7,3′,4′-Trihydroxyisoflavone, a Metabolite of the Soy Isoflavone Daidzein, Suppresses α-Melanocyte-Stimulating Hormone-Induced Melanogenesis by Targeting Melanocortin 1 Receptor

7,3′,4′-Trihydroxyisoflavone (7,3′,4′-THIF) is a metabolite of daidzein which is a representative isoflavone found in soybean. Recent studies suggested that 7,3′,4′-THIF exerts a hypopigmentary effect in B16F10 cells, however, its underlying molecular mechanisms and specific target protein remain unknown. Here, we found that 7,3′,4′-THIF, but not daidzein, inhibited α-melanocyte-stimulating hormone (MSH)-induced intracellular and extracellular melanin production in B16F10 cells by directly targeting melanocortin 1 receptor (MC1R). Western blot data showed that 7,3′,4′-THIF inhibited α-MSH-induced tyrosinase, tyrosinase-related protein-1 (TYRP-1), and tyrosinase-related protein-2 (TYRP-2) expressions through the inhibition of Microphthalmia-associated transcription factor (MITF) expression and cAMP response element-binding (CREB) phosphorylation. 7,3′,4′-THIF also inhibited α-MSH-induced dephosphorylation of AKT and phosphorylation of p38 and cAMP-dependent protein kinase (PKA). cAMP and Pull-down assays indicated that 7,3′,4′-THIF strongly inhibited forskolin-induced intracellular cAMP production and bound MC1R directly by competing with α-MSH. Moreover, 7,3′,4′-THIF inhibited α-MSH-induced intracellular melanin production in human epidermal melanocytes (HEMs). Collectively, these results demonstrate that 7,3′,4′-THIF targets MC1R, resulting in the suppression of melanin production, suggesting a protective role for 7,3′,4′-THIF against melanogenesis.


INTRODUCTION
Melanin, synthesized in human melanocytes, plays an important role in protecting the skin from the harmful effects of ultraviolet (UV) radiation (Solano et al., 2006;Park et al., 2009). However, the accumulation of abnormal melanin can cause skin pigmentary disorders such as melasma, freckles and age spots and senile lentigines (Briganti et al., 2003).
Epidemiologic studies have shown that the dietary consumption of soy may contribute to a reduced the risk of hyperpigmentation and several compounds found in soy have been studied in terms of their ability to promote skin health (Leyden and Wallo, 2011). Previous studies suggested that genistein induces cellular melanin synthesis and enhances tyrosinase activity (Yan et al., 1999). Whereas, daidzein isolated from Maackia fauriei inhibited tyrosinase activity, but the effect was weak . Unlike genistein, daidzein ( Figure 1A) is converted to 7,3 ,4 -trihydroxyisoflavone (7,3 ,4 -THIF or 3 -hydroxydaidzein, Figure 1B) and other compounds in human liver microsomes (Kulling et al., 2001), which can lead to bioactivation. Recently, previous studies have reported that 7,3 ,4 -THIF has a depigmenting effect on melanin production (Lee et al., 2006;Goh et al., 2011). Additionally, 7,3 ,4 -THIF was shown to inhibit melanin production more effectively than genistein or daidzein in melan-a cells (Park et al., 2010). However, the underlying molecular mechanisms and specific target protein of 7,3 ,4 -THIF remain unknown. Here, we report that 7,3 ,4 -THIF attenuates α-MSH-induced melanogenesis by targeting MC1R.

Cell Culture
The murine melanoma cell line B16F10 was obtained from the Korean Cell Line Bank (Seoul, South Korea). B16F10 cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37 • C in a humidified atmosphere with 5% CO 2 . Human epidermal melanocytes (HEMs) derived from moderately pigmented neonatal foreskins were purchased from Cascade Biologics. HEMs were cultured in Medium 254 supplemented with HMGS at 37 • C in a humidified atmosphere with 5% CO 2 .

