Molecular Cloning And Biochemical Characterization Of A New Coumarin Glycosyltransferase CtUGT1 From Cistanche Tubulosa

for more information:ali.ma@wecistanche.com

Xiping Xu, et al

ABSTRACT

UDP-glycosyltransferases (UGTs) are an important and functionally diverse family of enzymes involved in secondary metabolite biosynthesis. Coumarin is one of the most common skeletons of natural products with candidate pharmacological activities. However, to date, many reported GTs from plants mainly recognized flavonoids as sugar acceptors. Only limited GTs could catalyze the glycosylation of coumarins. In this study, a new UGT was cloned from Cistanche tubulosa, a valuable traditional tonic Chinese herb, which is abundant with diverse glycosides such as phenylethanoid glycosides, lignan glycosides, and iridoid glycosides. Sequence alignment and phylogenetic analysis showed that CtUGT1 is phylogenetically distant from most of the reported flavonoid UGTs and adjacent to phenylpropanoid UGTs. Extensive in vitro enzyme assays found that although CtUGT1 were not involved in the biosynthesis of bioactive glycosides in Cistanche tubulosa, it could catalyze the glucosylation of coumarins umbelliferone 1, esculetin 2, and hymecromone 3 in considerable yield. The glycosylated products were identified by comparison with the reference standards or NMR spectroscopy, and the results indicated that CtUGT1 can regiospecifically catalyze the glucosylation of hydroxyl coumarins at the C7- OH position. The key residues that determined CtUGT1’s activity were further discussed based on homology modeling and molecular docking analyses. Combined with site-directed mutagenesis results, it was found that H19 played an irreplaceable role as the crucial catalysis basis. CtUGT1 could be used in the enzymatic preparation of bioactive coumarin glycosides.

Keywords: Plant glycosyltransferases, Coumarin glycosides, Cistanche tubulosa, Homology modeling, Site-directed mutagenesis

Click to cistanche extract benefits products

Introduction

Glycosylation is one of the most common structure modification types of secondary natural products. Glycosylation could change the stability and/or solubility of the aglycon, and contribute a lot to the diversity of natural/unnatural products due to the multiple conjugation types. In plants, these reactions were usually conducted by uridine diphosphate (UDP)-dependent glycosyltransferases (UGTs), a kind of enzymes that catalyze the transfer of monosaccharide moieties from activated nucleotide sugar to a glycosyl acceptor molecule which can be a carbohydrate, glycoside, oligosaccharide, or a polysaccharide [1,2]

To date, many genes encoding UGTs have been isolated from several plants [3]. The glycosylation mainly happened at hydroxyl groups or carboxyl groups of a wide range of secondary products, such as flavonoids [4], phenylpropanoids [5], terpenoids [6], steroids [7], alkaloids [8], and so on. Coumarins are phenylpropanoids metabolites widely distributed in many plant species. Naturally occurring coumarins are often existed as glycoconjugates, and these derivatives exhibited a broad spectrum of pharmacological activities, including antioxidant [9], antiviral [10,11], hepatoprotective [12], anti-inflammatory [13], anti-cancer [14], antimutagenic [15], and cholinesterase (ChE) & monoamine oxidase (MAO) inhibitory activities [14,15], etc. However, to date, in comparison with the most frequently reported and extensively investigated flavonoids UGTs, only limited UGTs were reported able to transfer the sugar moiety to coumarins [16–18].

Plants rich in diverse glycosides could serve as an ideal gene pool of UGTs with novel catalysis activities. Furthermore, numerous UGTs have been reported to possess considerable acceptor promiscuity to catalyze the glycosylation of multiple substrates, some of which were even not separated in the plants [19]. In this article, a novel glycosyltransferase CtUGT1 was identified from Cistanche tubulosa, a traditional medicinal herb that is abundant with diverse glycosides such as phenylethanoid glycosides (PhGs), lignan glycosides, and iridoid glycosides. Heterologous expression and function characterization demonstrated that although recombinant CtUGT1 was not involved in the biosynthesis of these glycosides in Cistanche tubulosa, it can catalyze the glucosylation of hydroxyl coumarins, and the reaction regiospecifically happened on the 7- OH position of coumarins umbelliferone 1, escalating 2, and hymecromone 3 to produce three pharmacologically active compounds skimming (1a), cichoriin (2a), and 4-methylumbelliferyl glucoside (3a) with considerable yield, respectively. In addition, benzophenone substrate 4,4′ -dihydroxy benzophenone (4), isoflavone substrate genistein (5), and anthraquinone substrate aloe-emodin (6) could also be accepted by CtUGT1. The catalytic properties and key residues underlining the catalysis ability of CtUGT1 were also evaluated based on homology modeling, AutoDock analysis, and site-directed mutagenesis.

2. Experimental

2.1. Plant material and chemicals

Cistanche tubulosa used in this study was collected from desert areas in Hetian, Xinjiang autonomous regions. Their botanical identity was confirmed by Professor Pengfei Tu at Peking University. The tested coumarins and other substrates in this study were purchased from Sigma-Aldrich (St. Louis, USA), Chengdu Push Biotechnology Co., Ltd. (Chengdu, China), and Chengdu Biopurify Phytochemicals Co., Ltd. (Chengdu, China) unless otherwise stated. Reference standards of skimming (1a) and cichoriin (2a) were purchased from Wuhan ChemFaces Biochemical Co., Ltd. (Wuhan, China).

