Sulforaphane suppresses the activity of sterol regulatory element-binding proteins (SREBPs) by promoting SREBP precursor degradation

Sterol regulatory element-binding proteins (SREBPs) are transcription factors that regulate various genes involved in cholesterol and fatty acid synthesis. In this study, we describe that naturally occurring isothiocyanate sulforaphane (SFaN) impairs fatty acid synthase promoter activity and reduces SREBP target gene (e.g., fatty acid synthase and acetyl-CoA carboxylase 1) expression in human hepatoma Huh-7 cells. SFaN reduced SREBP proteins by promoting the degradation of the SREBP precursor. Amino acids 595–784 of SREBP-1a were essential for SFaN-mediated SREBP-1a degradation. We also found that such SREBP-1 degradation occurs independently of the SREBP cleavage-activating protein and the Keap1-Nrf2 pathway. This study identifies SFaN as an SREBP inhibitor and provides evidence that SFaN could have major potential as a pharmaceutical preparation against hepatic steatosis and obesity.


SFaN reduces SREBP-1 precursor forms in the SCAP-deficient SRD-13A cell line. A previous
study has shown that SCAP protein levels directly affect the SREBP precursor stability 19 . In addition, we previously reported that the inhibition of heat shock protein (HSP) 90 destabilizes the SREBP precursor forms preceded by SCAP protein degradation, and this SREBP precursor form destabilization could not be observed in the SCAP-deficient SRD-13A cell line 20 . This line of evidence led us to consider the possible contribution of SCAP in the SFaN-and SFeN-mediated degradation of the SREBP precursor forms. Figure 5A shows that SFaN and SFeN treatments for 6 h reduced the SCAP levels in Huh-7 cells, implying the involvement of the SCAPdependent pathway in SFaN-and SFeN-mediated SREBP precursor degradation. To investigate whether an SCAP-dependent pathway can contribute to the degradation, we took advantage of an SCAP-deficient cell line SRD-13A derived from Chinese hamster cell line CHO-7 19 . We confirmed that SFaN and SFeN treatments for 6 h reduced the SCAP and SREBP-1 precursor protein levels in CHO-7 cells (Fig. 5B). Consistent with a previous report, SRD-13A cells expressed no detectable SCAP and lower amounts of the SREBP-1 precursor than CHO-7 cells. Importantly, treatment with SFaN and SFeN for 6 h further reduced the levels of the SREBP-1 precursor www.nature.com/scientificreports/ protein in SRD-13A cells (Fig. 5B), suggesting that in addition to the SCAP-dependent pathway, there is also an SCAP-independent pathway in SFaN-and SFeN-mediated SREBP precursor degradation.
SFaN accelerates ubiquitin-proteasome-mediated SREBP precursor form degradation. The ubiquitin-proteasome system and the autophagy-lysosome system are the two major protein degradation pathways in eukaryotic cells. To examine whether SFaN and SFeN regulate the SREBP precursor degradation via these pathways, Huh-7 cells were preincubated with MG132 or NH 4 Cl, a proteasome and lysosome inhibitor, respectively. Consistent with a previous report describing that the mature SREBP forms were degraded through the ubiquitin-proteasome system, the amount of the mature SREBP forms increased in the MG132-, but not in the NH 4 Cl-treated cells (Fig. 6, lanes 1, 4, and 7). We also found that MG132, but not NH 4 Cl treatment increased the SREBP precursor forms. Importantly, the MG132 treatment partly attenuated SFaN-and SFeN-mediated SREBP precursor form reduction (Fig. 6, lanes 1-6), although the NH 4 Cl treatment did not affect them (Fig. 6, lanes 1-3 and 7-9). These results suggest that the degradation of the SREBP precursor form, by SFaN and SFeN, is mediated, in part, through the ubiquitin-proteasome pathway.  www.nature.com/scientificreports/ ciency was confirmed by quantitative RT-PCR (Fig. 7A). Figure 7B shows that the SFaN-and SFeN-mediated SREBP precursor degradation could still be observed in the case of HSP27-knockdown Huh-7 cells as well as siControl-transfected Huh-7 cells, suggesting that HSP27 is not involved in the SFaN-and SFeN-mediated SREBP precursor form degradation effect.

