Figure 5
   [117]Open in a new tab

   Effect of α-mangostin on cholesterol uptake in the HepG2 human
   hepatocellular liver carcinoma cell line. (A) Filipin staining of HepG2
   cells cultured for 24 h in cells treated with α-mangostin and
   cholesterol (10 μg/mL) as indicated. After 24 h, the cells were washed,
   fixed in paraformaldehyde, and stained with filipin. Intracellular
   filipin-cholesterol complexes were visualized by fluorescence
   microscopy, and images were captured with a fluorescence microscope.
   Representative images are shown. (Magnification: ×40; scale bars: 30
   μm). (B) Relative fluorescence of filipin and PI-stained cells treated
   as described in A that were fluorescence detected using UV excitation
   around 340 nm and emission around 380 nm in a microplate reader.
   Statistical significance of the differences between each treatment
   group and the normal group (*p < 0.05) was determined. (C) Effect of
   α-mangostin on intracellular accumulation of DiI-LDL in HepG2 cells
   incubated with DiI-LDL (10 μg/mL) under control conditions or treated
   with 2 or 10 μM α-mangostin. Representative images are shown.
   (Magnification: ×40; scale bars: 30 μm). (D) Relative fluorescence of
   DiI-LDL and DAPI-stained cells treated as described in C that were
   fluorescence detected using UV excitation around 554 nm and emission
   around 571 nm in a microplate reader. Statistical significance of the
   differences between each treatment group and the normal group (*p <
   0.05) was determined. (E) HepG2 cells treated with hydroxycholesterol
   (1 μg/mL) and/or treated α-mangostin (10 μM) for 24 h. The expressions
   of SREBP1, SREBP2, and PCSK9 were assayed by western blot analysis. (F)
   HepG2 cells treated with hydroxycholesterol and/or treated with
   α-mangostin (10 μM) for 24 h. The expressions of LDL-R, ABCA1, IDI1,
   SQLE, SREBP1, SREBP2, and PCSK9 were assayed by qRT-PCR.

   The formation of mature SREBP2 is prevented by 25-hydroxycholesterol
   (25-HC) via the suppression of SREBP2 cleavage, and 25-HC is also known
   to activate LXR, which in turn upregulates SREBP1 transcription.
   Therefore, the effect of α-mangostin on SREBP1 and SREBP2 expressions
   was evaluated in HepG2 cells in the presence or absence of exogenous
   cholesterol such as 25-HC. Treatment of HepG2 cells with 25-HC
   suppressed SREBP2 expression compared with that detected in the
   nontreated group, whereas SREBP1 expression was not changed
   ([118]Figures [119]5
   E and [120]S2C). Furthermore, in the presence of 25-HC, α-mangostin
   treatment appeared to not enhance the suppression of SREBP1 and SREBP2
   expressions. PCSK9 expression, which is regulated by SREBP2, appeared
   to be similar in HepG2 cells in both the presence and absence of 25-HC.
   PCSK9 expression was downregulated with α-mangostin treatment in both
   the absence and presence of 25-HC. In the qPCR analysis ([121]Figure
   [122]5F), α-mangostin was found to control the expression of SREBP2,
   which is associated with cholesterol homeostasis, as well as the
   downstream target genes of SREBP2 involved in cholesterol synthesis
   (SQLE, IDI1) and cholesterol trafficking (ABCA1, PCSK9).

Discussion

   As one of the major constituents in mangosteen fruit, α-mangostin is
   used as a popular dietary supplement and has been reported to possess
   lipid-lowering activities in both in vitro and in vivo studies. Despite
   extensive research on the properties of α-mangostin, transcriptional
   gene changes related to cellular cholesterol homeostasis regulated by
   α-mangostin have not been investigated. Owing to the complexity of
   cellular cholesterol homeostasis that is orchestrated by many
   interacting proteins and genes, it is not easy to determine the genes
   responsible for the effect of α-mangostin on cholesterol homeostasis.