2.2. Molecular cloning of CtUGT1 from Cistanche tubulosa

The total RNA of Cistanche tubulosa was extracted from the fleshy stem using an OMEGA Plant RNA Kit (GA, USA) and reverse-transcribed to cDNA with PrimeScript™ RT reagent Kit (TaKaRa, Japan) following the manufacturer’s instructions. A degenerate primer was designed for 3′ RACE based on the aminoacid sequences in the conserved PSPG (Plant Secondary Product Glycosyltransferases) -box of plant glycosyltransferases. The 3′ -end and 5′ -end amplifications of CtUGT1 were carried out using Smart RACE cDNA Amplification Kit (Clontech, USA) according to the manufacturer’s protocol using primers (Table S1). The full-length cDNA of the CtUGT1 was amplified by PCR using KOD-PlusNeo DNA Polymerase (TOYOBO, Japan) with gene-specific primer pairs (Table S1) under the following conditions: an initial denaturation step at 94 ◦C for 2 min, followed by 35 cycles of denaturation at 98 ◦C for 15 s, annealing at 55 ◦C for 30 s, and extension at 68 ◦C for 55 s, with a final extension step at 68 ◦C for 7 min. All the DNA fragments obtained were cloned into a pEASY-Blunt vector (TransGen Biotech, China) and sequenced.

2.3. Sequence alignment and phylogenetic analysis

Sequence alignment of CtUGT1 with reported glycosyltransferases (Table S2) was performed using DNAMAN 4.0 software package (Lynnon Biosoft, Canada). The motifs and domains were analyzed using a conserved domain tool (http://www.ncbi.nlm.nih.Gov/Structure/cdd/ wrpsb. cgi). The phylogenetic tree was constructed by bootstrapping MEGA 7.0.14 software using the neighbor-joining method with 1000 bootstrap replicates [20]. The translated protein sequence of CtUGT1 was aligned with the known plant glycosyltransferases deposited in the NCBI GenBank database (Table S3) with ClustalW.

2.4. Heterologous expression and protein purification of CtUGT1 in Escherichia coli

The coding region of CtUGT1 was amplified with BamH I and Not I as restriction sites using primers shown in Table S1. PCR reactions were conducted using KOD-Plus-Neo DNA Polymerase (TOYOBO, Osaka, Japan). The obtained sequences were verified by cloning into the pEASY-Blunt vector (TransGen Biotech, China). The confirmed amplification product of CtUGT1 was subsequently digested and sub-cloned into the expression vector pET-28a(+) (Novagen, USA). The verified construct was then transformed into E. coli Transetta (DE3) to obtain the recombinant strain. Overnight culture of recombinant strain was inoculated into Luria-Bertani (LB) medium containing 50 μg⋅mL− 1 kanamycin and 40 μg⋅mL− 1 chloromycetin at a ratio of 1:100. The cultures were grown at 37 ◦C, 200 rpm until the OD600 value reached 0.4–0.6. Isopropyl-β-D-thiogalactopyranoside (IPTG) was subsequently added to a final concentration of 0.5 mM and the cells were grown for 16 h at 18 ◦C, 180 rpm. Then cell pellets were harvested by centrifugation (7600 ×g, 10 min, 4 ◦C), and re-suspended in 3 mL/g chilled lysis buffer (Supplementary Data Note 1), and disrupted by sonication on ice. The cell debris was removed by centrifugation at 7600 ×g, 4 ◦C for about 30 min. A pre-equilibrated Histrap column (GE Healthcare, Uppsala, Sweden) was used for affinity chromatography according to the manufacturer's instructions. The recombinant protein was eluted by 10 column volumes of elution buffer (Supplementary Data Note 1) containing 250 mM imidazole. Protein purity was analyzed by 10% SDS-PAGE. The purified protein was concentrated by a 30 kDa ultrafiltration tube (Sigma-Aldrich, USA) and desalted using a PD–10 column (GE Healthcare, Uppsala, Sweden) with desalting buffer (Supplementary Data Note 1). Protein concentration was determined by the Bradford method using BSA as a standard.

2.5. Enzymatic activity assays and substrate specificity of CtUGT1

To investigate the glycosylation activity of CtUGT1, enzyme assays were performed in a reaction mixture (150 μL) composed of 0.4 mM aglycone, 0.8 mM UDP-glucose (UDPG), and 50 μg of purified CtUGT1 protein in the desalting buffer. All reactions were incubated at 30 ◦C for 12 h and terminated by adding 300 μL of cold methanol. The mixtures were centrifuged at 15,000 ×g for 30 min to collect the supernatant for HPLC-UV/ESI-MS analyses. Three parallel assays were routinely carried out for each reaction. HPLC (Agilent 1260, USA) was equipped with a diode array detector and a CAPCELL PAK C18 column (250 mm × 4.6 mm, 5 μm; Shiseido, Japan) at a flow rate of 1 mL⋅min− 1, and the column temperature was maintained at 30 ◦C. The mobile phase consisted of A (i.e., 0.1% formic acid aqueous solution) and B (i.e., acetonitrile). The gradient programs were listed in Table S4. HRESI-MS data was recorded on an LCMS-IT-TOF system, fitted with an ESI interface (Shimadzu, Kyoto, Japan) with ultra-high purity He as the collision gas, and N2 as the nebulizing gas. The optimized ESI source parameters were as follows: sheath gas flow rate, 1.5 mL⋅min− 1; auxiliary gas flow rate, 1.5 mL⋅min− 1; spray voltage, 4.5 kV; capillary temperature, 200 ◦C. The spectra were recorded in the 100–1500 m/z range for a full scan MS analysis.

