Molecular Identification of UDP-Sugar-Dependent Glycosyltransferase and Acyltransferase Involved in the Phenylethanoid Glycoside Biosynthesis Induced by Methyl Jasmonate in Sesamum indicum L.

Abstract

Sesame (Sesamum indicum L.) plants contain large amounts of acteoside, a typical phenylethanoid glycoside (PhG) that exhibits various pharmacological activities. Although there is increasing interest in the biosynthesis of PhGs for improved production, the pathway remains to be clarified. In this study, we established sesame-cultured cells and performed transcriptome analysis of methyl jasmonate (MeJA)–treated cultured cells to identify enzyme genes responsible for glucosylation and acylation in acteoside biosynthesis. Among the genes annotated as UDP-sugar-dependent glycosyltransferase (UGT) and acyltransferase (AT), 34 genes and one gene, respectively, were upregulated by MeJA in accordance with acteoside accumulation. Based on a phylogenetic analysis, five UGT genes (SiUGT1–5) and one AT gene (SiAT1) were selected as candidate genes involved in acteoside biosynthesis. Additionally, two AT genes (SiAT2–3) were selected based on sequence identity. Enzyme assays using recombinant SiUGT proteins revealed that SiUGT1, namely, UGT85AF10, had the highest glucosyltransferase activity among the five candidates against hydroxytyrosol to produce hydroxytyrosol 1-O-glucoside. SiUGT1 also exhibited glucosyltransferase activity against tyrosol to produce salidroside (tyrosol 1-O-glucoside). SiUGT2, namely, UGT85AF11, had similar activity against hydroxytyrosol and tyrosol. Enzyme assay using the recombinant SiATs indicated that SiAT1 and SiAT2 had activity transferring the caffeoyl group to hydroxytyrosol 1-O-glucoside and salidroside (tyrosol 1-O-glucoside) but not to decaffeoyl-acteoside. The caffeoyl group was attached mainly at the 4-position of glucose of hydroxytyrosol 1-O-glucoside, followed by attachment at the 6-position and the 3-position of glucose. Based on our results, we propose an acteoside biosynthetic pathway induced by MeJA treatment in sesame.

Introduction

Phenylethanoid glycosides (PhGs) are natural products mainly from Lamiales order plants and are widely used as traditional medicines worldwide (Alipieva et al. 2014). The core structure of PhGs consists of a C₆–C₂ skeleton with aldohexose/aldopentose attached by a glycosidic bond (Fig. 1A). The structures are modified with various substituents, such as aromatic organic acids (mainly hydroxycinnamoyl units like cinnamoyl, p-coumaroyl and caffeoyl moieties) and saccharides (glycosyl units like glucosyl, rhamnosyl and galactosyl moieties), through ester or glycoside bonds, respectively (Xue and Yang 2016).

Acteoside, which is also called kusagin or verbascoside, is the most typical PhG and consists of hydroxytyrosol 1-O-glucoside with rhamnose attached to the glucose moiety at the 3-position and caffeic acid at the 4-position (Fig. 1B). Acteoside has been reported in more than 150 plant species belonging to 20 families and 77 genera of mainly the order Lamiales such as Olea europaea, Rehmannia glutinosa and Plantago asiatica (He et al. 2011). Acteoside has various biological activities such as antioxidative, anti-inflammatory (Seo et al. 2013), hepatoprotective (Jing et al. 2015) and neuroprotective (Koo et al. 2005) activities. Previous research has shown that acteoside could be a potential therapeutic compound for Alzheimer’s disease (Korshavn et al. 2015, Shiao et al. 2017) and Parkinson’s disease (Yuan et al. 2016). Current knowledge of the acteoside biosynthetic pathway is based on feeding experiments (Ellis 1988, Saimaru and Orihara 2010) and transcriptome analyses in O. europaea L., R. glutinosa, Forsythia sp. and Cistanche tubulosa (Alagna et al. 2012, Wang et al. 2017, Sun et al. 2018, Hou et al. 2022, Yuan et al. 2022). As shown in Fig. 2A, the caffeoyl moiety of acteoside is biosynthesized from phenylalanine through the phenylpropanoid pathway, whereas the hydroxytyrosol moiety is biosynthesized from tyrosine with tyramine and/or dopamine intermediates. The phenylpropanoid pathway from phenylalanine to caffeoyl-CoA is common in flavonoid and lignin biosynthesis, and the enzyme genes involved in this pathway have been extensively characterized in various plant species (Winkel-Shirley 2001, Vanholme et al. 2019). On the other hand, genes encoding the enzymes responsible for the formation of hydroxytyrosol from tyrosine [namely, polyphenol oxidase (PPO), copper-containing amine oxidase (CuAO) and alcohol dehydrogenase (ALDH)] have not been identified. The in vivo enzyme activity of tyrosine decarboxylase (TyDC) has been reported in R. glutinosa, suggesting its involvement in acteoside biosynthesis (Yang et al. 2022). The downstream pathway from caffeoyl-CoA, salidroside and hydroxytyrosol to acteoside involves hydroxylation, glycosylation and acylation (Fig. 2B), but the order of these reactions and the corresponding enzyme genes have not yet been identified.