   By analyzing whole-transcriptome data using DEGs, GO, KEGG pathway, and
   genetic interactions, we found that some of the genes related to
   cholesterol synthesis (e.g., LSS, IDI1, DHCR7) and cholesterol
   trafficking (PCSK9) were downregulated or upregulated by α-mangostin
   treatment. Furthermore, specific genes, such as SQLE, DHCR7, and IDI1,
   and enzymes, such as HMGCR, SQLE, and LSS, that are involved in
   cholesterol synthesis were downregulated.^[123]21−[124]24 Moreover,
   multiple genes closely associated with lipid metabolism were apparently
   downregulated following α-mangostin treatment. Members of the fatty
   acid desaturase (FADS) family, including FADS1 and FADS2, regulate
   fatty acid desaturation by introducing double bonds between defined
   carbons of fatty acyl chains, thereby playing an essential role in the
   lipid metabolic pathway.^[125]25 In addition, acetyl-CoA
   acetyltransferase 2 (ACAT2), also known as cytosolic acetoacetyl-CoA
   thiolase, plays a role in regulating lipid metabolism by catalyzing the
   synthesis of acetoacetyl-CoA from two acetyl-CoA molecules, which is
   later converted into steroids.^[126]26 We found that α-mangostin
   inhibited FADS1, FADS2, and ACAT2 expressions, implying the potential
   regulation of a lipid metabolic pathway.

   Collectively, these observations suggest that α-mangostin might control
   the cholesterol metabolic process. In addition, among the
   α-mangostin-regulated genes, SQLE, LSS, HMGCR, and PCSK9 are well-known
   downstream targets of SREBPs. Therefore, SREBP1 and SREBP2 expressions
   in α-mangostin-treated cells were monitored in an effort to detect any
   dominant SREBP isoform. The genes encoding SREBP1 and SREBP2, the
   predominant SREBP isoforms expressed in the liver, contain sterol
   regulatory elements within their promoters that mediate feed-forward
   transcriptional regulation.^[127]27,[128]28 SREBP2 expression was more
   downregulated by α-mangostin treatment than SREBP1 at both
   transcriptional and post-transcriptional levels. Moreover, predominant
   interactions of SREBP2 with proteins relevant to cholesterol metabolic
   processes in α-mangostin-treated hepatocytes were predicted based on
   analysis of the STRING protein–protein interaction network database.
   Previous reports also suggested that SREBP2 preferably modulates the
   expression of genes associated with cholesterol metabolism, whereas
   SREBP1 primarily regulates lipogenic genes.^[129]29 In line with these
   previous reports, we found that α-mangostin seemed to preferentially
   downregulate the expression of SREBP2 mRNA compared to SREBP1 mRNA in
   two independent hepatocyte cell lines (HepG2 and Huh7). Therefore, we
   concluded that the differential suppression of SREBP2 by α-mangostin
   treatment controlled the cholesterol metabolic process as a central
   factor of transcription.

   Furthermore, downregulation of downstream targets of SREBPs is achieved
   through the translocation of SREBPs to the nucleus. Cholesterol,
   oxysterols, and fatty acids regulate the SREBP pathway through complex
   formation of SREBP with SCAP and the endoplasmic reticulum retention
   protein insulin-induced gene 1 protein (INSIG). The status of this
   complex controls the endoplasmic reticulum-to-Golgi trafficking of the
   SCAP–SREBP complex and thereby modulates the proteolytic activation of
   SREBPs.^[130]30 When cholesterol is abundant in a cell, SREBP remains
   in the endoplasmic reticulum in the form of the SREBP–SCAP–INSIG
   complex.^[131]31 In contrast, depletion of intracellular sterols
   enhances SREBP translocation from the endoplasmic reticulum to the
   Golgi apparatus, leading to subsequent cleavage by site-1 and site-2
   proteases, thereby generating mature SREBP and leading to mature SREBP
   nuclear translocation.^[132]32

   Moreover, 25-HC is known to inhibit SREBP maturation by blocking the
   SCAP-mediated movement of SREBPs from the endoplasmic reticulum to the
   Golgi^[133]33 through the binding of INSIGs to SCAP as a consequence of
   binding of 25-HC to INSIGs.^[134]22,[135]34 Treatment of cells with
   α-mangostin decreased the content of mature SREBP2, as did 25-HC.
   Moreover, the interaction of SREBP2 and SCAP was reduced by treatment
   with α-mangostin, and the amount of the SREBP2–SCAP complex was
   consequently reduced. In particular, cellular cholesterol is
   dynamically controlled by multiple processes, including the sensing of
   intracellular cholesterol levels, uptake of extracellular cholesterols
   via LDL receptors (LDLR) or efflux of intracellular cholesterols by
   transporters, and de novo synthesis of cholesterols inside a
   cell.^[136]18,[137]19 As shown in [138]Figures [139]2
   C and [140]5F, although α-mangostin did not regulate significantly the
   LDLR mRNA expression, cholesterol uptake increased, which was in part
   explained by the downregulation of PCSK9, a downstream target protein
   of SREBP2. PCSK9 facilitates LDLR degradation and thereby decreases LDL
   uptake.