2.6. Effects of reaction time, pH value and temperature on enzyme activity, and kinetic studies of CtUGT1

Effects of reaction time, pH value, and temperature on the enzyme activity of CtUGT1 were tested using UDPG as the sugar donor and 3 as the sugar acceptor in the reaction conditions as described above. For the determination of optimal pH, the enzyme activity was compared in 100 mM buffer (citric acid‑sodium citrate) with pH values ranging from 3.0 to 6.0, 100 mM buffer (KH2PO4-K2HPO4) with pH values ranging from 6.0 to 8.0, 100 mM buffer (Tris-HCl) with pH values ranged from 7.0 to 9.0, 100 mM buffer (Na2CO3-NaHCO3) with pH values ranged from 9.0 to 11.0. To assay for the optimal reaction temperature, the reactions were incubated at various temperatures ranging from 0 to 65 ◦C. The time courses of the reaction were evaluated at 12 different time points between 0 and 24 h. All experiments were performed in triplicate. The reaction mixtures were analyzed by HPLC-MS as described above. For kinetic studies of CtUGT1, enzymatic assays were performed in a final volume of 100 μL containing 100 mM KH2PO4-K2HPO4 buffer (pH 8.0), 25 μg of purified CtUGT1, 3 mM of UDPG, and varying concentrations (50–3000 μM) of 3. The reactions were conducted at 37 ◦C for 30 min and then terminated with 100 μL ice-cold methanol. All of these reactions were performed in triplicate, and enzyme activity was evaluated via quantification of the corresponding glucosylated derivatives. Kinetic parameters, including the Michaelis–Menten constant (Km) and Vmax, were calculated by nonlinear regression analysis using GraphPad Prism 5 software.

2.7. Substrate and sugar donor selectivity of CtUGT1

To explore the substrate preferences of CtUGT1, a total of 27 substrates belonging to different structural types were tested using UDPG as the donor. Their chemical information was listed in Table S5, and their structures were shown in Figs. S1 and S2. To investigate the sugar donors selectivity of CtUGT1, assays were conducted using umbelliferone (1), esculetin (2), hymecromone (3), 4,4′ -dihydroxy benzophenone (4), genistein (5), and aloe-emodin (6) as substrates with UDPG, UDP-Nacetyl galactosamine, UDP-galactose, GDP-mannose, GDP-fucose, UDP galacturonic acid, UDP-rhamnose as sugar donors, respectively. All experiments were performed in triplicate. The reaction mixtures were analyzed by HPLC-MS as described above.

2.8. Preparation of 4-methylumbelliferyl glucoside (compound 3a) by CtUGT1

The preparative reaction mixture consisted of 0.034 mmol substrate compound 3, 0.04 mmol UDPG, 5 mg purified CtUGT1 in 50 mL reaction buffer. The reactions were gently agitated at the optimum conditions (100 mM KH2PO4-K2HPO4 with pH 8.0; 37 ◦C). After incubation at 37 ◦C for 24 h, the reaction supernatant was separated by column chromatography with macroporous resin (MCI GEL CHP) after centrifuging at 12,000 ×g for 30 min. The mobile phase was a gradient elution of 100% water to 100% methanol. Each fraction was analyzed by HPLC-UV. Fractions contained only targeted products were evaporated to dryness under reduced pressure and characterized by 1 H nuclear magnetic resonance (NMR) spectroscopy and 13C NMR spectroscopy.

2.9. Molecular docking and site-directed mutagenesis of CtUGT1

The protein model of CtUGT1 was established by using SWISS-MODEL [21–25]. The crystal structure of Medicago truncatula glycosyltransferase UGT85H2 (PDB ID 2PQ6) was chosen as the template structure [26]. The established model was assessed by VERIFY-3D [27,28]. Auto-Dock Vina was used for molecular docking studies [29]. The crystal structure of UGT74F2 [30], a glycosyltransferase from Arabidopsis thaliana, in complex with ligand UDP and salicylic acid (PDB ID 5U6M) was used as the control structure for sugar donor pocket and sugar acceptor pocket analysis, respectively. The Auto-Dock-Tools 1.5.6 from the MGLTools (Molecular Graphics Laboratory tools) were used to prepare the sugar donor UDPG and substrate 1, 2 for molecular docking. The side chains of residues around the active-site cavity were set flexible, and the rotatable bonds of ligands were left free to rotate. The top-ranked pose as judged by the Vina docking model with the highest (kcat/ mol) affinity was subjected to visual analysis using PyMOL 2.0 software.

For mutagenesis investigations, the sequence optimized gene of CtUGT1 (Supplementary data Note 2) was synthesized and tested as the wild type. Site-directed mutagenesis of the CtUGT1 was conducted using the Fast Mutagenesis System (TransGen Biotech). The mutants were amplified from the template of the optimized gene which has been constructed into the pET-28a vector using the specific primers shown in Table S1. The constructed mutants were verified by PCR and sequencing before being transformed into E. coli (DE3) for heterologous expression. Protein purification and enzyme reactions were performed and analyzed as described above.

3. Results and discussion

3.1. Full-length cDNA cloning of the CtUGT1 gene from Cistanche tubulosa

Cistanche tubulosa (L.) (RouCongRong) is a traditional Chinese medicinal herb producing a wide variety of bioactive natural glycosides. There should be a considerable number of new glycosyltransferases genes that existed in its genome. To seek novel glycosyltransferases with interesting catalysis activities, a degenerate primer for 3′ -RACE was designed based on the conserved PSPG motif to clone permissive UGTs from Cistanche tubulosa. The obtained 3′ end sequence was further analyzed by NCBI blast and the specific primer for 5′ -RACE was designed based on the obtained sequence. The amplified 5′ cDNA ends and 3′ cDNA ends were spliced and the full-length of the CtUGT1 sequence, including 1739 nucleotides, was obtained (Fig. S3) and listed in Supplementary data Note 3. Subsequently, an open reading frame (ORF) of CtUGT1 was found by NCBI ORF Finder tools and the 1482 bp coding sequence was successfully cloned by RT-PCR amplification using specific primers in Table S1. The cDNA sequence of CtUGT1 has been submitted to NCBI (accession No. MW629113).

3.2. Sequence analysis and heterologous expression of CtUGT1

The CtUGT1 gene was expected to encode a protein of 493 amino acid residues with an estimated molecular mass of 55.78 kDa. NCBI blast revealed that the deduced protein sequence of CtUGT1 shares the highest identity (76.39%) with UGT86A1-like enzymes. The biochemical functions of this kind of enzyme have not been fully characterized. It is predicted that they may be involved in the plant adaptation to abiotic stresses. The alignment of the amino acid sequences of CtUGT1 and other plant glycosyltransferases is shown in Fig. S4.

