Cyclic Phthalate Esters as Liver X Receptor Antagonists with Anti-hepatitis C Virus and Anti-severe Acute Respiratory Syndrome Coronavirus 2 Properties

Abstract

The liver X receptor is a nuclear hormone receptor that regulates lipid metabolism. Previously, we had demonstrated the antiviral properties of a liver X receptor antagonist associated with the hepatitis C virus and severe acute respiratory syndrome coronavirus 2. In this study, we screened a chemical library and identified two potential liver X receptor antagonists. Spectroscopic analysis revealed that the structures of both antagonists (compounds 1 and 2) were cyclic dimer and trimer of esters, respectively, that consisted of phthalate and 1,6-hexane diol. This study is the first to report the structure of the cyclic trimer of phthalate ester. Further experiments revealed that the compounds were impurities of solvents used for purification, although their source could not be traced. Both phthalate esters exhibited anti-hepatitis C virus activity, whereas the cyclic dimer showed anti-severe acute respiratory syndrome coronavirus 2 activity. Cyclic phthalate derivatives may constitute a novel class of liver X receptor antagonists and broad-spectrum antivirals.

Introduction

Liver X receptors (LXRs) belong to the superfamily of nuclear hormone receptors and play an important role in lipogenesis,^1–3) cholesterol metabolism,^4,5) and glucose homeostasis.⁶⁾ LXRs regulate the expression of several genes, such as sterol regulatory element-binding protein 1c (SREBP-1c), fatty acid synthetase, and stearoyl CoA desaturase 1 (SCD-1).^7,8) SREBP-1c is a master regulator of lipid and cholesterol metabolism that mediates lipid accumulation in hepatocytes.^9–12) Oxysterols, such as 22R-hydroxycholesterol, 24S,25-epoxycholesterol, 24S-hydroxycholesterol, are endogenous agonists of LXRs.^13–15) LXR ligands can be used to treat several diseases including cancer,¹⁶⁾ rheumatoid arthritis,¹⁷⁾ Alzheimer’s disease,¹⁸⁾ atherosclerosis,¹⁹⁾ and nonalcoholic fatty liver disease²⁰⁾; several synthetic LXR agonists (e.g., T0901317 and GW3925) and antagonists (e.g., GSK2033 and 5CPPSS-50) have been reported.^21,22)

Recently, we revealed that the fungus-derived natural product neoechinulin B inhibited the replication of hepatitis C virus (HCV) by inhibiting LXR activation.²³⁾ The compound impairs the expression of LXR-regulated genes associated with double-membrane vesicle (DMV) formation, such as SCD-1, and disperses DMVs. As DMVs are the putative sites of viral RNA replication, HCV replication is blocked by neoechinulin B. Notably, neoechinulin B exhibits anti-poliovirus activity, whose genome is known to replicate in the DMVs.²³⁾ More recent studies revealed that neoechinulin B is effective against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that requires DMVs to replicate the RNA genome.²⁴⁾ These findings suggest that neoechinulin B is a promising compound for the development of broad-spectrum antiviral drugs, and LXR is an attractive target against multiple viruses. It has been suggested that broad-spectrum antivirals might help treat emerging infectious diseases caused by novel viral pathogens.^25–27) To expand the structural diversity of LXR-targeting antivirals, we screened a chemical library for identifying LXR antagonists and obtained cyclic phthalate esters 1 and 2 (Fig. 1). Our aim was to evaluate the antiviral activities of these discovered compounds.