Our previous studies revealed that acteoside is highly accumulated in sesame leaf blades, up to 12.3% (weight/dry weight) (Fuji et al. 2018a, 2018b). The accumulation in sesame leaves is among the highest levels reported in plants in nature, and thus, we chose sesame as an excellent experimental material for acteoside biosynthesis. In this study, we established a callus of sesame (Sesamum indicum L.) and its suspension culture. Based on transcriptome analysis of methyl jasmonate (MeJA)–treated cultured cells and phylogenetic analysis, we selected five genes annotated as UDP-sugar-dependent glycosyltransferase (UGT; SiUGT1–5) and three genes annotated as acyltransferase (AT; SiAT1–3) for the candidate genes involved in the downstream pathway of acteoside biosynthesis. Enzyme assays using their recombinant proteins revealed that SiUGT1, SiUGT2, SiAT1 and SiAT2 had enzymatic activities responsible for the skeletal formation of acteoside. The result suggests a possible route in the downstream pathway of acteoside biosynthesis in sesame.

Results

Establishment of PhG-producing cultured cell of sesame

To establish sesame-cultured cells, explants such as cotyledon and hypocotyl were prepared from 7-day-old sterile sesame plants and cultured on Murashige–Skoog solid medium supplemented with hormones to induce callus formation (Supplementary Fig. S1A–D). Then, the callus was transferred to Murashige–Skoog liquid medium (Supplementary Fig. S1E, F), and the compounds in the cultured cells incubated in liquid medium were analyzed by liquid chromatography coupled with photodiode array and mass spectrometry (LC-PDA–MS). Our previous studies (Fuji et al. 2018b) have revealed that sesame leaves produce specialized metabolites such as iridoids and flavonoids in addition to acteoside (Supplementary Fig. 1G), whereas sesame culture cells produced acteoside as the main metabolite (0.027 ± 0.004% fresh weight (FW) basis), with no iridoids or flavonoids detected (Supplementary Fig. 1H). The UV and mass spectra suggested that the minor components other than acteoside were PhGs in which sugar and/or hydroxyl groups were added to acteoside.

Induction of acteoside biosynthesis by MeJA in sesame-cultured cell

We examined the conditions under which acteoside accumulation is induced using established culture cells and various elicitors.

First, we analyzed time-dependent changes in acteoside content during the culture period (Supplementary Fig. S2). The acteoside content in cultured cells did not change for 7 d after transfer to liquid medium. From days 7 to 11 after transfer, the acteoside content increased. The content was stable from 11 to 16 d after transfer, gradually increased until 26 d after transfer reaching a maximum of 0.053% FW and then decreased greatly. We also analyzed the culture medium during the culture period by LC-PDA and confirmed that acteoside was not excreted into the medium (Supplementary Fig. S3).

Next, based on this result, we performed elicitor treatment during the period of 11 to 16 d after transfer. When cultured cells were treated with MeJA, salicylic acid (SA), abscisic acid (ABA), AgNO₃ or CuCl₂, only MeJA induced acteoside accumulation, reaching 0.199 ± 0.3% FW after 72 h of treatment (Fig. 3A). The LC-PDA–MS chromatograms showed that only the acteoside content increased significantly by MeJA with treatment time (Fig. 3B). On the other hand, the acteoside content decreased along with the treatment time by SA, ABA, AgNO₃ and CuCl₂ (Fig. 3A; Supplementary Fig. S4).

Transcriptome analysis of MeJA-treated sesame-cultured cells

The time-dependent change of the acteoside content by MeJA treatment was examined in more detail to determine appropriate time points for transcriptome analysis (Supplementary Fig. S5). Based on the results, the samples at 0, 6, 12 and 24 h after the onset of MeJA treatment were subjected to RNA sequencing (RNA-seq) analysis. By analyzing RNA-seq data, we found 218 genes with sequence lengths longer than 200 bp, which were annotated as enzymes presumably responsible for acteoside biosynthesis (Supplementary Table S2), namely, phenylalanine ammonia-lyase (PAL, five genes), cinnamate 4-hydroxylase (C4H, three genes), coumarate-3-hydroxylase (C3H, three genes), 4-coumarate-CoA ligase (4CL, 18 genes), TyDC (five genes), PPO (seven genes), CuAO (five genes), ALDH (18 genes), UGT (151 genes) and AT (three genes). To identify the genes involved in acteoside biosynthesis, we performed a hierarchical clustering analysis of the 218 genes. The genes were classified into eight clusters based on their time-dependent expression patterns (Fig. 4A, Supplementary Table S2). Genes in cluster 3 were upregulated by MeJA after 6 h, prior to the acteoside accumulation (Fig. 3), suggesting that they were involved in acteoside biosynthesis. For each enzyme, at least one gene was upregulated by MeJA (Fig. 4B). The genes in clusters 1 and 2 were upregulated by MeJA, but the induction was preceded by the accumulation of acteoside.

Selection of candidate UGTs and ATs involved in acteoside biosynthesis

In this study, we focused on UGTs and ATs to identify enzymes involved in the formation of core structure (Fig. 1A) and the acylation step in the downstream pathway of acteoside biosynthesis. To further narrow down the candidates among the 34 UGT genes in cluster 3, five genes were selected based on the amino acid identity with UGT85A1 of Arabidopsis thaliana and UGT72B14 of Rhodiola sachalinensis, which have UGT activity against tyrosol (Yu et al. 2011, Chung et al. 2017), namely, SiUGT1–5. In a phylogenetic tree of UGTs, UGTs SiUGT1–4 were classified into the UGT85 family, while SiUGT5 was classified into the UGT72 family (Supplementary Fig. S6). The predicted coding sequence (CDS) of SiUGT1–5 encoded 472, 474, 479, 443 and 477 amino acids, respectively. These amino acid sequences had a typical plant secondary product glycosyltransferases-box sequence including highly conserved key residues for substrate recognition (Supplementary Fig. S7). These UGTs were renamed UGT85AF10 (SiUGT1), UGT85AF11 (SiUGT2), UGT85A149 (SiUGT3), UGT85K53 (SiUGT4) and UGT72B79 (SiUGT5) by the UGT Nomenclature Committee according to Mackenzie et al. (1997).