The highly conserved 44 residues of the PSPG motif were found at the C-terminal of CtUGT1. Within the PSPG-box, a peptide sequence of HCGWNS was located at the 373rd amino acid, which has been detected in 95% of all glycosyltransferases [31]. The final residue of the PSPG motif was glutamine, which indicated that this enzyme tends to use UDPG as its sugar donor [32] (Fig. S4). The phylogenetic tree was constructed based on the predicted polypeptide sequences of CtUGT1 and different kinds of plant origin glycosyltransferases which were involved in the biosynthesis of various secondary metabolites. So far, most of the reported glycosyltransferases from plants recognized flavonoids as their substrates, such as the flavone 7-O-glycosyltransferases, flavonoid 3-O-glucosyltransferases, isoflavone 7-O-glycosyltransferases, and chalcone glycosyltransferases. Phylogenetic analysis results showed that CtUGT1 has not grouped into the clade of flavonoid GTs. Instead, it is located adjacent to UGT74T1 and UGT74S1 (lignan glycosyltransferases from Linum usitatissimum) and NtGT2 (a coumarin glycosyltransferase from Nicotiana tabacum), which indicated that CtUGT1 may have some novel catalysis activities towards phenylpropanoid substrates (Fig. 1), especially considered that only limited members have been functionally identified belonging to these clades.

To further reveal its catalysis activity, the CtUGT1 sequence was expressed under the control of the T7 lac promoter in E. coli Transetta (DE3) cells as His6-tagged fusion protein. The His-tagged recombinant CtUGT1 was efficiently purified to homogeneity by Ni–NTA affinity chromatography. The molecular weight of the purified recombinant CtUGT1 was approximately 55 kDa (Fig. S5), which is consistent with the predicted molecular mass.

3.3. In vitro acceptor substrate specificity of CtUGT1

To test the glycosylation activity of CtUGT1, in vitro enzymatic assays were performed. UDPG was used as the sugar donor. Sugar acceptors were selected based on different concerns. Firstly, the CtUGT1 gene was cloned from Cistanche tubulosa whose main chemical components are PhGs. To verify whether CtUGT1 is involved in the biosynthesis of PhGs, the common aglycones of PhGs in Cistanche tubulosa such as tyrosol (NP1, Fig. S2) and salidroside (NP2, Fig. S2) were firstly selected as substrates. Unfortunately, HPLC and HRESI-MS analyses didn’t find any glycosylated products, which indicated that CtUGT1 may not be involved in the biosynthesis of PhGs.

Secondly, based on the sequence alignment and phylogenetic analysis results (Fig. 1), CtUGT1 was phylogenetically related to lignanoid and coumarin glycosyltransferases, which suggested that CtUGT1 is probably a phenylpropanoid glycosyltransferase. To verify this prediction, three coumarin substrates 1, 2, and 3 together with one lignanoid substrate magnolol NP17 were tested as sugar acceptors, respectively. Heat-inactivated enzymes (100 ◦C, 10 min) were used as a negative control. HPLC-HRESI-MS analysis of reactions using NP17 as substrate did not find any product. By contrast, CtUGT1 showed obvious glucosylation abilities against the tested coumarin substrates (1,2,3). Using substrate 1 as an example, as shown in Fig. 2A, a new peak (1a) with a decreased retention time (11.46 min) compared with that of substrate 1 (19.75 min) was produced by CtUGT1 at a yield of 37.78%. Its UV absorption spectrum is corresponding to 1. HRESI-MS spectrum showed that peak 1a exhibited a [M + H]+ at m/z 325.1029, and a [M + Na]+ at m/z 347.0710 (calcd. 347.0737 for C15H16O8Na) with the predicted formula of C15H16O8, which was 162 amu greater than that of 1 and consistent with the theoretical formula of a glucosylated product (Fig. 2A). Similarly, the glucosylated product of substrate 2 (14.21 min, Fig. 2B) and 3 (21.92 min, Fig. 2C) was also found at the retention time of 10.85 min and 13.80 min, with the conversion rate of 22.03% and 68.67%, respectively. The molecular mass of the products matched with the calculated mass of glucosylated products exactly, which confirmed that CtUGT1 could catalyze the glucosylation of coumarin compounds 1–3 (Table 1).

Subsequently, considering that most of the recently reported enzymes exhibited catalysis promiscuity in vitro, to further explore the substrate spectrum of CtUGT1, expansive enzyme assays were conducted towards diverse compounds belonging to different kinds of natural products, including flavonoids, phenolic compounds, coumarins, stilbene, anthraquinones, benzophenone, naphthalene, and iridoid (Figs. S1-S2), using UDPG as the sugar donor. It was found that the benzophenone substrate 4,4′ -dihydroxy benzophenone (4), isoflavone substrate genistein (5), and anthraquinone substrate aloe-emodin (6) could also be catalyzed by CtUGT1 to form the corresponding glucosylated product 4a, 5a, 6a with higher polarity (Fig. 3). HRESI-MS analysis showed that the ion peaks of the products were all 162 amu greater than that of the substrates, suggesting that CtUGT1 could catalyze the glucosylation of these compounds to form their corresponding monoglucoside product. The conversion rates of 4, 5 and 6 were 4.12%, 14.14% and 23.33%, respectively (Table 1).

The sugar donor selectivity of CtUGT1 was also explored. In addition to UDPG, six other donors, including UDP-N-acetyl galactosamine, UDPgalactose, GDP-mannose, GDP-fucose, UDP-galacturonic acid, and UDPrhamnose, were tested using substrates 1–6 as acceptors. The results showed that CtUGT1 exclusively selected UDPG as the sugar donor. No product was observed when other activated sugars were tested (data not shown).