Cell Viability
B16F10 cells (7 × 10 3 ) were cultured in a 96-well plate for 6 h. Cell culture media containing 7,3 ,4 -THIF was added at final concentrations of 25, 50, and 100 µM, and then the cells were cultured for 96 h. HEMs (1 × 10 4 ) were cultured in a 96well plate for 6 h. Cell culture media containing 7,3 ,4 -THIF was added at final concentrations of 20, 40, and 80 µM, and then the cells were cultured for 72 h. MTT solution (5 mg/mL) were plused 20 µL/well and incubated 2 h. The media were removed and replaced 200 µL of dimethyl sulfoxide (DMSO) per well to dissolve the MTT formazan. After 2 h, absorbance was measured at 570 nm using a microplate reader (Sunrise-Basic Tecan; Grodig, Austria). The cells were pretreated with samples at the indicated concentrations (10, 20, or 40 µM) for 1 h before being exposed to α-MSH (100 nM) for 3 days. The secreted melanin levels were determined as described in "Materials and Methods." Asterisks indicate a significant difference (*p < 0.05; **p < 0.01; ***p < 0.001) compared with α-MSH treated groups.

Fontana-Masson Staining
Intracellular melanin accumulation was visualized by Fontana-Masson staining with a slight modification (Joshi et al., 2007;An et al., 2010). Cells were fixed in 100% ethanol for 30 min at room temperature and stained for melanin using a Fontana-Masson staining kit from American Master * Tech Scientific, Inc. (Lodi, CA, United States), according to the manufacturer's instructions. In brief, cells were stained with ammoniacal silver for 60 min at 60 • C, followed by incubation in 0.1% gold chloride and then in 5% sodium thiosulfate. Cell morphology and pigmentation were examined under a Nikon phase-contrast microscope (Tokyo, Japan). The images were analyzed using NIS-Elements 3.0 software.

Measurement of the Extracellular Melanin Content
The melanin content was measured using a slight modification of a previously reported method (Eisinger and Marko, 1982;Friedmann and Gilchrest, 1987;Gordon et al., 1989). Briefly, cells (8 × 10 3 ) were cultured in a six-well plate for 24 h. The culture media was replaced with the media containing daidzein or 7,3 ,4 -THIF at the indicated concentrations (10, 20, or 40 µmol L −1 ) for 1 h before being exposed to 100 nmol L −1 α-MSH and harvested 3 days later. After treatment, media were collected and the melanin levels therein were determined by measuring the absorbance at 405 nm using an ELISA reader.

Western Blotting
B16F10 cells (1.5 × 10 4 ) were cultured in a 6 cm dish for 48 h then starved in serum-free medium for an additional 24 h to eliminate the influence of FBS on kinase activation. HEMs (9.6 × 10 4 ) were cultured in 9 cm dish for 6 days. The culture media was replaced with the media containing 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH for 3 days. The harvested cells were disrupted and the supernatant fractions were boiled for 5 min. The protein concentration was determined using a dye-binding protein assay kit (Bio-Rad Laboratories Inc.), as described in the manufacturer's manual. Lysate protein (20-40 µg) was subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretically transferred to a polyvinylidene fluoride membrane (Millipore). After blotting, the membrane was incubated with primary antibodies at 4 • C overnight. After hybridization with secondary antibodies, the protein bands were visualized using an ECL plus Western blotting detection system (Amersham TM , Piscataway, NJ, United States). The relative intensities were quantified by Image J program.
cAMP Immunoassay cAMP levels were measured using a cAMP immunoassay kit (Cayman). Briefly, B16F10 cells were treated with 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 1 µM forskolin for 30 min. Next, the cells were lysed in 0.1 M HCl to inhibit phosphodiesterase activity. The supernatants were then collected, neutralized, and diluted. After neutralization and dilution, a fixed amount of cAMP conjugate was added to compete with cAMP in the cell lysate for sites on rabbit polyclonal antibodies immobilized on a 96-well plate. After washing to remove excess conjugated and unbound cAMP, substrate solution was added to the wells to determine the activity of the bound enzyme. The color development was then stopped, after which the absorbance was read at 415 nm. The intensity of the color was inversely proportional to the cAMP concentration in the cell lysate.