Concerning ATs, only one AT gene was classified into cluster 3 and we named it SiAT1. In addition, two genes showing high sequence identity and having the same annotation as SiAT1 (Supplementary Table S2) were selected for further analysis and named SiAT2 and SiAT3, although these genes were not upregulated by MeJA (Supplementary Fig. S8). The deduced amino acid sequences of three genes had the HXXXD and DFGWG motifs conserved in the BAHD-AT family (Supplementary Fig. S9), and they belonged to the BAHD-AT family in the phylogenetic tree (Supplementary Fig. S10).

Enzymatic activity and substrate specificity of SiUGTs

The glucosyltransferase activity of the recombinant SiUGT1–5 toward C₆–C₂ compounds, hydroxytyrosol, tyrosol and phenethyl alcohol (Fig. 5A), was investigated. When the purified recombinant SiUGT1 was incubated with hydroxytyrosol and UDP-glucose and the reaction mixture was analyzed by LC-PDA–MS, a new peak was observed at a retention time (Rt) of 5.84 min, which was consistent with that of the authentic hydroxytyrosol 1-O-glucoside (Fig. 5B, left). To determine the structure of the product by LC–tandem mass spectrometry (LC–MS/MS) and NMR analyses, the product was produced on a large scale by bioconversion assay in which SiUGT1-expressing Escherichia coli was cultured in the medium supplemented with hydroxytyrosol (Supplementary Fig. S11A) and the product compound was purified. In the LC–MS/MS analysis, the product gave a mass spectrum similar to that of the authentic hydroxytyrosol 1-O-glucoside (Supplementary Fig. S12A, left). NMR analysis further confirmed that the product was hydroxytyrosol 1-O-glucoside (Supplementary Table S3). When the purified recombinant SiUGT1 was incubated with tyrosol, another peak was observed at Rt 6.73 min in the LC–UV chromatogram, which was considered to be salidroside (tyrosol 1-O-glucoside) based on comparison with the standard compound (Fig. 5B, middle). LC–MS/MS analysis of the bioconversion product supported the presumption (Supplementary Fig. S12A, right). The in vitro assay using phenethyl alcohol showed that SiUGT1 was slightly active to this compound (Fig. 5B, right). In the same way, the activity of SiUGT2–5 toward C₆–C₂ compounds was analyzed (Fig. 5C–E). SiUGT2 showed similar substrate specificity to SiUGT1 (Fig. 5C), but the activity was weaker than SiUGT1 (Fig. 5D). SiUGT3–5 exhibited considerably weaker activity toward C₆–C₂ compounds (Fig. 5D, E). Among the members of the UGT85 family, SiUGT1 and SiUGT2 showed higher specific activity, while SiUGT3 and SiUGT4 showed almost no activity. SiUGT5 belonging to the UGT72 family was slightly active only to hydroxytyrosol.

Furthermore, when the recombinant ATs SiUGT1–5 were incubated with C₆–C₃ compounds (Fig. 5A) as a substrate, no new peaks were observed in LC–UV chromatograms (data not shown). Fig. 5E summarizes the substrate specificity of SiUGT1–5 toward C₆–C₂ and C₆–C₃ compounds tested. The result suggested that SiUGT1–5 preferentially transfer the glucose moiety to the alcoholic hydroxyl group of the C₆–C₂ compounds such as tyrosol and hydroxytyrosol.

Enzymatic activity and substrate specificity of SiATs

As shown in Fig. 2B, it remains to be clarified whether rhamnosylation or acylation occurs first in the downstream pathway leading to acteoside. In this study, we investigated substrates of the candidate ATs (SiAT1–3) by in vitro assay with possible acyl acceptors (hydroxytyrosol 1-O-glucoside, decaffeoyl-acteoside and salidroside shown in Fig. 6A) using caffeoyl-CoA as an acyl donor. First, the purified recombinant ATs SiAT1–3 were incubated with hydroxytyrosol 1-O-glucoside and the reaction mixture was analyzed by LC–MS/MS in multiple reaction monitoring (MRM) mode. Using the MRM conditions optimized for detection of a presumed product calceolarioside A (Fig. 6C; MRM transition m/z 477.2 > 161.15), no peaks were observed in the SiAT3 assay, while three peaks were observed at Rt 16.76, 17.35 and 17.68 min in the SiAT1 and SiAT2 assays (Fig. 6B). The major peak at Rt 17.35 and the second major peak at 17.68 min were considered to be calceolarioside A (hydroxytyrosol 1-O-glucoside with the caffeoyl moiety attached to the 4-position of glucose) and calceolarioside B (that attached to the 6-position of glucose), respectively, in comparison with Rt and mass spectrum of the standard compounds (Fig. 6B, Supplementary Fig. S12B). The peak at Rt 16.76 min showed a mass spectrum similar to that of the other two peaks (Supplementary Fig. S11B), suggesting that these three peaks were regioisomers with a different caffeoyl binding position. To confirm the structure of these products by LC–MS/MS and NMR analyses, bioconversion assay was performed. The SiUGT1_At4CL2_SiAT1 gene-expressing E. coli was cultured in the medium supplemented with hydroxytyrosol and caffeic acid, where hydroxytyrosol 1-O-glucoside and caffeoyl-CoA were produced via the reactions catalyzed by SiUGT1 and At4CL2, respectively. The assay revealed three peaks in the LC–UV chromatogram, which corresponded to the peaks observed in the in vitro assay (Supplementary Fig. S11B). These three products were purified and identified as plantainoside A, calceolarioside A and calceolarioside B, based on mass spectra, ¹H-NMR spectra and ¹³C-NMR spectra (Fig. 6C, Supplementary Fig. 12B, Supplementary Table S3) (Damtoft and Jensen 1994, Jensen 1996).