HSP27 is not involved in
SFaN accelerates SREBP precursor form ubiquitination. Next, we investigated whether SFaN and SFeN accelerate the SREBP precursor form ubiquitination. Huh-7 cells were transfected with expression plasmids for full-length human SREBP-1a with a C-terminal FLAG tag [amino acids 2-1147; pCMV-SREBP-1a-(2-1147)-3 × FLAG] and ubiquitin with an N-terminal HA tag (HA-Ub) and cultured in the presence of MG132. Next, SREBP-1a-3 × FLAG was immunoprecipitated with an anti-FLAG antibody and the immunoprecipitates were subjected to immunoblotting using an anti-HA antibody. Although the ladder band signal of ubiquitin-conjugated SREBP-1a was barely detected in the absence of SFaN and SFeN, the SFaN and SFeN treatments intensified it (Fig. 8), suggesting that SREBP-1a was ubiquitinated, stimulated by SFaN and SFeN.
Nrf2 activation does not contribute to SFaN-mediated SREBP precursor form degradation. SFaN reportedly stimulates the expression of genes involved in antioxidant enzymes by inducing Nrf2 nuclear translocation and thereby protects against oxidative stress. To examine whether Nrf2 activation could be involved in the SFaN-mediated SREBP precursor form degradation, Huh-7 cells were transfected with an Nrf2-specific siRNA. The Nrf2 knockdown did not affect the SFaN-mediated SREBP precursor form reduction in Huh-7 cells (Fig. 10A). The Nrf2 expression knockdown efficiency was confirmed by quantitative real-time PCR (Fig. 10B). These results indicate that Nrf2 activation is not required for the SFaN-mediated SREBP precursor form degradation.
SFaN does not interact directly with SREBP-1a. SFaN reportedly interacts directly with Keap1 and thereby activates the transcription factor Nrf2 10,22 . To examine whether SFaN interacts directly with SREBP, SFaN-fixed magnetic beads were generated. An alkyne molecule was introduced into the methylsulfinyl moiety of SFaN ( Fig. 11A; alkynyl-SFaN). Alkynyl-SFaN reduced SREBP activity and the levels of SREBP precursor forms to the same extent as SFaN (Fig. 11B,C). Alkynyl-SFaN molecules were fixed to azide-type magnetic beads (FG beads) using click chemistry (Fig. 11D). HEK293 cells were transfected with FLAG-tagged full-length SREBP-1a and Keap1 expression plasmids; then, the cell lysates were exposed to the SFaN beads. The resulting www.nature.com/scientificreports/ pull-down samples were subjected to immunoblotting using an anti-FLAG antibody. Figure 11E displays that the proteins pulled down with the SFaN beads contained Keap1, whereas the control beads did not (  www.nature.com/scientificreports/ forms. These results suggest that SFaN functions as an SREBP inhibitor, at least in part, by stimulating SREBP precursor form degradation. SREBP precursor forms reportedly become unstable when the binding partner SCAP is lost 19,23 . We previously reported that HSP90 is required for SCAP protein stability and its inhibition reduced SCAP, promoting SREBP precursor form degradation in mammalian cells 20 . Asano et al. reported that 25-hydroxyvitamin D interacts with SCAP and induces ubiquitin-mediated SCAP degradation, thereby reducing SREBP-2 precursor forms in CHO-K1 cells 24 . In this study, we showed that SFaN and SFeN reduced SCAP protein levels in Huh-7 and CHO-7 cells (Fig. 5). These results suggest that SCAP reduction could be involved in SFaN-and SFeN-mediated degradation of the SREBP precursor forms. We also determined that SFaN and SFeN SCAP independently stimulate the degradation of the SREBP precursor forms (Fig. 5). To the best of our knowledge, ours is the first study describing SCAP-independent SREBP precursor degradation. Further studies are needed to elucidate the contribution of SCAP-dependent and SCAP-independent pathways to the degradation of SREBP precursor by SFaN and SFeN.