Molecular Modeling and Energy Minimization
To investigate the detailed molecular basis of MC1R inhibition by 7,3 ,4 -THIF, modeling study with 7,3 ,4 -THIF and MC1R was performed twice. First, the sequence based search showed that MC1R has a similarity of 47% with a MC4R theoretical model (PDB entry: 2IQP) (Pogozheva et al., 2005), then the MC1R structure was built by using Prime v3.2 from Schrödinger suite 2013 (Jacobson et al., 2002(Jacobson et al., , 2004 based this MC4R theoretical model, the loops of the MC1R homology structure were refined and minimized for docking studying. 7,3 ,4 -THIF and daidzein were prepared under LigPrep with default parameter. The grid for docking was generated based on the binding sites that were predicted by SiteMap, finally, Flexible docking was performed in the extra precision (XP) mode. The number of poses per ligand was set to 10 in the post-docking minimization. Second, the sequence regions for the catalytic domains in human MC1R were identified using the PFAM profile database. Unfortunately, no experimental structure exists for the human MC1R. Therefore, models of the catalytic domains in human MC1R were built by homology modeling using closest templates available in the protein data bank (PDB). A BLAST search against the PDB identified the X-ray crystal structures 1CJK and 1AB8 with 67.8 and 38.8% sequence identity, respectively. After a careful multiple-sequence alignment with ClustalW V2.1, 100 models were generated with Modeller V9.10 (Sali and Blundell, 1993;Thompson et al., 1994). The modeled regions were 1-282 (MC1R). The best model, according to the intrinsic Modeller DOPE function, was chosen for docking studies (Krivov et al., 2009). Stereochemistry was assessed with PROCHECK (Laskowski et al., 1996). The stereochemistry of the best model was manually inspected with COOT (Emsley and Cowtan, 2004). Electrostatics calculations were performed with APBS V1.2.1; the molecular surfaces with the electrostatics properties (blue: positive charges; red: negative charges, with unit + 5/-5 kT/e) were rendered using VMD V1.8.9 and PyMOL V1.5 (Oberoi and Allewell, 1993;Holst et al., 1994;Wallace et al., 1995;Trott and Olson, 2010).

Docking
Prior to the docking calculation, the missing hydrogen atoms on the receptor were added using PDB2PQR in conjunction with the CHARMm force field (Oberoi and Allewell, 1993). AutoDock Vina was used to perform the dockings (Trott and Olson, 2010). Ligands were prepared with Spartan'10 (Wavefunction, Inc.) and the scripts provided by the AutoDock tools (v1.5.4 r29). The docking grid sizes and positions were set using AutoDock. The grid dimensions were large enough to cover the whole biological unit and centered onto previously defined relevant catalytic and biological areas. The grid size was 26 × 28 × 30 Å (MC1R). In MC1R, the loop region closing the receptor (a.a. 155-163) was set as the flexible region for docking. The lowest docking energy and predicted free binding energy for each compound were retained. These energies were used to sort the VLS results. Top docking solutions were manually inspected using COOT and PyMOL V1.5 to verify and confirm their compatibility with existing knowledge on the receptor (e.g., active site location, binding pockets, and conserved amino acid residues potentially involved in binding).

Model Analysis and Validation
The models were analyzed using Ligplot (Wallace et al., 1995). Confirmation of the interaction maps was performed manually by visual inspection of the models in COOT (Emsley and Cowtan, 2004).

Statistical Analysis
Where applicable, the data are expressed as means ± SD; Student's t-test was used for single statistical comparisons.
Frontiers in Molecular Biosciences | www.frontiersin.org A probability value of p < 0.05 was used as the criterion for statistical significance.