   Collectively, the current study provides a comprehensive mRNA
   expression profile of hepatocytes treated with α-mangostin. Among the
   mRNAs regulated, SREBP2 was strongly associated with the effect of
   α-mangostin on the cholesterol biological process and appeared to
   function as an important hub gene in α-mangostin-treated hepatocytes
   for controlling the networks of cholesterol metabolism. The present
   study of mining the mRNA expression profile generated in response to an
   active component present in dietary supplements and edible fruit
   suggests that attempts to discover new potential of functional foods
   may be complemented by mining the whole transcriptome.

Materials and Methods

Cell Culture, Drugs, and Chemicals

   The HepG2 human hepatocellular liver cell line was obtained from the
   Korea Research Institute of Bioscience and Biotechnology (Daejeon,
   South Korea) and grown in Eagle’s minimum essential medium (EMEM)
   containing 10% fetal bovine serum and 100 U/mL penicillin/streptomycin
   sulfate. Cells were incubated in a humidified 5% CO[2] atmosphere at 37
   °C. EMEM, penicillin, and streptomycin were purchased from Hyclone
   (Logan, UT, USA). Bovine serum albumin (BSA) and 25-hydroxycholesterol
   were purchased from Sigma-Aldrich (St. Louis, MO, USA).
   4′,6-Diamidino-2-phe-nylindole (DAPI) and DiI-LDL were purchased from
   Thermo Fisher Scientific (Waltham, MA, USA). Antibodies against LSS,
   SREBP1, SREBP2, HMGCR, SQLE, and β-actin were purchased from Abcam,
   Inc. (Cambridge, MA, USA). Antibodies against PCSK9 were purchased from
   Cell Signaling Technology. (Beverly, MA, USA). SREBF1, SREBF2, HMGCR,
   PCSK9, IDI1, SQLE, ABCA1, ACSL6, DHCR7, FDFT1, FDPS, SOAT1, and
   glyceraldehyde-3-phosphate dehydrogenase (GAPDH) oligonucleotide
   primers were purchased from Bioneer Corp. (Daejeon, South Korea).
   α-Mangostin was isolated from the chloroform fraction of G. mangostana,
   as previously described, and confirmed by a spectroscopic analysis. The
   purity of α-mangostin was determined to be over 95% by high-performance
   liquid chromatography using an ultraviolet detector.^[141]4

Filipin Staining

   To measure the cholesterol content in the cultured cells, filipin
   staining was used, which enables semiquantification of free cholesterol
   in biological membranes.^[142]35 Briefly, HepG2 cells were treated with
   500 μL of fixation/permeabilization solution (BD Bioscience, San Jose,
   CA, USA). Propidium iodide (PI) dye (50 μM) dissolved in
   phosphate-buffered saline (PBS) was then added and incubated for 5 min
   to stain the cells and nucleic acids. The cells were then incubated in
   500 μL of a PBS-based filipin working solution (100 μg/mL; Sigma, St.
   Louis, MO, USA) for 2 h at room temperature. Filipin stained cells were
   imaged using a Nikon Eclipse Ti-U inverted microscope equipped with a S
   Plan Fluor 40× (N.A. 0.6) objective lens and a DS-F1i digital
   microscope camera in conjunction with NIS-Elements F software (Nikon,
   Tokyo, Japan). An excitation wavelength of 340 nm and an emission
   wavelength of 380 nm were used. Illumination was provided by a Lambda
   DG-4 wavelength switching illumination system (Sutter Instruments,
   Novato, CA, USA). The acquired images were analyzed using a Meta
   Imaging System (Molecular Devices, Sunnyvale, CA, USA). Filipin
   fluorescence was detected using a microplate reader (Molecular Devices,
   Sunnyvale, CA, USA) with UV excitation of approximately 340 nm and an
   emission wavelength of approximately 380 nm.

Measurement of DiI-LDL Uptake

   The amount of DiI-LDL uptake was measured according to the
   manufacturer’s instruction (Thermofisher Scientific). In brief, the
   culture medium of HepG2 cells was replaced with serum-free medium, and
   the cells were incubated with 10 μg/mL DiI-LDL at 37 °C for 24 h,
   followed by observation with a fluorescence Nikon Eclipse Ti-U inverted
   microscope (Nikon, Tokyo, Japan).