Fig. 1. Structures of Compounds 1 and 2

Results and Discussion

We used a reporter assay with an LXR-responsive element (LXRE)-driven luciferase reporter plasmid to screen for LXR antagonists. The human liver carcinoma cell line, Huh-7, which is susceptible to HCV infection, was used for this screening. In our previous study, we constructed a library containing more than 1000 fractions, which were partially purified from the culture broths of fungi. Reporter assay-based screening of the library yielded several hit fractions that showed LXR antagonistic activity. A large-scale culture broth of a fungus producing the hit fraction was subjected to bioassay-guided fractionation to identify compounds 1 and 2 as LXR antagonists, as described in Experimental. We analyzed the structures of compounds 1 and 2 using NMR and MS spectroscopy. Compound 1 was a cyclic ester dimer consisting of phthalate and 1,6-hexane diol (Fig. 1). Compound 1 was previously isolated from the extract of Villebrunea pedunculata leaves as an antileishmanial compound and from earthworm fecal soil as an antifungal compound.^28,29) More recently, Yoshimoto et al. isolated a fluorescent compound (compound 1) from the exuviae of scorpions.³⁰⁾ In addition, the current ¹H- and ¹³C-NMR and MS data of compound 1 were consistent with those reported previously.³⁰⁾ Moreover, the ¹H- and ¹³C-NMR data of compound 2 were very similar to those of compound 1; however, slight differences in chemical shift values were observed between these compounds. High-resolution mass spectrometry (HRMS) revealed the molecular formula of compound 2 to be C₄₂H₄₈O₁₂, whereas that of compound 1 to be C₂₈H₃₂O₈. These data and heteronuclear multiple bond correlation (HMBC) spectral findings suggested that compound 2 has a symmetrical and cyclic structure (Supplementary Fig. S6), and it was identified as an unknown cyclic trimer of phthalate ester (Fig. 1). All ¹H and ¹³C signals were assigned on the basis of ¹H–¹H correlation spectroscopy (COSY), HMBC, and heteronuclear multiple quantum coherence (HMQC) spectra (Supplementary Table S1).

To the best of our knowledge, compounds 1 and 2 are not used as general plasticizers, while various other phthalate esters are used as general plasticizers. Therefore, we investigated whether compounds 1 and 2 were impurities that were originally present in the solvent. The chloroform used for the purification of compounds was concentrated and the resulting impurities were purified and analyzed using NMR and MS. We confirmed that compounds 1 and 2 were derived from concentrated chloroform, indicating that they were not fungal metabolites. Yoshimoto et al. conducted an experiment without a scorpion sample and excluded the possibility that compound 1 is a contaminant in solvents, suggesting that it is a naturally occurring compound derived from the scorpion.³⁰⁾ The structure of compound 2 has not been reported. In this study, we confirmed the structure; however, the source of these compounds could not be found.

Compounds 1 and 2 inhibited the LXR agonist T0901317-dependent transactivational function of LXR in the luciferase assay system²³⁾ (Fig. 2A). Compound 1 showed a concentration-dependent activity, whereas compound 2 showed a bell-shaped activity. In the same reporter assay system, high concentrations (>20 µM) of compound 2 inhibited the expression of both Fluc and βGal, suggesting that the activity of compound 2 was not specific for LXR at high concentrations, although the LXR-specific activity was observed at relatively lower concentrations. Thus, the bell-shaped dose response of compound 2 could be attributed to the non-specific activity. To exclude the possibility that LXR antagonism was caused by cytotoxicity, we evaluated the viability of cells treated with both compounds. Neither compound showed cytotoxicity at concentrations of 1–30 µM, whereas 100 µM of compound 1 resulted in a slight decrease in cell viability (Fig. 2B). Future structure–activity relationship studies may provide useful insights for decreasing non-specificity and cytotoxicity of these phthalate derivatives.

Fig. 2. LXR Antagonistic Activity (A) and Cytotoxicity (B) of Compounds 1 and 2

Huh-7 cells transfected with LXRE-Fluc, pSV-βGal, and expression plasmids for LXRα and retinoid X receptor alpha (RXR) were treated with the indicated concentrations of compounds 1 and 2 in the presence of T0901317 (0.3 µM) for 48 h (A). The cells were treated with various concentrations of compounds 1 and 2 for 72 h. The viability of each cell was evaluated using a WST-8 assay (B). Values are mean ± S.D. (n = 3).

Thereafter, compounds 1 and 2 were evaluated for their anti-HCV activity according to a previously described procedure. Compound 1 exhibited anti-HCV activity in a dose-dependent manner, and the corresponding IC₅₀ value was 2.6 µM (Fig. 3 and Supplementary Fig. S7). Meanwhile, the IC₅₀ value of sofosbuvir, a nucleotide analog used to treat hepatitis C, was 65 nM. Cyclic trimer 2 also displayed anti-HCV activity (Fig. 3). These findings suggest that LXR antagonism is closely related to the inhibition of HCV replication.