Effect of 7,3 ,4 -THIF and Daidzein on the α-MSH-Induced Melanin Content in B16F10 Melanoma Cells
To investigate the whitening effect of 7,3 ,4 -THIF and daidzein, we first examined the melanin content in B16F10 melanoma cells using Fontana-Masson staining and a melanin content assay. 7,3 ,4 -THIF showed cytotoxicity at 50 µM concentration when treated in B16F10 for 96 h, and 7,3 ,4 -THIF were incubated with B16F10 under non-cytotoxic conditions for all experiments (Supplementary Figure S1). Treatment with 7,3 ,4 -THIF, but not daidzein, significantly reduced the intracellular melanin content of B16F10 cells in a dose-dependent manner (Figures 1Cg-i). Consistent with this, 7,3 ,4 -THIF inhibited the α-MSH-induced extracellular melanin content in a dosedependent manner ( Figure 1D). Additionally, 7,3 ,4 -THIF lightened the color of the extracellular culture media to a greater degree than did arbutin, a well-known whitening agent. In contrast, daidzein had no effect on the α-MSH-induced extracellular melanin content at all indicated concentrations ( Figure 1D). These results suggest that 7,3 ,4 -THIF, but not daidzein, exerts a strong depigmenting effect on B16F10 cells.
Frontiers in Molecular Biosciences | www.frontiersin.org FIGURE 3 | Effect of 7,3 ,4 -THIF on the α-MSH-induced dephosphorylation of AKT signaling or phosphorylation of p38 and PKA signaling in B16F10 melanoma cells. (A) 7,3 ,4 -THIF activated the α-MSH-induced dephosphorylation of AKT. Cells were treated with 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH and harvested 1 h later. (B) 7,3 ,4 -THIF inhibited the α-MSH-induced phosphorylation of p38. Cells were treated with 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH and harvested 15 min later. (C) 7,3 ,4 -THIF inhibited the α-MSH-induced phosphorylation of PKA. Cells were treated with 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH and harvested 15 min later. The cells were disrupted and the levels of phosphorylated and total proteins were determined by Western blot analysis, as described in section "Materials and Methods," using specific antibodies against the respective phosphorylated and total proteins. The data are representative of three independent experiments that gave similar results. The levels of indicated proteins were determined by Western blot analysis, as described in section "Materials and Methods," using specific antibodies. The data are representative of more than two independent experiments that gave similar results. Asterisks indicate a significant difference (*p < 0.05; **p < 0.01; ***p < 0.001) compared with α-MSH treated group.