3.4. Biochemical properties and kinetic parameters of CtUGT1

Biochemical properties of CtUGT1 were investigated using 3 as the sugar acceptor and UDPG as the sugar donor as indicated in Fig. 4. Catalysis activity CtUGT1 was tested in a temperature range of 4–60 ◦C, the optimal activity was detected at 37 ◦C (Fig. 4A). Analysis of the enzyme activity between pH 3.0 and 9.0 showed that the optimal pH value was observed at pH 8.0 (100 mM buffer KH2PO4-K2HPO4, Fig. 4C). The yield of the glucosylation product 3a increased linearly within 1 h and the growth rate was leveling off after 5 h (Fig. 4B). The apparent Km value of CtUGT1 for 3 was 218.7 ± 7.176 μM, and Vmax was 0.02255 ± 0.0001796 nmol⋅min− 1 ⋅μg− 1.

3.5. Enzymatic preparation and structural identification of glucosylated products

Based on the enzyme assays results, CtUGT1 tends to be a specific glucosyltransferase that could catalyze the glucosylation of different kinds of natural products. HPLC-HR-MS information of the products was summarized in Table 1. Among them, coumarins substrates exhibited a considerable conversion rate. The enzyme activity against coumarins was analyzed further. It should be noted that all three tested coumarin substrates accepted by CtUGT1 have a hydroxyl group at the C-7 position, while esculetin (2) has an extra hydroxyl group at the C-6 position. The HPLC profile of the glucosylated product of esculetin is shown in Fig. 2. Only a single product was observed, suggesting that the glucosylation reaction regiospecifically happened at the C-6 or C-7 position. To confirm the glycosylation site, product 2a was identified by comparison with the authentic glucosides by HPLC-UV/HRESI-MS analysis. When adding the authentic glucoside cichoriin (glucosylated on the 7-OH position of 2) in the reaction system, the peak overlapped with 2a, co-HPLC profile further confirmed that the glycosylated product of 2 catalyzed by CtUGT1 is cichoriin [33], which proved that the glucosyl moiety was specifically transferred on C7-OH position. Product 1a was also identified by comparison with the reference standard. UV-HPLC spectra together with co-HPLC analysis (Fig. 2) revealed that the chemical structure of 1a was attributed as skimming [34], which is the glucosylated product of 1 on C7-OH position.

Product 3a was prepared from a scaled-up reaction and structurally characterized by HRESI-MS, 1 H NMR, and 13C NMR spectroscopy (Figs. S13, S14, and Note 4). In the 1 H NMR signal, compared with substrate 3, a phenolic hydroxyl signal at the C-7 position disappeared, and the carbon signal of C-7 shifted δ 1.0 ppm to the high field, while C-6 and C-8 signals shifted to the low field in 13C NMR. These findings confirmed that the glucopyranosyl residue was attached to the phenolic hydroxyl group at the C-7 position of 3 [35]. Besides, 1 H NMR spectrum showed anomeric proton signals at δH 5.00 (1H, d, J = 7.0 Hz); the sugar component was indicated to be β-D-glucopyranose. This result demonstrated CtUGT1 is a coumarin O-glucosyltransferase that catalyzed the glucosylation of hydroxyl coumarin at the 7-OH position.

3.6. Homology modeling, molecular docking, and site-directed mutagenesis of CtUGT1 protein

Extensive enzyme reactions revealed that CtUGT1 could catalyze the glucosylation of three coumarins into their corresponding glucosylated product, and the glucosylated position specifically happened at the 7- OH position. To identify the key residues that determined the catalysis activities of CtUGT1, homology modeling, molecular docking, and site-directed mutagenesis investigations of CtUGT1 were performed. Homologous molecular modeling for CtUGT1 was established using SWISS-MODEL [21–25]. The crystal structure of Medicago truncatula glycosyltransferase UGT85H2 (PDB ID 2PQ6), determined at a resolution of 2.1 Å, showed the highest total score with maximum sequence homology and a low E-value, was chosen as the template structure [26]. Rationality assessment of the established model was shown in the Ramachandran plot (Fig. S6), 88.5% of the residues in the established model are in the most favored region. The homology model of CtUGT1 consists of two similar Rossmann-type folds motifs at the N- and C-terminal, respectively (Fig. 5A). The N-terminal domain has seven-stranded parallel β sheets around by ten α helices. The C-terminal domain contains six-stranded β sheets flanked by nine α helices (Fig. 5A, S7).

To further clarify the active pocket of CtUGT1, molecular docking was performed. Since UGT85H2 only has the apo form structure, another homolog UGT74F2 with both UDP and salicylic acid as two ligands were selected as the control template. The UDPG was prepared as the sugar donor. Two coumarin substrates 1 and 2 were prepared as the sugar acceptors respectively. The binding sites of glycosyl donor and acceptor were calculated and docked separately as shown in Fig. 5. The results indicated that both of the two binding pockets are located inside the deep cleft in the interaction area of the tightly packed N-terminus domain and C-terminus domain of CtUGT1 and spatially close to each other (Fig. 5A). The C-terminus of CtUGT1 is mainly involved in contact with the glycosyl donor as described in other plants' UGTs, and most of the surrounding residues are hydrophilic and highly conserved. Such as the typical PSPG-box, residue W377 in this motif could form hydrogen bonds with the glucose group of UDPG (Fig. 5B); N378 and S379 in this motif could form hydrogen bonds with the diphosphate group of UDPG (Fig. 5B). In addition, according to the docking results, residues of W356 and Q359 may interact with the uridine group through hydrogen bond interaction to stabilize the donor in the active pocket. To verify the function of these predicted key residues, a total of 16 sites were selected and subjected to alanine scanning mutagenesis. The catalysis activities of these mutants were tested using all six positive substrates mentioned above. HPLC analyses showed that, when residues of G18, H19, S295, F296, W356, C357, Q359, H374, G376, W377, N378, S379, E382, Y397, and D398 were mutated into alanine, the glycosylation activity of the protein was totally lost towards all tested substrates. These results suggested that the above 16 residues are the key residues that determined the binding of the UDPG donor and the glycosylation activity of CtUGT1.