We have previously reported that the mature SREBP forms are rapidly degraded by the ubiquitin-proteasome pathway 25 . Sundqvist et al. described that SREBP-1a phosphorylation on Thr426 and Ser430 resulted in Skp1-Cul1-F box protein ubiquitin ligase (SCF FBW7 ) binding, leading to SREBP-1a ubiquitination. However, the SREBP-1a precursor form was not phosphorylated on Thr426 and Ser430 and was insensitive to Fbw7-dependent degradation 26 . In this study, we showed that SFaN stimulates ΔN SREBP-1a (amino acids 479-1147) ubiquitination and degradation, which does not contain these amino acid residues (Fig. 9). Therefore, we considered that SFaN caused SREBP-1a ubiquitination outside the previously reported N-terminal region. Further studies would be required to determine the ubiquitin ligase and amino acid residues involved in the SFaN-mediated SREBP ubiquitination.
Recently, a degradation signal at the SREBP-2 C-terminus has been identified 27 . This signal, comprising seven noncontiguous amino acids, mediates SREBP-2 proteasomal degradation in the absence of SCAP. A degradation signal is also presented at the SREBP-1 C-terminus (amino acids 1034-1071), although the specific amino acids have not been identified. Considering that SFaN-mediated degradation of SREBP-1a also occurs in the absence of SCAP (Fig. 5) and that degradation occurs with ΔC SREBP-1a (amino acids 2-968 and 2-784) (Fig. 9A), we believe that SFaN-mediated SREBP-1a degradation does not require the degradation signal. SFaN-mediated SREBP precursor degradation was partly inhibited by MG132, a proteasome inhibitor (Fig. 6), indicating the involvement of multiple proteolytic systems in the process. Further studies are required to determine whether the degradation signal at the SREBP-2 C-terminus or the proteolytic systems other than the ubiquitin-proteasome system involved in the SFaN-mediated SREBP precursor degradation. www.nature.com/scientificreports/ Several small molecules reportedly regulate SREBP activity 4 . Betulin, an abundant compound in birch bark, suppresses proteolytic SREBP processing 28 . Dipyridamole, a phosphodiesterase inhibitor, blocks the ER-to-Golgi transport of the SCAP-SREBP complex independently of its phosphodiesterase inhibitor activity 29 . We have previously reported that XN, the most abundant prenylated flavonoid in hops, impairs the ER-to-Golgi translocation of the SCAP-SREBP complex by binding to Sec23/24 and blocking SCAP/SREBP incorporation into common coated protein II vesicles 18 . In addition, the XN isomer isoxanthohumol (IXN), generated non-enzymatically during the brewing process for beer production, also reduced SREBP activity. However, unlike XN, IXN stimulates the ubiquitin-proteasome-dependent SREBP precursor form degradation 17 . The difference between the SFaN-and IXN-mediated SREBP precursor degradation is their dependence on intracellular cholesterol levels. Although SFaN reduces the SREBP precursor forms regardless of intracellular cholesterol levels (Fig. 3), IXNmediated reduction is completely abolished under sterol-supplemented conditions 17 . Considering that when the cells are sterol-depleted, the SCAP-SREBP complex is incorporated into COPII-coated vesicles and transported to the Golgi apparatus from the ER, IXN-mediated SREBP precursor degradation might occur during transport or in the Golgi apparatus. However, SFaN-mediated SREBP precursor degradation is presumed to occur on the ER. We have also previously reported that AITC reduces the mature SREBP forms 15 . Although AITC and SFaN are classified isothiocyanates that contain an-N=C=S reactive group, these food factors suppress SREBP activity in different ways, through the reduction of the mature or degradation of the precursor SREBP forms. At present, it remains unclear what causes this difference, although the isothiocyanate group must be required to suppress SREBP activity, and the sulfoxide group, which is not present in AITC, might be involved in SFaN-mediated SREBP precursor degradation. Further studies are required to determine whether other isothiocyanates reduce SREBP activity and could affect precursor or mature SREBPs. www.nature.com/scientificreports/ SFaN reported anticancer effects by directly binding to Keap1 and activating the transcription factor Nrf2, thereby detoxifying carcinogenesis 8 . In the present study, we observed the SFaN-mediated degradation of the SREBP precursor forms in Nrf2-knockdown cells (Fig. 10), indicating that this degradation is not due to the SFaN-mediated activation of the Keap1-Nrf2 pathway. This speculation was supported by the fact that degradation was observed within 1 h after the SFaN treatment. Therefore, we considered that SFaN binds to factors other than Keap1 and degrades the SREBP precursor forms. SFaN beads were used to determine whether SFaN binds to SREBP, but no such interaction could be observed (Fig. 11E). Currently, SFaN beads are used to identify novel SFaN-binding proteins directly involved in SFaN-mediated SREBP precursor degradation.