Library Preparation and Sequencing

   Total RNA was isolated using a TRIzol reagent (Thermo Fisher
   Scientific, Waltham, MA, USA) according to the manufacturer’s
   protocols. The quantity and quality of the total RNA were tested using
   a model Agilent 2100 bioanalyzer system RNA kit (Agilent Technologies,
   Santa Clara, CA, USA). The isolated total RNA was processed for
   preparing mRNA sequencing library using a TruSeq Stranded mRNA Library
   Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer’s
   instructions. The quality and size of the libraries were assessed using
   an Agilent 2100 bioanalyzer DNA kit (Agilent Technologies). All
   libraries were quantified by qPCR using a CFX96 Real Time System
   (Bio-Rad Hercules, CA, USA) and sequenced on the NextSeq500 sequencers
   (Illumina) with a paired-end 75bp plus single 6bp index read run.

Preprocessing and Genome Mapping

   To monitor the quality and soundness of the raw RNA sequence reads,
   base quality distribution and inclusion of the adapter sequences of the
   raw reads were evaluated using FastQC
   ([143]http://www.bioinformatics.babraham.ac.uk/projects/fastqc)
   software. Potentially existing sequencing adapters and raw quality
   bases in the raw reads were trimmed by Cutadapt software.^[144]36 The
   options −a AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC and −A
   AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAGATCTCGGTGGTCGCCGTATCATT were used
   for the common adapter sequence of the Illumina TruSeq adapters, and
   the options −q 0, −m 20, and −O 3 were used for trimming low-quality 5′
   and 3′ ends of the raw reads. The cleaned high-quality reads after
   trimming the low-quality bases and sequencing adapters were mapped to
   the human reference genome mm10 of the UCSC genome
   ([145]https://genome.ucsc.edu) by STAR software.^[146]37,[147]38 Since
   the sequencing libraries were prepared strand-specifically by using
   Illumina’s strand-specific library preparation kit, the strand-specific
   library option, —library-type fr-firststrand, was applied in the
   mapping process.

Quantifying Gene Expression and DEG Analysis

   Cufflinks software was used to quantify the mapped reads on the human
   reference genome into the gene expression values.^[148]38,[149]39 The
   gene annotation of the human reference genome mm10 from UCSC genome
   ([150]https://genome.ucsc.edu) in GTF format was used as the gene
   model, and the expression values were calculated in Fragments Per
   Kilobase of transcript per Million fragments mapped unit. The DEGs
   between the two selected biological conditions were analyzed by
   Cuffdiff software in Cufflinks package.^[151]40 To compare the
   expression profiles among the samples, the normalized expression values
   of the few hundred selected genes with the most variable expression
   were unsupervised clustered using MeV software of the TM4 microarray
   software suite.^[152]41 The scatter plots for the gene expression
   values and the volcano plots for the expression-fold changes and
   p-values between the two selected samples were drawn by in-house R
   scripts.

Functional Category Analysis

   The biological functional role of the analyzed DEGs between the
   compared biological conditions was assessed using the gene set
   overlapping test. Functional categorized genes were determined using
   the DAVID tool and included the biological processes of GO, KEGG
   pathways, and transcription factor binding target gene sets.^[153]40

Immunofluorescence

   HepG2 cells cultured on slides were fixed with ethanol for 30 min at 4
   °C. Following washing with PBS and blocking with 3% BSA in PBS for 30
   min, the samples were incubated overnight at 4 °C with rabbit
   polyclonal anti-SREBP1 and mouse polyclonal anti-SREBP2 (1:500
   dilution). The excess primary antibody was removed, the slides were
   washed with PBS, and the samples were incubated with Alexa
   488-conjugated and Alexa 594-conjugated secondary antibody (Invitrogen
   Molecular Probes, Burlington, ON, Canada) for 2 h at room temperature.
   Following washing with PBS, the slides were mounted using ProLong Gold
   Antifade reagent containing DAPI (Thermo Scientific, Waltham, MA, USA)
   to visualize the nuclei. Specimens were covered with coverslips and
   evaluated under a confocal laser scanning microscope (Nikon Eclipse,
   Nikon, Japan).