Next, the AT activity toward salidroside was investigated. Because the standard compound for caffeoyl salidroside was not available, we analyzed using the presumed MRM transition of this compound (m/z 461.3 > 161.2). When salidroside was incubated in vitro with the recombinant SiAT1 or SiAT2, three peaks were observed as in the case of hydroxytyrosol 1-O-glucoside and were considered to be regioisomers of caffeoyl salidroside (Fig. 6D). No peak was observed in the SiAT3 assay with salidroside. On the other hand, when decaffeoyl-acteoside was used as a substrate in the SiAT1–3 assays, no peak was observed using the MRM conditions optimized for acteoside (MRM transition m/z 623.2 > 161.2) (Fig. 6E). The result indicated that SiAT1 and SiAT2 catalyzed caffeoylation of hydroxytyrosol 1-O-glucoside and salidroside. The activity of SiAT1 seemed higher than that of SiAT2, because the SiAT1 assay gave higher peak intensity (Fig. 6B, D) under the same reaction conditions using 30 μg of recombinant protein for 60 min.

Overall, these results suggested that SiAT1 and SiAT2 were able to transfer the caffeoyl moiety to the 3-, 4- or 6-position of the glucose moiety of hydroxytyrosol 1-O-glucoside and salidroside, while SiAT3 did not produce any compound detected under the MRM conditions used when assayed with hydroxytyrosol 1-O-glucoside, decaffeoyl-acteoside or salidroside.

Discussion

In this study, we identified MeJA-inducible SiUGT1 (UGT85AF10) and SiUGT2 (UGT85AF11), which transferred the glucose moiety to the alcoholic hydroxyl group of tyrosol and hydroxytyrosol in the in vitro and bioconversion assays. Regarding the known functions of the UGT85 family, UGT85A1 of Arabidopsis thaliana is involved in the regulation of a plant hormone trans-zeatin through O-glucosylation (Jin et al. 2013). UGT85A24 of Gardenia jasminoides showed glucosylation activity toward iridoid and preferentially glucosylated the 1-O-hydroxyl group of 7-deoxyloganetin and genipin (Nagatoshi et al. 2011). UGT85K11 from Camellia sinensis was reported to catalyze the formation of geranyl glucoside and have volatile glycosylatoin function (Ohgami et al. 2015). Thus, UGT85 family enzymes play a role in the formation of aromatic glycosides in various plants. The specificity of SiUGT1 and SiUT2 was in contrast to UGT73B6 identified from R. sachalinensis (Crassulaceae, Saxifragales) (Ma et al. 2007), which attached glucose to both the phenolic hydroxyl group and the alcoholic hydroxyl group of tyrosol, indicating that UGT73B6 may not be specific to salidroside synthesis (Bai et al. 2014). We also identified SiUGT5 (UGT72B79). Although it showed 60% identity with UGT72B14, overexpression of which in the transgenic hairy roots of R. sachalinensis led to a 420% increase of the salidroside content (Yu et al. 2011), it exhibited much lower activity than SiUGT1 and SiUGT2 only toward hydroxytyrosol (Fig. 5C).

The specific activity of SiUGT1 toward hydroxytyrosol (18.6 ± 0.56 nkatal/mg) was approximately 2-fold higher than that toward tyrosol (10.1 ± 1.9 nkatal/mg) and SiUGT2 toward hydroxytyrosol and tyrosol (9.5 ± 0.3 and 10.1 ± 1.9 nkatal/mg, respectively). In addition, the expression level of SiUGT1 was higher by 2- to 3-fold than that of SiUGT2 in the MeJA-treated sesame-cultured cells (Supplementary Table S2). Among acteoside-producing plants, both hydroxytyrosol 1-O-glucoside and salidroside (tyrosol 1-O-glucoside) were detected in olive (Romero et al. 2002, Melliou et al. 2015), while only salidroside was detected in Plantaginaceae plants (Taskova et al. 2006). In our previous study, neither hydroxytyrosol 1-O-glucoside nor salidroside was detected in any part of the sesame plant (Fuji et al. 2018a). Taking all these observations into account, it is possible that the substrates of the C₆–C₂ glycosides may differ among the Lamiales plants.