Fig. 3. Anti-HCV Activities of Compounds 1 and 2

Huh-7 cells, which were infected with HCV JFH-1, were cultured in a growth medium in the presence of various concentrations of compound 1, compound 2, or sofosbuvir for 72 h. The infectivity of HCV in the medium was quantified using a focus-forming assay with Huh-7.5.1 cells. Values are mean ± S.D. (n = 3).

As neoechinulin B functions as an LXR antagonist and is effective against both HCV and SARS-CoV-2, we evaluated the anti-SARS-CoV-2 activity of compounds 1 and 2 using VeroE6/TMPRSS2 cells. The production of SARS-CoV-2 RNA was reduced to 20% when compound 1 was used at a concentration of 40 µM, without causing cytotoxicity (Fig. 4). Meanwhile, compound 2 exerted toxicity against VeroE6/TMPRSS2 cells at a concentration of 40 µM. Moreover, it did not exhibit significant anti-SARS-CoV-2 activity at concentrations of 1–20 µM; no cytotoxicity was observed in this concentration range. Compound 1 showed both anti-HCV and -SARS-CoV-2 activities, suggesting that it may exhibit stable LXR antagonistic activity and inhibit DMV formation.

Fig. 4. Anti-SARS-CoV-2 Activity of Compounds 1 and 2

VeroE6/TMPRSS2 cells inoculated with SARS-CoV-2 were cultured in a medium with various concentrations of compounds 1 and 2 for 24 h. Extracellular viral RNA was quantified.

LXR activation may enhance the accumulation of lipid droplets (LDs) in cells by inducing the expression of SREBP-1c. In our previous study, we demonstrated that the LXR antagonist neoechinulin B suppresses the accumulation of LDs.²³⁾ To examine the effect of compound 1 on lipid accumulation, we analyzed the number of LDs in cells treated with T0901317 using our previously reported method. Figure 5A presents a typical image of BODIPY493/503-stained LDs of cells untreated or treated with compound 1. The LDs were classified into three categories based on size (1–2, 2–4, and >4 µm²). We counted LDs per size category in the treated and untreated cells (Fig. 5B). Compound 1 significantly reduced the LD count in all three size categories. The average number of LD per cell also reduced after treatment with compound 1 (Fig. 5C). These findings suggest that compound 1 reduced the accumulation of LDs in cells by suppressing the LXR transactivation function.

Fig. 5. Effect of Compound 1 on Lipid Accumulation in Cells

Huh-7 cells were treated with compound 1 (10 µM) in the presence of T0901317 (1 µM) for 72 h. The cells were stained with BODIPY493/503 (green) and DAPI (blue) and observed under a confocal microscope. A typical image of untreated or compound 1-treated Huh-7 cells. Scale bar = 10 µm (A). LDs in untreated or compound 1-treated cells were counted per LD size category (1–2, 2–4, and >4 µm²) and is shown as relative LD count per cell (B). Mean total LD counts in untreated or treated cells (C). Values are mean ± S.E. (n = 20). Statistical significance was determined using Welch’s t-test (*, p < 0.05; **, p < 0.01).

Compound 1 was synthesized from phthalic anhydride, following the method reported by Yoshimoto et al.³⁰⁾ Synthetic compound 1 was evaluated for the anti-HCV activity, revealing potency comparable to that of compound 1 obtained from concentrated chloroform (Supplementary Fig. S8). Non-cyclic phthalate esters, such as bis(2-ethylhexyl)phthalate and diethyl phthalate, are commonly used as plasticizers. To investigate whether the cyclic structure of compounds 1 and 2 is responsible for the anti-HCV activity, we assessed the anti-HCV activity of diisopropyl phthalate, a small non-cyclic phthalate ester. The compound did not inhibit HCV replication to 50% evaluated at concentrations of 1.25–20 µM, suggesting that the cyclic structures of compounds 1 and 2 are important to their activity (Supplementary Fig. S9). In addition, some plasticizers of phthalate esters may exhibit agonistic activity against LXRα through direct binding.^31,32) Further structure–activity relationship research is required to elucidate the relevance of cyclic structures of compounds 1 and 2 to exhibit LXR agonistic or antagonistic activity and antiviral properties.