Quantitative Real-Time RT-PCR

   Total cellular RNA was isolated using a TRIzol RNA extraction kit
   according to the manufacturer’s instructions. The first-strand cDNA was
   synthesized with M-MLV reverse transcriptase (Promega, Madison, WI,
   USA). cDNA was then subjected to quantitative real-time PCR using a
   SYBR-Green PCR Master Mix (Bio-Rad) and the CFX384 Real-Time PCR
   Detection System (Bio-Rad). The specificity of the amplification was
   confirmed using a melting curve analysis. Data were collected and
   recorded by CFX Manager Software (Bio-Rad) and expressed as a function
   of the threshold cycle (C[T]). The relative quantity of the gene of
   interest was then normalized to the relative quantity of GAPDH (ΔΔCT).
   The mRNA abundance in the sample was calculated using the equation
   2^-(ΔΔCT). The following specific primer sets were used (5′ to 3′):
   human—GAPDH: GAAGGTGAAGGTCGGAGTCA (forward), AATGAAGGGGTCATTGATGG
   (reverse); human—SREBF1: GGAGGATGGACTGACTTCCA (forward),
   GGCCTTTCACAGAACAGGAA (reverse); human—SREBF2: ACCACGCAGAGCACCAAG
   (forward), GGGAGGAGAGGAAGGAGAGG (reverse); human—HMGCR:
   TGATTGACCTTTCCAGAGCAAG (forward), CTAAAATTGCCATTCCACGAGC (reverse);
   human—PCSK9: GGTACTGACCCCCAACCTG (forward), CCGAGTGTGCTGACCATACA
   (reverse); human—IDI1: AACACCGAAAATAAGCTTCTGC (forward),
   GCGTCACTTTCCTCAAGCTC (reverse); human—SQLE: GGAAAAGCCTGGTCTCCAAT
   (forward), GAGAACTGGACTCGGGTTAGC (reverse); human—ABCA1:
   GTGTTGTCAAGGAGGGGAGA (forward), GCCATCCTAGTGCAAAGAGC (reverse);
   human—ACSL6: GAACTACTGGGCCTGCAAAG (forward), TCCGATGTCTCCAGTGTGAA
   (reverse); human—DHCR7: CTGGACCCTCATCAACCTGT (forward),
   AGGTACCAGGTTTCGTTCCA (reverse); human—FDFT1: ATAACCAATGCACTGCACCA
   (forward), ATAACAGGCAGCCAAAGTGG (reverse); human—FDPS:
   GGAGATGGGGGAGTTCTTTC (forward), GTCCCCAAAGAGGTCAAGGT (reverse);
   human—SOAT1: TTCTATCCCGTGCTCTTCGT (forward), GACTCCATTGCCCAAGAAAA
   (reverse). Gene-specific primers were custom-synthesized by Bioneer.

Immunoblot Analysis

   Protein expression was assessed by western blotting according to
   standard procedures. Briefly, HepG2 cells were cultured in 60 mm dishes
   (2 × 10^6 cells/mL) and then treated with various concentrations of
   α-mangostin. Cells were washed twice in ice-cold PBS, the cell pellets
   were resuspended in RIPA buffer and incubated on ice for 15 min, and
   the cell debris was then removed by centrifugation in 5 min and 1300
   rpm. Protein concentrations were determined using a protein assay dye
   (Bio-Rad) according to the manufacturer’s instructions. Protein (20–30
   μg) was mixed 1:1 with 2 × sodium dodecyl sulfate (SDS) sample buffer,
   loaded onto 8 or 15% SDS-polyacrylamide gel electrophoresis (SDS-PAGE)
   gels, and electrophoretically separated at 150 V. Proteins were then
   transferred onto ImmunoBlot polyvinylidene difluoride membranes
   (Bio-Rad) using a Bio-Rad transfer system according to the
   manufacturer’s instructions. Then, the membranes were blocked with 5%
   skim milk in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1
   h at room temperature and then incubated overnight at 4 °C in primary
   antibodies. The next day, the blots were washed three times with TBST
   and incubated for 1 h with a horseradish peroxidase conjugated
   secondary anti-IgG antibody (diluted 1:2000–1:20,000). The blots were
   then washed three times with Tris-buffered saline (0.1% Tween 20), and
   the immunoreactive bands were developed using the chemiluminescent
   substrate ECL Plus (Amersham Biosciences, Piscataway, NJ, USA). Images
   were acquired by using a ChemiDoc Imaging system (ChemiDoc XRS system;
   Bio-Rad) with Image Lab software 3.0 (Bio-Rad).

Immunoprecipitation

   Total cell extracts were immunoprecipitated overnight at 4 °C using
   rabbit polyclonal anti-SCAP antibody. Immune complexes (typically 600
   μg) were collected using protein A beads prior to being incubated with
   the appropriate antibody for 2 h. This was followed by the addition of
   more protein A beads and incubated for an additional hour. The beads
   were washed with lysis buffer prior to being boiled in SDS sample
   buffer and then subjected to SDS-PAGE and western blotting. Blots were
   developed using the ECL reagent.

Statistical Analysis

   The experimental data are presented as the mean ± standard error of the
   mean. The level of statistical significance was obtained from analysis
   of variance (ANOVA) followed by Dunnett’s t-test for multiple
   comparisons. P values less than 0.05 were considered to be significant.

Acknowledgments