We also identified MeJA-inducible SiAT1, which transferred the caffeoyl moiety to hydroxytyrosol 1-O-glucoside and salidroside (tyrosol 1-O-glucoside) in the in vitro and bioconversion assays. When decaffeoyl-acteoside was used as a substrate in the SiAT1 assay, acteoside was not detected in the reaction mixture (Fig. 6E). This result strongly suggested that rhamnosylation occurs after caffeoylation (Fig. 7). It is likely that rhamnose bound to the glucose moiety of hydroxytyrosol 1-O-glucoside sterically inhibited the caffeoylation reaction. Based on these results, we propose the downstream pathway of acteoside biosynthesis in sesame (Fig. 7). Identification and characterization of oxygenase and rhamnosyltransferase would clarify the order of the reactions. In the SiAT1 assay, three products were produced: in the case of hydroxytyrosol 1-O-glucoside as a substrate, calceolarioside A (caffeoyl group attached to the 4-position of glucose) and calceolarioside B (caffeoyl group attached to the 6-position of glucose) were the major and second major products, respectively, while plantainoside A (caffeoyl group attached to the 3-position of glucose) was a minor one (Fig. 6B–D). In planta, acylation to the 4-O position of glucose is likely to be the main reaction, because acteoside (caffeoyl moiety at the 4-O position of glucose) is the main component and isoacteoside (caffeoyl moiety at the 6-O position of glucose) is a minor one (Fuji et al. 2018b). On the other hand, there is a possibility that those regioisomers in the SiAT1 assay were non-enzymatically formed because the intramolecular transition of the 4-O caffeoyl group of acteoside to the 6-O position was observed by using sodium hydroxide (Birkofer et al. 1968, Schilling et al. 1982) or in alkaline solutions (Kawada et al. 2002).

As the last enzyme involved in rhamnosylation, Yang et al. (2021) reported that UGT79G7 of Ligustrum robustum (Oleaceae, Lamiales) acted as 1,3-rhamnosyltransferase of osmanthuside A (p-coumaroyl salidroside). However, the sesame gene (XM_020694616.1 in Supplementary Table S1), which is highly homologous (81%) to UGT79G7 (GenBank accession number, MZ734611), was not induced by MeJA, and its expression at 0 h was extremely low. This suggests that other glycoside-specific glycosyltransferases, rather than a homolog of UGT79G7, is likely to be involved in rhamnosylation, at least in sesame (Ono et al. 2020).

Among the current methods for obtaining acteoside, extraction from natural plants is particularly difficult due to its extremely low content (Alipieva et al. 2014). Chemical synthesis mainly relies on multiple tedious, low-yielding steps (Duynstee et al. 1999, Kawada et al. 1999, 2002). Then, cell suspension culture has been used for the improvement of acteoside production. In this study, we established a sesame cell culture and induced acteoside production by MeJA treatment. To date, studies have been performed mainly in the cell suspension culture of Cistanche deserticola elicited by chitosan, MeJA, SA or fungal infection (Lu and Mei 2003, Xu et al. 2005, Cheng et al. 2006). The increase in PhG accumulation in C. deserticola cell suspension cultures was attributed to the increase in PAL activity stimulated by the chitosan elicitor (Lu and Mei 2003). Wang et al. (2017) performed elicitation of R. glutinosa hairy roots and reported that SA and MeJA induce acteoside accumulation, with SA being particularly effective. However, in the present study, SA was not effective in sesame-cultured cells, suggesting that the acteoside accumulation in S. indicum occurs through different regulation mechanisms than in other plant species. In this study, acteoside accumulation dramatically increased by 6.6-fold by MeJA and production reached 199.5 ± 30.6 mg/l within 3 d after the onset of treatment. This suggests that sesame-cultured cells have the potential to be an excellent system for acteoside bioproduction. To further improve acteoside production, it would be important to understand the PhG biosynthesis mechanism in sesame.

Materials and Methods

Chemicals

Acteoside was purchased from BP Biochemicals Inc. (San Diego, CA, USA). Pedalitin, pedaliin, lamalbid, sesamoside and shanzhiside methyl ester were isolated previously (Fuji et al. 2018b). Decaffeoyl-acteoside and calceolarioside B were purchased from ChemFaces (Wuhan, China). Calceolarioside A was purchased from Namiki Shoji Co. (Tokyo, Japan). All reagents were used without further purification.

Establishment of callus and cell suspension culture of sesame

Callus induction was performed as follows with reference to the method of Heidaeifar and Nayeri (2015). Sesamum indicum L. variety Myanmar Black Sesame, which is lignan-rich black sesame, was used. Sesame seeds were soaked with 70% ethanol, washed three times with sterilized water, then, sterilized with 2% sodium hypochlorite for 5 minutes and washed three times with sterilized water. The seeds were cultured on Murashige–Skoog medium (30 g/l sucrose, 9 g/l agar, pH 5.7–5.8) for 7 d in the dark at 25 ± 2°C in a growth chamber. Explants such as cotyledon and hypocotyl were prepared from 7-day-old aseptic sesame plants and cultured on the same medium as earlier supplemented with 3 mg/l naphthalene acetic acid and 0.6 mg/l 6-benzyl amino purine in a growth chamber at 25°C with a 16-h photoperiod to induce callus. Callus cultures were maintained with subculturing every 4 weeks.

Cell suspension cultures were initiated using a 4-week-old fresh friable callus. Callus (15 ± 0.1 g) as initial inoculum was transferred to a 300 ml Erlenmeyer flask containing 100 ml of the same liquid medium. Cultures were incubated in the dark at 25 ± 2°C with continuous agitation at 110 rpm in a shaker, and pre-culture was performed for 2 weeks. Then, 25 ml of pre-culture was transferred to a 300 ml Erlenmeyer flask containing 100 ml of the same medium, and the main culture was performed under the same conditions as the pre-culture. The main culture was used in various experiments.