Binding of 7,3 ,4 -THIF to MC1R
We next investigated whether 7,3 ,4 -THIF affected adenylyl cyclase, an upstream effector of the AKT, MAPK, and PKA cascades. After the stimulation of MC1R by α-MSH, adenylyl cyclase converts ATP to cAMP, resulting in the activation of FIGURE 4 | Binding of 7,3 ,4 -THIF to MC1R. (A) 7,3 ,4 -THIF reduced the forskolin-induced cAMP level in B16F10 melanoma cells. The intracellular cAMP level was determined by a cAMP immunoassay as described in section "Materials and Methods." The results are expressed as the cAMP level relative to the forskolin-treated control. All data are presented as the mean ± SD of three independent determinations. Asterisks (*) indicate a significant difference (p < 0.05) compared with the forskolin-treated group. (B) 7,3 ,4 -THIF binds MC1R directly in vitro and ex vivo. MC1R-7,3 ,4 -THIF binding was confirmed by immunoblotting using antibodies against human MC1R (upper panel) or mouse MC1R (lower panel). Lane 1 (input control), human MC1R protein standard or whole B16F10 cell lysates; lane 2 (negative control), Sepharose 4B was used to pull-down MC1R, as described in section "Materials and Methods," or a B16F10 cell lysate precipitated with Sepharose 4B beads; and lane 3, 7,3 ,4 -THIF-Sepharose 4B affinity beads were used to pull-down MC1R or whole B16F10 cell lysates precipitated with 7,3 ,4 -THIF-Sepharose 4B affinity beads. (C) Daidzein did not bind with MC1R in vitro and ex vivo. MC1R-daidzein binding was confirmed by immunoblotting using antibodies against human MC1R (upper panel) or mouse MC1R (lower panel). Lane 1 (input control), human MC1R protein standard or whole B16F10 cell lysates; lane 2 (control), Sepharose 4B was used to pull-down MC1R, as described in section "Materials and Methods," or a B16F10 cell lysate precipitated with Sepharose 4B beads; and lane 3, daidzein-Sepharose 4B affinity beads were used to pull-down MC1R or whole B16F10 cell lysates precipitated with daidzein-Sepharose 4B affinity beads. (D) 7,3 ,4 -THIF binds to MC1R competitively with α-MSH. B16F10 cellular supernatant fraction (1,000 µg) was incubated with α-MSH at the concentrations indicated (0, 0.1, 1, 10, or 100 µM) and 100 µL of 7,3 ,4 -THIF-Sepharose 4B or Sepharose 4B (as a negative control) in a reaction buffer to a final volume of 500 µL. The pulled-down proteins were detected by western blot analysis as described in "Materials and Methods": lane 1 (input control), whole B16F10 cell lysates; lane 2 (negative control), indicating that neither MC1R binds with Sepharose 4B and lane 3 is the positive control, which indicates that MC1R binds with 7,3 ,4 -THIF-Sepharose 4B. Each experiment was performed three times; representative blots are shown. Asterisks indicate a significant difference.  downstream signaling pathways (Solano et al., 2006;Yamaguchi and Hearing, 2009;Gillbro and Olsson, 2011). Therefore, to identify the effect of 7,3 ,4 -THIF on the intracellular cAMP level, we performed a cAMP immunoassay using forskolin, a direct activator of adenylyl cyclase. At 40 µM, 7,3 ,4 -THIF reduced the forskolin-induced intracellular cAMP level by up to 23.8% (Figure 4A). These results indicate that the inhibition of tyrosinase expression by 7,3 ,4 -THIF involves the inhibition of intracellular cAMP formation and adenylyl cyclase activity. Accumulating data suggest that adenylyl cyclase or MC1R is the potential molecular target of 7,3 ,4 -THIF, and that binding results in the inhibition of melanogenesis and tyrosinase expression. To confirm whether 7,3 ,4 -THIF binds directly to MC1R, we performed an in vitro and ex vivo pull-down assay using 7,3 ,4 -THIF-conjugated and non-conjugated Sepharose 4B beads. MC1R (Figure 4B and Supplementary Figure S2A, To determine the expression level, the cells were pretreated with 7,3 ,4 -THIF at the indicated concentrations (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH for 3 days. (C) Effect of 7,3 ,4 -THIF on the α-MSH-regulated phosphorylation of AKT, p38 and PKA. Cells were treated with 7,3 ,4 -THIF (10, 20, or 40 µM) for 1 h before being exposed to 100 nM α-MSH and harvested 0.5 h later. The cells were disrupted and the levels of phosphorylated and total proteins were determined by Western blot analysis, as described in section "Materials and Methods," using specific antibodies against the respective phosphorylated and total proteins. The data are representative of more than two independent experiments that gave similar results. Asterisks indicate a significant difference (*p < 0.05; **p < 0.01; ***p < 0.001) compared with α-MSH treated group.