Conclusion

Cyclic phthalates 1 and 2 were identified as antagonists for LXR with antiviral properties. The synthesis of several cyclic phthalate derivatives is in progress. Structure–activity relationship studies may yield more potent and specific LXR antagonists and antiviral compounds. We believe that the LXR antagonistic phthalate derivatives constitute a new class of broad-spectrum antivirals.

Experimental

General Experimental Procedures

The IR spectra were recorded using a JASCO FT/IR-4600 spectrophotometer (Jasco Corp., Tokyo, Japan). The ¹H- and ¹³C-NMR spectra were recorded in CDCl₃ using a Bruker 400 MHz spectrometer (Avance DRX-400 or Avance III-400; Bruker, Billerica, MA, U.S.A.) with tetramethylsilane (TMS) and CDCl₃ as internal references for ¹H- and ¹³C-NMR measurements, respectively. Chemical shifts are expressed in δ (ppm) relative to the TMS or residual solvent resonance and coupling constants (J) are expressed in Hz. Mass spectra were obtained using a JEOL mass spectrometer (JMS-700; JEOL, Tokyo, Japan). Analytical TLC was performed on precoated silica gel 60 F254 plates (Merck, Darmstadt, Germany). Silica gel 60N (Kanto Chemical, Tokyo, Japan) was used for silica gel column chromatography.

Identification of LXR Antagonists

Five liters of fungal culture broth was extracted using CH₂Cl₂. The crude extract was separated using silica gel column chromatography with CHCl₃–MeOH (10 : 0–9 : 1) and fractions 1–7 were obtained. Among them, fraction 2 exhibited LXR antagonistic activity and was further purified using HPLC (PEGASIL ODS SP100; Senshu, Tokyo, Japan; 10 × 150 mm, 5 µm; solvent 70–100% MeOH, flow rate 3.0 mL/min) to yield compound 1 (0.4 mg) and compound 2 (0.6 mg) as LXR antagonists.

Purification of Compounds 1 and 2 Using Chloroform

Chloroform (10 L, Kanto Chemical) was concentrated and the resultant mixture was separated via HPLC (PEGASIL Silica SP100; 10 × 250 mm; solvent 10–60% hexane/ethyl acetate, flow rate 4.0 mL/min) to obtain compounds 1 (1.7 mg) and 2 (0.6 mg).

Compound 1

¹H-NMR (CDCl₃) δ: 1.46 (8H, m), 1.75 (8H, m), 4.30 (8H, t, J = 6.7 Hz) 7.53 (4H, m), 7.72 (4H, m); ¹³C-NMR (CDCl₃) δ: 25.7, 28.5, 65.7, 128.9, 131.0, 132.2, 167.7; electrospray ionization (ESI)-MS m/z: 519 [M + Na]⁺

Compound 2

Colorless oil; ¹H- and ¹³C-NMR see Table S1; IR (film) cm⁻¹: 3020, 2937, 1722, 1522, 1289, 1216, 1136, 759; ESI-MS m/z 767.3051 (Calcd for C₄₂H₄₈O₁₂Na: 767.3043).

Luciferase Reporter Assay

On day 0, Huh-7 cells were plated on a 96-well plate in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, St. Louis, MO, U.S.A.), with 5% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, U.S.A.), 100 units/mL penicillin, and 100 µg/mL streptomycin sulfate (Thermo Fisher Scientific). On day 1, the cells were cotransfected with LXRE-FLuc (a reporter plasmid carrying the binding element of LXR upstream of the firefly luciferase), pSV-β-Gal (a reporter plasmid carrying SV40 promoter upstream of the β-galactosidase) along with expression plasmids for LXRα and RXR, using Lipofectamine reagent (Thermo Fisher Scientific). After incubation for 6 h, the cells were treated with the compound in the presence of T0901317 (Sigma-Aldrich). After 48 h of incubation, the cells in each well were lysed, and the luciferase activity and β-galactosidase activity were measured.