Elicitor treatment of cell suspension culture

Cell suspension cultures of 11- to 16-days-after-passage samples were used for elicitor treatment. MeJA, SA and ABA were dissolved in dimethyl sulfoxide, while AgNO₃ and CuCl₂ were dissolved in distilled H₂O. All the solutions were filter-sterilized through 0.22 μm filters and added to cultures to a final concentration of 250 μM. After 0, 24, 48 and 72 h treatment with elicitors, cell suspension cultures were collected for determination of acteoside contents. MeJA-treated cell suspension cultures at the 0, 6, 12 and 24 h time points were used for RNA-seq analysis. The collected cell suspension cultures were separated into cells and culture medium by suction filtration, and only the cells were used for various analyses.

Total RNA isolation and cDNA synthesis

Total RNA was isolated from the powdered frozen samples using the SV Total RNA Isolation System (Promega, Madison, WI, USA). cDNA for cloning was synthesized using the SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA, USA).

Transcriptome analysis

RNA-seq using extracted total RNAs was outsourced to Eurofins Genomics (Tokyo, Japan) and performed with a HiSeq2500 under the following conditions: sequence mode, 100 bp paired-end; software, HiSeq Control Software 2.2.58, RTA 1.18.64 and bcl2fastq 1.8.4; library kit, HiSeq SBS Kit v4. Obtained reads were cleaned by FastQ Quality Control Software (Lo and Chain 2014), and transcripts per kilobase million (TPM) values were calculated by Kallisto (Bray et al. 2016) using the transcript sequence (GCF_000512975.1_S_indicum_v1.0_rna) constructed from S_indicum_v1.0 genome (Wang et al. 2014). Finally, data of 38,621 genes were obtained and used for further analysis (Supplementary Table S1). The TPM values were Z-score-transformed and scaled across each row, and heatmap was generated using R with the pheatmap package.

Cloning

All CDSs were amplified by PCR using PrimeSTAR MAX DNA Polymerase (Takara, Shiga, Japan) with specific primer sets and cDNA (12-h MeJA-treated culture cells) as the template. Primer sequences are as follows: SiUGT1, 5'-ATGAAGCCGCACGCTGTATTG-3' (forward) and 5'-CTAGTTCTTGAGGAGTATCTC-3' (reverse); SiUGT2, 5'-ATGGGTTCCCTACCAAGACC-3' and 5'-TTATTGCAAAAGCAATTCATTGATC-3'; SiUGT3, 5'-ATGAAGCCTCACGCAGTAG-3' and 5'-CTAGTTGTTGTTGAGGAGAG-3'; SiUGT4, 5'-ATGATGCAACTGGCCAGAC-3' and 5'-TCAAAGTAAATGGGGCCTTAATG-3'; SiUGT5, 5'-ATGGAAACCGCCTCC-3' and 5'-TTATATGCAAGTGTTACTCTTCC-3'; SIAT1, 5'-ATGGTGACTTTCAAGG-3' and 5'-TCAGATATCATCGTAGAAAC-3'; SiAT2, 5'-ATGGTGGTGGTTGTC-3' and 5'-TTACTTATCAAATATGTCATAG-3'; SiAT3, 5'-ATGCACCTCACTATCC-3' and 5'-TCATATTGCTGAGCC-3'. Amplified fragments were introduced into pCR8/GW/TOPO vector (Thermo Fisher Scientific), and CDSs were introduced into the pET-53-DEST vector (Merck Millipore, Bedford, MA, USA) using a Gateway LR Clonase II Enzyme Mix (Thermo Fisher Scientific).

Heterologous expression and purification of the candidate UGTs and ATs

E. coli strain C41 (DE3) (Lucigen, Middleton, WI, USA) was transformed with the pET-53-DEST vector to generate an N-terminal 6 × histidine fusion protein. The transformed cells were pre-cultured in LB medium supplemented with 50 μg/ml carbenicillin overnight at 37°C. A 10 ml aliquot of the overnight culture was used to inoculate the same medium (200 ml). The fresh culture was incubated at 37°C until the Optical Density (OD)₆₀₀ reached 0.4–0.6. Isopropyl 1-β-d-thiogalactoside (IPTG) was then added to the medium at a final concentration of 1.0 mM, and the culture was incubated at 18°C for 24 h.

All subsequent steps were conducted at 0–4°C. E. coli cells induced overnight by IPTG were collected by centrifugation (5,000×g, 5 min). For purification of recombinant proteins, the cells were resuspended in 5 ml of buffer (buffer A for SiUGTs, 50 mM potassium phosphate, pH 7.0, 0.05% 2-mercaptoethanol; buffer B for SiATs, 100 mM Tris-HCl buffer pH8.0, 10 mM EDTA) supplemented with 25 μl of lysozyme (TaKaRa, Shiga, Japan) and disrupted by vortex with glass beads for 5 min. Cell debris was removed by centrifugation (11,000×g at 4°C for 5 min), and the supernatant was filtered through a 0.22 μm filter and applied to a HisTrap HP column (1 ml; Cytiva, Tokyo, Japan) equilibrated with buffer A. The column was extensively washed with wash buffer (buffer A containing 20 mM imidazole, 500 mM NaCl), and the recombinant proteins were eluted with elution buffer (buffer A containing 200 mM imidazole, 500 mM NaCl). The eluted fractions were desalted and concentrated with buffer A by ultrafiltration with an Amicon Ultra-4 centrifugal filter device (10,000 molecular weight cut-off; Millipore). The amount of recombinant proteins was quantified using a Qubit™ Protein Assay Kit (Thermo Fisher Scientific) with bovine serum albumin as the standard.