To identify the crucial residues in the sugar acceptor binding pocket, molecular docking analyses were performed on 1 and 2, respectively. These two substrates have the same coumarin core. 2 have two hydroxyl groups at both C-6 and C-7 position while 1 only has one hydroxyl group at the C-7 position. The overlapped docking results of these two compounds were shown in Fig. 5C. They exhibited almost the same confirmation in the glycosyl acceptor binding pocket (1 in cyan color and 2 in green color). Two hydrogen bonds were observed between the NE2 atom of the imidazole ring of H19 and the OH group of coumarins at positions C-6 (~3.1 Å) and C-7(~3.0 Å) respectively (Fig. 5D). It was calculated that H19 is much closer to the 7-OH group which will make the hydroxyl group at the C-7 position easily to be deprotonated by H19 to form the nucleophilic oxyanion that could attack the C1′ carbon of UDPG to initiate the glycosylation reaction. This may explain why the glycosylation position was preferred to happen at the 7-OH position of coumarins substrates. Besides, it can be seen that H19 is located at the cleft of the two active pockets, and interacts with both the substrate and UDPG (Fig. 5B, C, D). Mutation of His19 to alanine will lead to the complete loss of enzyme activity, which confirmed the irreplaceable role of H19 that plays a role as the crucial catalysis basis, meanwhile, stabilizing the UDPG donor. Alanine scanning mutagenesis of some other sites in the acceptor binding pocket also resulted in no detectable enzyme activity, including Y16, L213, F296, or significantly decrease of the enzyme activity (> 44%) such as V14, V132, E153, T212, and I209 (Fig. S8). It is predicted that these hydrophobic residues surrounding the aromatic ring could form van der Waals interactions with the substrate in the binding pocket.

4. Discussion

Glycosylation is one of the most important modification steps in many secondary metabolites’ biosynthesis pathways. Glycosylation could increase the bioavailability of natural products by improving their water solubility and reducing their toxicity. Coumarin is considered the simplest form within a huge class of naturally-occurring phenolic sub-stances constructed from an α-pyrone ring fused with a benzene ring. Both natural and synthetic coumarin-based compounds have attracted much attention from medicinal chemists and drug design scientists as a consequence of their diverse pharmacological and biological properties [36]. Further glycosylated modification of coumarins could help to enrich the structural diversity of these compounds and also improve their solubility and bioavailability. In plants, such a glycosylation process is catalyzed by UDP-glycosyltransferase. Although multiple UGTs have been identified from different plants, to date, most of the reported UGTs were involved in the glycosylated modification of flavonoids. Only limited UGTs were reported to be able to accept coumarins as substrates. In this study, a new glycosyltransferase gene CtUGT1 was cloned from the traditional Chinese herb Cistanche tubulosa. Sequence alignment and phylogenetic analysis results showed that CtUGT1 is phylogenetically distant from most of the reported flavonoid UGTs, and much close to phenylpropanoid UGTs. Extensive enzyme assays revealed that CtUGT1 could not catalyze the glycosylation of phenyl ethanol compounds to form PhGs, which is the main chemical component of Cistanche tubulosa. Instead, it could recognize coumarins as its substrates, and conduct the glycosylation reaction. Structure elucidation of the products showed that the glycosylation site catalyzed by CtUGT1 regiospecifically happened at the C7-OH position of coumarins. Using CtUGT1, three coumarin glycosides including skimming (1a), cichoriin (2a), and 4-methylumbelliferyl glucoside (3a) were successfully enzymatic synthesized and structurally identified. All these three compounds were reported as pharmaceutically active molecules. For example, 1a has been considered to possess various pharmacological activities, including renoprotective activity, anti-inflammatory, anti-cancer, and anti amoebic properties [37]. 2a showed high antibacterial activity against some Gram-positive bacteria like Bacillus cereus and Staphylococcus aureus [38]. 3a can help the plants to improve their response abilities to chemicals and toxins [39].

In addition to coumarins, extensive enzyme assays found that CtUGT1 also showed observable glucosylation activity towards benzophenone substrate 4, isoflavone substrate 5, and anthraquinone substrate 6, which indicated that CtUGT1 has certain substrate tolerance. Nevertheless, CtUGT1 could not catalyze the biosynthesis of glycosides components in Cistanche tubulosa such as phenylethanoid glycosides (PhGs), lignanoid glycosides, and iridoid glycosides. Besides, according to the previous investigations, no coumarin glycosides were isolated from Cistanche tubulosa. So, the in vivo function of CtUGT1 still needs to be further explored. The glucosylation activity of CtUGT1 towards coumarin substrates makes this enzyme a useful tool for the enzymatic glycosylation modification of coumarins and also indicated that that plant secondary metabolic enzymes could perform inestimable catalysis activities that are far beyond their native functions.

Declaration of Competing Interest

The authors declare that there are no conflicts of interest.

Acknowledgment

This work was financially supported by Beijing Natural Science Foundation (Grant No. 7192112); Young Elite Scientists Sponsorship Program by CAST (Grant No. CACM− 2018− QNRC1− 02); National Natural Science Foundation of China (Grant No. 81402809); China Scholarship Council State Scholarship Fund (Grant No. 201906555010) and Science Foundation for Elite Young Scholars of Beijing University of Chinese Medicine (Grant No. 2018-JYB-XJQ006).

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.fitote.2021.104995.