SFaN administration to obesity model animals shows an anti-obesity effect, with a decrease in serum cholesterol and TG levels [30][31][32][33] . At present, it remains controversial how SFaN exerts these beneficial effects. One of the promising candidates is the Keap1-Nrf2 pathway 34 . The Keap1-KD-mediated activation of this pathway reportedly reduces weight gain and improves insulin resistance in the short term (8-9 weeks) in HFD-fed mice; however, it conversely causes weight gain and the development of insulin resistance in the long-term in HFDfed mice 35 . In addition, a decrease in weight gain and improvement in insulin resistance have been observed in Nrf2KO mice [36][37][38] . These results imply that the anti-obesity effect of SFaN cannot be explained by the Keap1-Nrf2 pathway activation alone. As SREBP-1 activity suppression has been reported to reduce body weight and improve insulin resistance in obese model mice 4,18,28,39 , it is likely that dietary SFaN exerts beneficial effects, at least in part, by the suppression of SREBP-1 activity. Further studies are required to determine whether dietary SFaN suppresses SREBP activity by promoting the SREBP precursor form degradation in the liver of obese mice. SFaN administration also reportedly suppresses adipocyte differentiation, with AMPK activation and decreased peroxisome proliferator-activated receptor γ and CCAAT/enhancer-binding protein α expression in the adipose tissue 31 . More recently, SFaN administration reportedly exerted anti-obesity effects by promoting white adipose tissue browning 40 . These findings suggest that SFaN administration might exert its effects by regulating multiple pathways in multiple organs.
Here, we analyzed the effects of SFaN using human hepatoma Huh-7 cells. We demonstrated that 30 and 100 μM SFaN treatment for 3 h did not affect the viability of Huh-7 cells as well as LDH release from Huh-7 cells, whereas 20-60 μM SFaN treatment for 24 h reportedly decreased the viability of Huh-7 cells 41 , indicating that a longer SFaN treatment period is expected to increase cytotoxicity. Since the degradation of the SREBP precursor forms by SFaN is observed within 3 h of SFaN treatment, it is unlikely that SFaN cytotoxicity is responsible for this effect. We also showed that SFaN treatment with 10 μM SFaN decreased the FAS gene promoter activity and 30 μM SFaN promoted the degradation of the SREBP precursor. Further studies are required to verify whether the similar effects are observed in normal hepatocytes and animal models. Considering that the maximum concentration of SFaN metabolites in serum reaches 7.4 μM when 150 mL of test meal containing 100 g florets of super broccoli is consumed by humans 42 , it is unlikely that the consumption of broccoli leads to a SFaN serum levels of 30 μM, and dietary supplements containing highly stable SFaN and SFaN derivatives need to be developed. However, it is believed that compounds derived from food that supplemental intake exist at higher concentrations in the intestinal tract than in the serum. The expression of bitter taste-sensing G protein-coupled receptors [type 2 taste receptors (T2Rs)] is regulated by SREBP-2 in enteroendocrine STC-1 cells 43 . T2Rs are known to sense dietary toxins 44 and promote cholecystokinin secretion 45 , thereby suppressing food intake. Therefore, it has been hypothesized that T2Rs play a role in limiting toxin absorption. We demonstrated that SFaN promotes SREBP-2 precursor degradation and thereby reduces their activity. Thus, SFaN consumption might downregulate the expression of bitter taste receptors in the intestinal tract and affect the sensing of dietary toxins.