Recombinant enzyme assay

The assay of SiUGTs was performed using C₆–C₂ or C₆–C₃ compounds as a sugar acceptor and UDP-glucose as a sugar donor. A standard reaction mixture (50 µl) consisted of 500 µM sugar acceptor, 500 µM UDP-glucose, 50 mM potassium phosphate buffer (pH 7.0, containing 0.05% 2-mercaptoethanol) and recombinant enzyme (final concentration, SiUGT1, 0.875 μg; SiUGT2, 3 μg; SiUGT3, 17.5 μg; SiUGT4, 15 μg and SiUGT5, 30 μg). The mixture without an enzyme was preincubated at 30°C for 10 min, and the reaction was started with the addition of an enzyme. After incubation at 30°C for an appropriate time (SiUGT1, 15 min; SiUGT2, 15 min; SiUGT3, 30 min; SiUGT4, 30 min and SiUGT5, 120 min), the reaction was stopped by the addition of 50 µl of acetonitrile.

The assay of SiATs was performed using hydroxytyrosol 1-O-glucoside, salidroside and decaffeoyl-acteoside as acyl acceptors and caffeoyl-CoA as an acyl donor. A standard reaction mixture (50 µl) consisted of 500 µM acyl acceptor, 500 µM caffeoyl-CoA, 100 mM Tris-HCl buffer (pH8.0, containing 10 mM EDTA) and recombinant enzyme (final concentration, SiUGT1–3: 30 μg). The mixture without an enzyme was preincubated at 30°C for 10 min, and the reaction was started with the addition of recombinant enzyme. After incubation at 30°C for 60 min, the reaction was stopped by the addition of 50 µl of acetonitrile.

Bioconversion of phenolic substrates by E. coli

For isolation of the products by SiUGT1 reaction, bioconversion was performed according to the method described by Ito et al. (2014) with some modifications. SiUGT1 (pET-53-DEST) was transformed into E. coli strain C41 (DE3) and overnight cultured in LB liquid medium with 50 μg/ml carbenicillin at 37°C. A 10 ml aliquot of the overnight culture was inoculated in 200 ml of fresh medium and cultured at 37°C until the OD₆₀₀ reached 0.4–0.6. IPTG was added to the culture at a final concentration of 1.0 mM and then incubated at 18°C for 24 h. The SiUGT1-expressing cells were collected by centrifugation at 3,000×g for 10 min, and the pellet was suspended in M9 minimal salts medium (M9; BD Difco, Franklin Lakes, USA) containing 2% glucose, adjusting the cell density to an OD₆₀₀ value of 3.0–6.0. The cell suspension culture was supplemented with hydroxytyrosol (dissolved in dimethyl sulfoxide at 200 mM) at a final concentration of 200 μM every 2 h and incubated at 30°C for 6 h with shaking at 160 rpm. The culture medium containing the product was collected by centrifugation, and crude purification was carried out by washing with water and eluting with acetonitrile using the COSMOSIL 75C₁₈-OPN (Nacalai Tesque, Inc., Kyoto, Japan) packed open columns (2.7 × 10 cm).

For the isolation of the SiAT1 reaction products, a multi-expression vector was constructed by a partial modification of the method of Uchida et al. (2020). Briefly, each expression cassette was prepared by PCR, assembled in a configuration to SiUGT1-Arabidopsis thaliana 4CL2 (At4CL2)-SiATs and introduced into pET-53-DEST by NEBuilder HiFi DNA assembly (New England Biolabs, Ipswich, USA). The codon-optimized At4CL2 CDS was prepared by GeneArt® Gene Synthesis (Thermo Fisher Scientific). This vector was transformed into E. coli strain C41 (DE3), and the bioconversion was performed in the same method as described earlier using hydroxytyrosol and caffeic acid as a substrate.

The products were purified by the following methods: HPLC analysis was performed on a Prominence™ system (Shimadzu, Kyoto, Japan) with a PDA detector and fraction collector (FRC-10A). A Waters Atlantis T3 OBD Prep column (10 × 250 mm, 5 μm) was used, and the column oven temperature was kept at 40°C. The mobile phase consisted of 0.1% (v/v) formic acid/water (A) and 0.1% (v/v) formic acid/acetonitrile (B) with gradient elution [1–30% (B) at 0–40 min, 30–100% (B) at 40–45 min, 100–100% (B) at 45–46 min, 100–1% (B) at 46–47 min and 1% (B) at 47–60 min]. The flow rate of the mobile phase was 3.0 ml/min, the sample injection volume was 60 μl and the eluent was monitored at 280 nm. For NMR analysis, the fraction containing the product was concentrated and crystallized.

Phylogenetic analysis

Multiple nucleotide and amino acid sequence alignment was performed using MUSCLE (Edgar 2004) algorithm in the MEGA11 software (Tamura et al. 2021). Phylogenetic trees were inferred using the neighbor-joining approach implemented within MEGA11 software, and the topology was estimated by bootstrapping over 1,000 replicates. The accession numbers of the amino acid sequences of the genes subjected to phylogenetic analysis are listed in the legend of each figure.