From: ' Molecular cloning and biochemical characterization of a new coumarin glycosyltransferase CtUGT1 from Cistanche tubulosa' by Xiping Xu, et al

---Fitoterapia 153 (2021) 104995 https://doi.org/10.1016/j.fitote.2021.104995

References

[1] Y. Li, S. Baldauf, E.K. Lim, D.J. Bowles, Phylogenetic analysis of the UDPglycosyltransferase multigene family of Arabidopsis thaliana, J. Biol. Chem. 276 (2001) 4338–4343, https://doi.org/10.1074/jbc.M007447200.
[2] L.F. Mackenzie, Q. Wang, R.A.J. Warren, S.G. Withers, Glycosynthases: mutant glycosidases for oligosaccharide synthesis, J. Am. Chem. Soc. 120 (1998) 5583–5584, https://doi.org/10.1021/ja980833d.
[3] C.J. Thibodeaux, C.E. Melancon, H.W. Liu, Unusual sugar biosynthesis and natural product glycodiversification, Nature 446 (2007) 1008–1016, https://doi.org/ 10.1038/nature05814.
[4] Y. Ji, B. Li, M. Qiao, J. Li, H. Xu, L. Zhang, X. Zhang, Advances on the in vivo and in vitro glycosylations of flavonoids, Appl. Microbiol. Biotechnol. 104 (2020) 6587–6600, https://doi.org/10.1007/s00253-020-10667-z.
[5] T. Koeduka, Y. Ueyama, S. Kitajima, T. Ohnishi, K. Matsui, Molecular cloning and characterization of UDP-glucose: volatile benzenoid/phenylpropanoid glucosyltransferase in petunia flowers, J. Plant Physiol. 252 (2020) 153245, https://doi.org/10.1016/j.jplph.2020.153245.
[6] S. Rahimi, J. Kim, I. Mijakovic, K.H. Jung, G. Choi, S.C. Kim, Y.J. Kim, Triterpenoid-biosynthetic UDP-glycosyltransferases from plants, Biotechnol. Adv. 37 (2019) 107394, https://doi.org/10.1016/j.biotechadv.2019.04.016.
[7] G. Singh, Y.V. Dhar, M.H. Asif, P. Misra, Exploring the functional significance of sterol glycosyltransferase enzymes, Prog. Lipid Res. 69 (2018) 1–10, https://doi. org/10.1016/j.plipres.2017.11.001.
[8] K.F. McCue, P.V. Allen, L.V. Shepherd, A. Blake, J. Whitworth, M.M. Maccree, D. R. Rockhold, D. Stewart, H.V. Davies, W.R. Belknap, The primary in vivo steroidal alkaloid glucosyltransferase from potato, Phytochemistry. 67 (2006) 1590–1597, https://doi.org/10.1016/j.phytochem.
[9] J. Yu, L. Wang, R.L. Walzem, E.G. Miller, L.M. Pike, B.S. Patil, Antioxidant activity of citrus limonoids, flavonoids, and coumarins, J. Agric. Food Chem. 53 (2005) 2009–2014, https://doi.org/10.1021/jf0484632.
[10] H.J. Cho, S.G. Jeong, J.E. Park, J.A. Han, H.R. Kang, D. Lee, M.J. Song, Antiviral activity of angelicin against gammaherpesviruses, Antivir. Res. 100 (2013) 75–83, https://doi.org/10.1016/j.antiviral.2013.07.009.
[11] Z. Xu, Q. Chen, Y. Zhang, C. Liang, Coumarin-based derivatives with potent anti-HIV activity, Fitoterapia. 150 (2021) 104863, https://doi.org/10.1016/j. fitote.2021.104863.
[12] A.T. Mossa, T.M. Heikal, M. Belaiba, E.G. Raoelison, H. Ferhout, J. Bouajila, Antioxidant activity and hepatoprotective potential of Cedrelopsis grevei on cypermethrin induced oxidative stress and liver damage in male mice, BMC Complement. Altern. Med. 15 (2015) 251, https://doi.org/10.1186/s12906-015-0740-2.
[13] Y. Bansal, P. Sethi, G. Bansal, Coumarin: a potential nucleus for anti-inflammatory molecules, Med. Chem. Res. 22 (2013) 3049–3060, https://doi.org/10.1007/s00044-012-0321-6.
[14] M. Kumar, R. Singla, J. Dandriyal, V. Jaitak, Coumarin derivatives as anticancer agents for lung cancer therapy: a review, Anti Cancer Agents Med. Chem. 18 (2018) 964–984, https://doi.org/10.2174/1871520618666171229185926.
[15] A. Stefanachi, F. Leonetti, L. Pisani, M. Catto, A. Carotti, Coumarin: a natural, privileged and versatile scaffold for bioactive compounds, Molecules 23 (2018) 250, https://doi.org/10.3390/molecules23020250.
[16] T. Taguchi, N. Ubukata, H. Hayashida, M. Yamamoto, Okazaki, cloning and characterization of a glucosyltransferase that reacts on 7-hydroxyl group of flavonol and 3-hydroxyl group of coumarin from tobacco cells, Arch. Biochem. Biophys. 420 (2003) 95–102, https://doi.org/10.1016/j.abb.2003.09.027.
[17] L. Zhou, T. Tian, B. Xue, L. Song, L. Liu, R. Yu, Biosynthesis of coumarin glycosides by transgenic hairy roots of Polygonum multiflorum, Biosci. Biotechnol. Biochem. 76 (2012) 1008–1010, https://doi.org/10.1271/bbb.110347.
[18] H. Kanoh, M. Kawauchi, M. Kuroyanagi, T. Arima, Molecular cloning and characterization of coumarin glucosyltransferase in hairy roots of Pharbitis nil (Ipomoea nil), Plant Tissue Cult. Lett. 31 (2014) 21–28, https://doi.org/10.5511/plantbiotechnology.13.1203a.
[19] J.B. He, P. Zhao, Z.M. Hu, S. Liu, Y. Kuang, M. Zhang, B. Li, C.H. Yun, X. Qiao, M. Ye, Molecular and structural characterization of a promiscuous Cglycosyltransferase from Trollius Chinensis, Angew. Chem. Int. Ed. Eng. 58 (2019) 11513–11520, https://doi.org/10.1002/anie.201905505.