WST-8 Assay

WST-8 assay was performed as described previously.³³⁾

Anti-HCV Assay

The anti-HCV assay was performed as described previously.^23,24)

Anti-SARS-CoV-2 Assay

SARS-CoV-2 was handled at biosafety level 3 (BSL3). SARS-CoV-2 Wk-521 strain, a clinical isolate from a patient with coronavirus disease 2019 (COVID-19), was used for the anti-SARS-CoV-2 assay.³⁴⁾ The anti-SARS-CoV-2 assay was performed as described previously.^24,35)

Quantification of LDs

LDs were quantified as described previously.³⁶⁾ Huh-7 cells were treated with compound 1 (10 µM) in the presence of T0901317 (1 µM). After 72 h of incubation, the cells were fixed by incubating them in 4% paraformaldehyde followed by permeabilization with 0.05% saponin. The LDs and the nucleus were stained with BODIPY493/503 (Thermo Fisher Scientific) and 4′,6-diamidino-2-phenylindole (DAPI) (Nacalai Tesque, Kyoto, Japan), and the cells were observed under a confocal microscope (TCS SP5 II; Leica, Wetzlar, Germany). LDs were categorized into three groups based on their size (area) (1–2, 2–4, and >4 µm²) and their numbers were quantified using ImageJ software. The data were obtained by observing 20 randomly selected cells in each experiment.