In summary, the present data demonstrate that SFaN reduces SREBP activity by promoting the degradation of SREBP precursors. Furthermore, we showed that C-terminal region is crucial for SFaN-and SFeN-mediated SREBP-1a precursor form degradation. Additional studies are required to elucidate the direct SFaN targets of these effects. . The polyclonal anti-SREBP-2 (RS004) antibody has been previously described 46 . Peroxidase-conjugated affinity-purified donkey anti-mouse IgGs and peroxidase-conjugated affinity-purified donkey anti-rabbit IgGs were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). www.nature.com/scientificreports/ Cell culture. Huh-7 cells and HEK293 cells were obtained from ATCC. CHO-7 cells and SRD-13A cells, an SCAP-deficient line derived from CHO-7 cells, were kindly provided by Debose-Boyd RA (University of Texas Southwestern Medical Center) 19 . Huh-7 cells and HEK293 cells were maintained in medium B. Huh-7/FAS-luc cells (a stable Huh-7 cell line expressing a luciferase reporter driven by an SRE-containing FAS promoter) 15,18,47 cells were maintained in medium B containing 2 μg/ml blasticidin S. CHO-7 cells and SRD-13A cells were maintained in medium F. SRD-13A cells were originated from the Brown and Goldstein Laboratory 19 . The cells were incubated at 37 °C under a 5% CO 2 atmosphere.

Media and buffers. Medium
Luciferase assays. Huh-7/FAS-luc cells were plated in 12-well plates at a density of 1.0 × 10 5 cells/well, cultured with medium B for 24 h, and the cells were then switched to medium C for 16 h. After incubation for another 24 h in the absence or presence of 10 μM SFaN or SFeN, 30 μM SFaN or SFeN, or 100 µM SFaN or SFeN, the luciferase activity and protein contents of the cell extracts were measured as described previously 48 . The normalized luciferase values were determined by dividing the luciferase activity by the protein content in the cell extracts quantified using the BCA protein assay (Pierce).
Cell viability assay and determination of plasma membrane damage. Cell viability and the plasma membrane damage assays were performed as described previously 49  www.nature.com/scientificreports/ obtained from Bonac. Huh-7 cells were transfected with siRNA (24 pmol per six-well plate) using lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions.
Immunoprecipitation. Immunoprecipitation was performed as described previously 18 . Huh-7 cells were plated in 100-mm dishes at a density of 2 × 10 6 cells/dish and cultured with medium B for 24 h. The cells were then transfected with 15 μg of plasmids including 12 μg of pCMV-SREBP-1a-3 × FLAG and 3 μg of HA-Ub. After incubation for 24 h, the cells were trypsinized and seeded in 100 mm dishes and cultured further under the same conditions. After a 24-h incubation, the cells were treated with 10 μM MG132 for 30 min. After incubation for another 3 h in the absence or presence of 100 μM SFaN or SFeN, the cells were harvested and lysed with Nonidet P-40 lysis buffer [50 mM Hepes-KOH (pH 7.6), 100 mM NaCl, 1.5 mM MgCl 2 , 1% (v/v) Nonidet P-40, 1 mM DTT] supplemented with a protease inhibitor cocktail. The cell lysate was passed through a 25-gauge needle 20 times, rotated at 4 °C for 1.5 h, and centrifuged at 20,000×g for 30 min. The supernatant was rotated overnight with anti-FLAG M2 Affinity gel (Sigma) at 4 °C, and the pelleted gel was washed 3 times with Tris-buffered saline (TBS). The bound proteins were eluted with 3 × FLAG peptide and were subjected to immunoblotting. Fig. S2). N-(4-mercaptobutyl)phthalimide (2) was prepared according to the reported procedure by Ren's group 52 . Under Ar atmosphere, to a solution of thiol 2 (900 mg, 3.8 mmol) in THF (13 mL) were added NaH (60%, 250 mg, 6.2 mmol) and 4-bromobut-1-yne (420 µL, 4.5 mmol) at room temperature. After stirring for 2 h at room temperature, the reaction mixture was poured into water at 0 °C and the resulting mixture was extracted with ethyl acetate. The combined organic layer was washed with brine, dried over anhydrous magnesium sulfate, and concentrated in vacuo. The residue was subjected to silica gel column chromatography (Hex/EtOAc = 10:1) to give N-(4-(but-3-ynylthio)butyl)phthalimide (3, 400 mg, 36%). Under Ar atmosphere, to a solution of phthalimide 3 (300 mg, 1.0 mmol) in MeOH (5 mL) was added hydrazine monohydrate (78 mg, 1.5 mmol) and the mixture was refluxed for 3 h. Then conc. H 2 SO 4 (200 µL) was added to the reaction mixture, which was refluxed for further 1 h. After cooling down to room temperature, water (6.5 mL) was added to the mixture, and the resulting precipitate was removed by filtration. Solvent was evaporated in vacuo to give crude 4-(but-3-yn-1-ylthio)butan-1-amine (80 mg), which was used in the next reaction without further purification. This crude amine (80 mg) was dissolved in CH 2 Cl 2 (1.7 mL) and treated with di-2-pyridyl thionocarbonate (120 µg, 0.52 mmol), followed by 5% aqueous NaOH solution (1.2 mL). The mixture was stirred for 4 h at room temperature, then diluted with H 2 O, and extracted with CHCl 3 . The combined organic layer was dried over anhydrous magnesium sulfate, and concentrated in vacuo. The residue was subjected to silica gel column chromatography (Hex/EtOAc = 10:1-3:1) to give but-3-yn-1-yl(4-isothiocyanatobutyl)sulfane (4, 60 mg, 29% in two steps) as a pale yellow oil. Under Ar atmosphere, to a solution of sulfide 4 (60 mg, 0.30 mmol) in CH 2 Cl 2 (2.7 mL) was slowly added m-chloroperoxybenzoic acid (77%, 74 mg, 0.33 mmol) at -78 °C. The solution was After stirring for 1 h at − 78 °C, saturated sodium bicarbonate solution and 10% aqueous sodium thiosulfate solution were added to the reaction mixture and the mixture was extracted with CH 2 Cl 2 . The combined organic layer was washed with saturated sodium bicarbonate solution and brine successively, dried over anhydrous magnesium sulfate, and concentrated in vacuo. The residue was subjected to silica gel column chromatography (Hex/EtOAc = 3:1-1:2) to give 4-((4-isothiocyanatobutyl)sulfinyl)but-1-yne (1, 25 mg, 39%) as a colorless oil.

Alkynyl-SFaN-immobilized beads (SFaN beads) preparation. SFaN beads preparation was per-
formed according to a previously published procedure 53 , with minor modifications. Immobilization was performed with FG beads (Tamagawa Seiki, Nagano, Japan) according to the manufacturer's instructions, with modifications. Azide beads (1 mg) were incubated with 62. www.nature.com/scientificreports/ erol], SDS sample buffer was added to the beads, and then the suspensions were heated at 96 °C for 5 min. The elution samples were subjected to immunoblotting.
Statistical analysis. All data are represented as mean ± S.E. Statistical analysis was performed using the Ekuseru-Toukei Ver.2.0 (Social Survey Research Information). Comparisons between treatments were made by a Student's t test for two groups. One-way ANOVA followed by the Bonferroni procedure was used to compare more than two groups. Differences were considered significant at P < 0.05.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.