[20] S. Kumar, G. Stecher, K. Tamura, MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets, Mol. Biol. Evol. 33 (2016) 1870–1874, https://doi. org/10.1093/molbev/msw054.
[21] A. Waterhouse, M. Bertoni, S. Bienert, G. Studer, G. Tauriello, R. Gumienny, F. T. Heer, T.A.P. de Beer, C. Rempfer, L. Bordoli, R. Lepore, T. Schwede, SWISS-MODEL: homology modeling of protein structures and complexes, Nucleic Acids Res. 46 (2018) W296–W303, https://doi.org/10.1093/nar/gky427.
[22] N. Guex, M.C. Peitsch, T. Schwede, Automated comparative protein structure modeling with SWISS-MODEL and Swiss-Pdb viewer: a historical perspective, Electrophoresis 30 (2009) S162–S173, https://doi.org/10.1002/elps.200900140.
[23] S. Bienert, A. Waterhouse, T.A.P. de Beer, G. Tauriello, G. Studer, L. Bordoli, T. Schwede, The SWISS-MODEL repository-new features and functionality, Nucleic Acids Res. 45 (2017) D313–D319, https://doi.org/10.1093/nar/gkw1132.
[24] G. Studer, C. Rempfer, A.M. Waterhouse, G. Gumienny, J. Haas, T. Schwede, QMEANDisCo-distance constraints applied on model quality estimation, Bioinformatics. 36 (2020) 1765–1771, https://doi.org/10.1093/bioinformatics/ btz828.
[25] M. Bertoni, F. Kiefer, M. Biasini, L. Bordoli, T. Schwede, Modeling protein quaternary structure of homo- and hetero-oligomers beyond binary interactions by homology, Sci. Rep. 7 (2017) 10480, https://doi.org/10.1038/s41598-017-09654-8.
[26] L. Li, L.V. Modolo, L.L. Escamilla-Trevino, L. Achnine, R.A. Dixon, X. Wang, the Crystal structure of Medicago truncatula UGT85H2-insights into the structural basis of a multifunctional (iso)flavonoid glycosyltransferase, J. Mol. Biol. 370 (2007) 951–963, https://doi.org/10.1016/j.jmb.2007.05.036.
[27] R. Lüthy, J.U. Bowie, D. Eisenberg, Assessment of protein models with three-dimensional profiles, Nature 356 (1992) 83–85, https://doi.org/10.1038/356083a0.
[28] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, PROCHECK: a program to check the stereochemical quality of protein structures, J. Appl. Crystallogr. 26 (1993) 283–291, https://doi.org/10.1107/S0907444998006684.
[29] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455–461, https://doi.org/10.1002/jcc.21334.
[30] A.M. George Thompson, C.V. Iancu, K.E. Neet, J.V. Dean, J.Y. Choe, Differences in salicylic acid glucose conjugations by UGT74F1 and UGT74F2 from Arabidopsis thaliana, Sci. Rep. 7 (2017) 46629, https://doi.org/10.1038/srep46629.
[31] T. Vogt, P. Jones, Glycosyltransferases in plant natural product synthesis: characterization of a supergene family, Trends Plant Sci. 5 (2000) 380–386, https://doi.org/10.1016/s1360-1385(00)01720-9.
[32] A. Kubo, Y. Arai, S. Nagashima, T. Yoshikawa, Alteration of sugar donor specificities of plant glycosyltransferases by a single point mutation, Arch. Biochem. Biophys. 429 (2004) 198–203, https://doi.org/10.1016/j. abb.2004.06.021.
[33] H. Kanho, S. Yaoya, T. Itani, T. Nakane, N. Kawahara, Y. Takase, K. Masuda, M. Kuroyanagi, Glucosylation of phenolic compounds by Pharbitis nil hairy roots: I. Glucosylation of coumarin and flavone derivatives, Biosci. Biotechnol. Biochem. 68 (2004) 2032–2039, https://doi.org/10.1271/bbb.68.2032.
[34] J. Bjerre, E.H. Nielsen, M. Bols, Hydrolysis of toxic natural glucosides catalyzed by cyclodextrin dicyanohydrins, Eur. J. Org. Chem. 2008 (2010) 745–752, HTTPS:// doi.org/10.1002/ejoc.200700954.
[35] B. Xue, L.B. Zhou, J.W. Liu, R.M. Yu, Biotransformation of hydroxycoumarin derivatives by cultured suspension cells of Catharanthus roseus, Pharmazie 67 (2012) 467–471, https://doi.org/10.1691/ph.2012.1741.
[36] T. Al-Warhi, A. Sabt, E.B. Elkaeed, W.M. Eldehna, Recent advancements of coumarin-based anticancer agents: an up-to-date review, Bioorg. Chem. 103 (2020) 104163, https://doi.org/10.1016/j.bioorg.2020.104163.
[37] X. Li, M.Y. Gruber, D.D. Hegedus, D.J. Lydiate, M.J. Gao, Effects of a coumarin derivative, 4-methylumbelliferone, on seed germination and seedling establishment in Arabidopsis, J. Chem. Ecol. 37 (2011) 880–890, https://doi.org/10.1007/s10886-011-9987-3.
[38] K. Xu, Z.M. Feng, Y.N. Yang, J.S. Jiang, P.C. Zhang, Eight new eudesmane- and eremophilane-type sesquiterpenoids from Atractylodes lancea, Fitoterapia. 114 (2016) 115–121, https://doi.org/10.1016/j.fitote.2016.08.017.
[39] Y. Lou, H. Wu, J. Zheng, X. He, Z. Wu, X. Lu, Y. Qiu, Determination and pharmacokinetic study of skimming by UHPLC-MS/MS in rat plasma, J. Pharm. Biomed. Anal. 179 (2020) 112969, https://doi.org/10.1016/j.jpba.2019.112969.