Acknowledgments

We acknowledge the support of Mr. Yoshihisa Sei and the Materials Analysis Division, Open Facility Center, Tokyo Institute of Technology, for NMR analysis. This work was supported by the Ministry of Education, Culture, Sports, Science and Technology-Supported Program for the Private University Research Branding Project (2016–2020), the Japan Society for the Promotion of Science [KAKENHI 18K05343, 20H03499, 20K05868, and 21K05299], the Agency for Medical Research and Development (AMED) [JP21fk0310114, JP21jm0210068, JP22fk0310504, JP22wm0325007, JP21fk0108589, and JP20fk0210036], Takeda Science Foundation, and the Center for Human and Animal Symbiosis Science, Azabu University.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1) Cha J. Y., Repa J. J., J. Biol. Chem., 282, 743–751 (2007).
• 2) Li A. C., Glass C. K., J. Lipid Res., 45, 2161–2173 (2004).
• 3) Edwards P. A., Kast H. R., Anisfeld A. M., J. Lipid Res., 43, 2–12 (2002).
• 4) Tobin K. A., Ulven S. M., Schuster G. U., Steineger H. H., Andresen A. M., Gustafsson J. A., Nebb H. I., J. Biol. Chem., 277, 10691–10697 (2002).
• 5) Millatt L. J., Bocher V., Fruchart J. C., Staels B., Biochim. Biophys. Acta, 1631, 107–118 (2003).
• 6) Mitro N., Mak P. A., Vargas L., Godio C., Hampton E., Molteni V., Kreusch A., Saez E., Nature (London), 445, 219–223 (2007).
• 7) Joseph S. B., Laffitte B. A., Patel P. H., Watson M. A., Matsukuma K. E., Walczak R., Collins J. L., Osborne T. F., Tontonoz P., J. Biol. Chem., 277, 11019–11025 (2002).
• 8) Zhang Y., Repa J. J., Gauthier K., Mangelsdorf D. J., J. Biol. Chem., 276, 43018–43024 (2001).
• 9) Eberlé D., Hegarty B., Bossard P., Ferré P., Foufelle F., Biochimie, 86, 839–848 (2004).
• 10) Fang D., Wan Y., Shen W., Coa J., Sun Z. X., Yu H. H., Zhang Q., Cheng W. H., Chen J., Ning B., Mol. Cell. Biochem., 381, 127–137 (2013).
• 11) Brown M. S., Goldstein J. L., Proc. Natl. Acad. Sci. U.S.A., 96, 11041–11048 (1999).
• 12) Brown M. S., Goldstein J. L., Cell, 89, 331–340 (1997).
• 13) She J., Gu T., Pang X., Liu Y., Tang L., Zhou X., Front. Pharmacol., 12, 772435 (2022).
• 14) Lehmann J. M., Kliewer S. A., Moore L. B., Smith-Oliver T. A., Oliver B. B., Su J. L., Sundseth S. S., Winegar D. A., Blanchard D. E., Spencer T. A., Willson T. M., J. Biol. Chem., 272, 3137–3140 (1997).
• 15) Janowski B. A., Willy P. J., Devi T. R., Falck J. R., Mangelsdorf D. J., Nature (London), 383, 728–731 (1996).
• 16) Lin C., Gustafsson J., Nat. Rev. Cancer, 15, 216–224 (2015).
• 17) Chintalacharuvu S. R., Sandusky G. E., Burris T. P., Burmer G. C., Nagpal S., Arthritis Rheum., 56, 1365–1367 (2007).
• 18) Wang Y., Rogers P. M., Stayrook K. R., Su C., Varga G., Shen Q., Nagpal S., Burris T. P., Mol. Pharmacol., 74, 1716–1721 (2008).
• 19) Varin A., Thomas C., Ishibashi M., Menegaut L., Gautier T., Trousson A., Bergas V., de Barros J. P. P., Narce M. A., Lobaccaro J. M., Lagrost L., Masson D., Arterioscler. Thromb. Vasc. Biol., 35, 1357–1365 (2015).
• 20) Griffett K., Solt L. A., El-Gendy B. E. M., Kamenecka T. M., Burris T. P., ACS Chem. Biol., 8, 559–567 (2013).
• 21) El-Gendy B. E. M., Goher S. S., Hegazy L. S., Arief M. M. H., Burris T. P., J. Med. Chem., 61, 10935–10956 (2018).
• 22) Noguchi-Yachide T., Miyachi H., Aoyama H., Aoyama A., Makishima M., Hashimoto Y., Chem. Pharm. Bull., 55, 1750–1754 (2007).
• 23) Nakajima S., Watashi K., Ohashi H., et al., J. Virol., 90, 9058–9074 (2016).
• 24) Nishiuchi K., Ohashi H., Nishioka K., Yamasaki M., Furuta M., Mashiko T., Tomoshige S., Ohgane K., Kamisuki S., Watashi K., Kuramochi K., J. Nat. Prod., 85, 284–291 (2022).
• 25) Andersen P. I., Ianevski A., Lysvand H., Vitkauskiene A., Oksenych V., Bjoras M., Telling K., Lutsar I., Dumpis U., Irie Y., Tenson T., Kantele A., Kainov D. E., Int. J. Infect. Dis., 93, 268–276 (2020).
• 26) Ianevski A., Zusinaite E., Kuivanen S., et al., Antiviral Res., 154, 174–182 (2018).
• 27) Adamson C. S., Chibale K., Goss R. J. M., Jaspars M., Newman D. J., Dorrington R. A., Chem. Soc. Rev., 50, 3647–3655 (2021).
• 28) Matsunami K., Ochiba S., JP 2016160237 (2016).
• 29) Sugawara F., Kamakura T., JP 2013180995 (2013).
• 30) Yoshimoto Y., Tanaka M., Miyashita M., Abdel-Wahab M., Megaly A. M. A., Nakagawa Y., Miyagawa H., J. Nat. Prod., 83, 542–546 (2020).
• 31) Mozzicafreddo M., Cuccioloni M., Bonfili L., Cecarini V., Palermo F. A., Cocci P., Mosconi G., Capone A., Ricci I., Eleuteri A. M., Angeletti M., J. Steroid Biochem. Mol. Biol., 152, 1–7 (2015).
• 32) Kanayama T., Kobayashi N., Mamiya S., Nakanishi T., Nishikawa J., Mol. Pharmacol., 67, 766–774 (2005).
• 33) Myobatake Y., Kamisuki S., Tsukuda S., Higashi T., Chinen T., Takemoto K., Hachisuka M., Suzuki Y., Takei M., Tsurukawa Y., Maekawa H., Takeuchi T., Matsunaga T. M., Sahara H., Usui T., Matsunaga S., Sugawara F., Bioorg. Med. Chem., 27, 115149 (2019).
• 34) Matsuyama S., Nao N., Shirato K., Kawase M., Saito S., Takayama I., Nagata N., Sekizuka T., Katoh H., Kato F., Sakata M., Tahara M., Kutsuna S., Ohmagari N., Kuroda M., Suzuki T., Kageyama T., Takeda M., Proc. Natl. Acad. Sci. U.S.A., 117, 7001–7003 (2020).
• 35) Ohashi H., Watashi K., Saso W., et al., iScience, 24, 102367 (2021).
• 36) Ohashi H., Nishioka K., Nakajima S., Kim S., Suzuki R., Aizaki H., Fukasawa M., Kamisuki S., Sugawara F., Ohtani N., Muramatsu M., Wakita T., Watashi K., J. Biol. Chem., 293, 19559–19571 (2018).