DISCUSSION
Soybean, one of the most important foods in Asia, is important as a source of protein, as well as an important source of isoflavone (Genovese et al., 2007). In particular, genistein and daidzein are major active compounds of soybeans and their many pharmacological activities have been known. The isoflavones are metabolized in the body, which requires evaluation of physiological activity in the form of their metabolites. 7,3 ,4 -THIF is considered one of the main oxidized metabolites of daidzein (Heinonen et al., 2003). Recently, 7,3 ,4 -THIF, a metabolite of daidzein, was reported to have beneficial effects on hypopigmentation. Previous studies showed that 7,3 ,4 -THIF inhibits melanin production in melan-a (Yan et al., 1999) and B16 melanoma (Goh et al., 2011) cells. Although accumulating evidence suggests that 7,3 ,4 -THIF can inhibit hyperpigmentation, the underlying molecular mechanisms and its specific target protein have not been reported yet. Here, we report the marked depigmentation effect of 7,3 ,4 -THIF on α-MSH-induced hyperpigmentation and suggest the underlying molecular mechanism and targets.
To determine the effect of 7,3 ,4 -THIF on α-MSH-induced melanogenesis, we first investigated its effect on intracellular and extracellular melanin levels in B16F10 melanoma cells. 7,3 ,4 -THIF, but not daidzein, inhibited α-MSH-induced intracellular and extracellular melanin production in B16F10 cells. These results indicate that 7,3 ,4 -THIF plays a more important role than daidzein in the depigmenting effect of soy.
Previous studies have established the role of tyrosinase and/or TYRPs in B16F10 melanoma cells (Goding and Fisher, 1997;Sato et al., 1997;Yasumoto et al., 1997). Upregulation of the level of several melanogenic enzymes, including tyrosinase and TYRPs, promotes melanin synthesis (Goding and Fisher, 1997;Sato et al., 1997;Yasumoto et al., 1997). Additionally, the expression of tyrosinase is primarily regulated by transcription factors such as MITF and CREB. Therefore, the inhibition of MITF and/or CREB leads to the suppression of melanin synthesis through reduced tyrosinase expression (Goding and Fisher, 1997;Sato et al., 1997;Yasumoto et al., 1997). Our results showed that 7,3 ,4 -THIF inhibited the α-MSH-induced tyrosinase, TYRP-1, TYRP-2 and MITF expression and CREB phosphorylation. This transcriptional regulation of tyrosinase by 7,3 ,4 -THIF was mediated by the activation of AKT and inhibition of p38 and PKA phosphorylation. 7,3 ,4 -THIF also inhibited cAMP production by directly binding with MC1R both in vitro and ex vivo, competitively with α-MSH. Taken together, these results indicate that the inhibition of tyrosinase, TYRP-1, and TYRP-2 expression by 7,3 ,4 -THIF was attributable to the suppression of FIGURE 7 | Hypothetical mechanism of the depigmenting activity of 7,3 ,4 -THIF.
MC1R activity. The melanocytes for our western blot data had different treatment times of 7,3 ,4 -THIF with α-MSH, depending on the proteins to be confirmed. Induction of tyrosinase, TYRPs, and MITF expression required a relatively long treatment of α-MSH for up to 3 days, and a relatively short treatment time (15-60 min) was required to regulate kinase phosphorylation. This is consistent with the sequence of processes in which the signaling pathways regulated by external stimulation regulate the transcriptional activity of MITF, followed by mRNA regulation, and finally the expression of TYR and TYRPs.
Considering the experimental result showing that 7,3 ,4 -THIF binds to MC1R, we carried out modeling study to investigate the binding mode of 7,3 ,4 -THIF to MC1R. In the previous study, Prusis et al. (1997) suggested that Glu94 of MC1R has great influence on ligand binding and receptor function. Based on our docking model results, only 7,3 ,4 -THIF can bind at Glu94 with binding affinities of -4.31 kcal/mol (Figures 5A-D Table 1). Additionally, in the another model structure of MC1R complexed with 7,3 ,4 -THIF, the hydroxyl group at the 3 position of 7,3 ,4 -THIF can potentially form a hydrogen bond with the backbone oxygen of Gly239 and the backbone nitrogen of Lys243 in the MC1R pocket ( Figure 5D). Because of the lack of a hydroxyl group at the 3 position of daidzein, its interaction with the ligand binding site of MC1R would be weaker than that of 7,3 ,4 -THIF; thus, daidzein is unable to effectively inhibit MC1R. These results suggested the possible reasons why only 7,3 ,4 -THIF can inhibit the activity of MC1R, but not daidzein. Further investigation using X-ray crystallography to determine the structure of the inhibitor complex will elucidate the exact binding modes of 7,3 ,4 -THIF with MC1R. Because primary human melanocytes are physiologically relevant to human skin, they have frequently been used for the in vitro screening of skin-whitening compounds. Consistent with the above result, 7,3 ,4 -THIF significantly reduced intracellular melanin production in HEMs.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
JK and NK contributed to conceptualization of the study. JK, J-EL, EL, HC, and ZD conducted experiment. MY, JP, and KL helped with analyzing the data. JK, J-EL, and NK wrote and edited the manuscript. J-EL, TK, and NK contributed to manuscript revision and approved the submitted version. All authors contributed to the article and approved the submitted version.