HPLC and LC-electrospray ionization–MS/MS analyses

Qualitative analysis of components in sesame-cultured cells and quantitative analysis of acteoside were performed by HPLC or LC-electrospray ionization (ESI)–MS/MS analyses as follows. HPLC analysis was performed on a Prominence™ system with a PDA detector. A Waters XBridge C18 column (4.6 × 150 mm, 5 μm) was used, and the column oven temperature was kept at 40°C. The mobile phase consisted of 0.1% (v/v) formic acid/water (A) and 0.1% (v/v) formic acid/acetonitrile (B) with gradient elution [5–35%(B) at 0–15 min, 35–100% (B) at 15–40 min, 100–5% (B) at 40–41 min and 5% (B) at 41–45 min]. The flow rate of the mobile phase was 0.8 ml/min, the sample injection volume was 10 μl and the eluent was monitored at 254 and 340 nm. For LC-ESI–MS/MS, a Waters Quattro Premier XL mass spectrometer coupled to an ACQUITY UPLC system with a Waters AQUITY PDA detector was used. A Waters HSS T3 column (2.1 × 50 mm, 1.7 μm) was used, and the column oven temperature was kept at 40°C. The mobile phase consisted of 0.1% (v/v) formic acid/water (A) and 0.1% (v/v) formic acid/acetonitrile (B) with gradient elution (5–35% (B) at 0–3 min, 35–100% (B) at 3–8 min, 100–5% (B) at 8–9 min and 5% (B) at 9–10 min]. The flow rate of the mobile phase was 0.3 ml/min, and the sample injection volume was 1 μl. A triple quadrupole mass spectrometer with an ESI source was used for mass determination. The ion source conditions were optimized as follows: source temperature, 120°C; capillary voltage, 3.5 kV; desolvation temperature, 400°C; flow rate of desolvation gas, 850 l/h; flow rate of cone gas, 50 l/h. The mass spectrometer was operated in both the positive and negative ion modes, with a scan range from m/z 200 to 1,200.

The analysis of the reaction products in the UGT and AT enzyme activity assays was performed as follows. LC-PDA–MS/MS analysis was performed on a Shimadzu LCMS-8050 system coupled to a Nexera X2 UPLC system with a PDA detector (Shimadzu). A Waters Atlantis T3 column (4.6 × 150 mm, 3.6 μm) was used, and the column oven temperature was kept at 40°C. The mobile phase consisted of 0.1% (v/v) formic acid/water (A) and 0.1% (v/v) formic acid/acetonitrile (B) with gradient elution [UGT reaction analysis method: 1–20% (B) at 0–6 min, 20–100% (B) at 6–12 min, 100–100% (B) at 12–12.5 min, 100–1% (B) at 12.5–13 min and 1% (B) at 13–15 min; AT analysis method: 1–30% (B) at 0–15 min, 30–100% (B) at 15–40 min, 100–100% (B) at 40–42 min, 100–1% (B) at 42–43 min and 1% (B) at 43–45 min]. The flow rate of the mobile phase was 1.0 ml/min (UGT analysis method) or 0.5 ml/min (AT reaction analysis method), and the sample injection volume was 2 μl. The MS spectrometer operated in the positive (+) and negative (−) ESI mode. Parameters for the interface were set as follows: nebulizing gas flow, 3 ml/min; heating gas flow, 10 l/min; drying gas flow, 10 l/min; interface temperature, 300°C; dilution line temperature, 250°C; heat block temperature, 400°C; collision-induced dissociation gas, 270 kPA and capillary voltage, 4 kV (+), −3.5 kV (−). The mass spectrometer operated in the scheduled MRM mode for MS/MS measurements. MRM transitions of each compound are described in each figure legend.

LC-PDA analysis to confirm the conversion was performed on the Nexera X2 UPLC system with a PDA detector (Shimadzu). Waters Atlantis T3 column (4.6 × 150 mm, 3.6 μm) was used, and the column oven temperature was kept at 40°C. The mobile phase consisted of 0.1% (v/v) formic acid/water (solvent A) and 0.1% (v/v) formic acid/acetonitrile (solvent B) with gradient elution (5–20% B at 0–6 min, 20–60% B at 6–8 min and 60–100% B at 8–10 min). The flow rate of the mobile phase was 0.5 ml/min, and the sample injection volume was 2 μl.

NMR analysis

¹H- and ¹³C-NMR spectra were obtained using a JEOL ECA-500 spectrometer (Tokyo, Japan) at 500 MHz in deuterium oxide (D₂O) and methanol-d₄. The ¹H- and ¹³C-NMR spectral assignments of the isolated compounds were achieved using COSY, HMQC and HMBC.

Supplementary Data

Supplementary data are available at PCP online.

Data Availability

Raw reads data of RNA-seq are available in the DDBJ Sequence Read Archive database with accession number (DRA016022). The nucleotide sequences of the SiUGTs and SiATs reported in this article have been submitted to the DDBJ database under the following accession numbers: SiUGT1 (LC763815), SiUGT2 (LC763816), SiUGT3 (LC763817), SiUGT4 (LC763818), SiUGT5 (LC763819), SiAT1 (LC763820), SiAT2 (LC763821) and SiAT3 (LC763822). The other data underlying this article will be shared on reasonable request to the corresponding author.

Funding

The Japan Society for the Promotion of Science KAKENHI Grant-in-Aid for Challenging Exploratory Research (JP18K19186) to H.M. and Grant-in-Aid for Early-Career Scientists (JP21K14788) to Y.F.

Acknowledgments

We are grateful to Ms. Junko Takanobu and Mr. Yutaka Yamada (RIKEN Center for Sustainable Resource Science) for technical support and to Prof. P. I. Mackenzie (Flinders University, Australia) for naming the Sesamum UGT numbers in this study. The authors would like to thank Editage for the English language editing.

Disclosures

The authors have no conflicts of interest to declare.

References