English

• Paris polyphylla var. yunnanensis is a popular medicinal plant with anti-tumor, anti-bacterial, anti-inflammatory, anti-tussive, and analgesic activities [1-6]. Its major bioactive compounds include steroidal saponins, also known as polyphyllins [7]. These paris saponins exhibit anticancer, antimicrobial, anti-inflammatory, anthelmintic and hemostatic activities [8,9]. Among them, the major compounds paris saponins Ⅱ and Ⅶ contain a four-residue sugar chain at C-3 of diosgenin and pennogenin, respectively (Fig. S1 in Supporting information).

  Biosynthesis of natural products through synthetic biology is a novel, environment-friendly and sustainable approach [10,11]. Although some enzymes participated in the biosynthesis of polyphyllins have been reported, the biosynthetic pathways of paris saponin Ⅱ and Ⅶ have not been fully elucidated. Hua and Yin et al. characterized ten enzymes from Paris polyphylla var. yunnanensis, which are responsible for the synthesis of cholesterol (Fig. S2 in Supporting information) [12,13]. Christ et al. identified a series of CYPs including CYP90G4 and CYP94D108 from Paris polyphylla var. yunnanensis which catalyzes spiroketalization of cholesterol to produce diosgenin [14]. Glycosylation plays a critical role in the biosynthesis of steroidal saponins. Several 3-O-glucosyltransferases including UGT80A40 have been identified [13,15,16]. Recently, a rhamnosyltransferase (UGT73CE1) was identified to catalyze 2′-O-rhamnosylation at the 3-O-glucose group to produce diglycosides [16]. However, the 4′-O-rhamnosylation and 4′′-O-rhamnosylation which are the last two steps for the biosynthesis of paris saponin Ⅱ and Ⅶ have not been elucidated (Fig. 1A).

  Figure 1

  

  Figure 1.  Bioinformatic analyses of candidate rhamnosyltransferase genes from Paris polyphylla var. yunnanensis. (A) Putative biosynthetic pathway of paris saponin Ⅱ (9) and paris saponin Ⅶ (10). (B) Expression levels of candidate genes in the transcriptome of different parts of the plant. (C) Cluster analysis of five candidate genes with complete reading frames.

  DownLoad: Full-Size Img PowerPoint

  l-Rhamnose is a popular sugar unit of plant glycosides [17,18]. Thus far, only about 30 rhamnosyltransferases have been identified from the plant kingdom, and most of them are flavonoid and terpenoid glycosyltransferases [19,20]. Up to now, only two steroidal rhamnosyltransferases have been reported, including DzGT1 from Dioscorea zingiberensis and UGT73CE1 from P. polyphylla var. yunnanensis [16,21]. In this work, we identified two rhamnosyltransferases PpRhaGT1 and PpRhaGT2 from P. polyphylla var. yunnanensis, which catalyze a cascade of 4′-O- and 4′′-O-rhamnosylation to produce paris saponin Ⅱ and Ⅶ, respectively. We also elucidated mechanisms for the substrate selectivity of PpRhaGT1 through molecular dynamics.

  To discover the functional UDP-glycosyltransferase (UGT) genes, CYP90G4, CYP90D108, UGT80A40 and UGT73CE1, which participated in the biosynthesis of polyphyllins, were used as baits for gene co-expression analysis (Table S1 in Supporting information) [22]. As shown in Fig. 1B, 14 candidate glycosyltransferase genes PpGT1–14 exhibited highly correlated expression with bait genes in different tissues of Paris polyphylla var. yunnanensis [14,16]. Among them, 5 candidate genes had complete open reading frames. In addition, PpGT3, PpGT5, PpGT8 and PpGT11 were clustered with terpenoid and steroid glycosyltransferases, while PpGT6 was clustered with flavonoid glycosyltransferases (Fig. 1C, Table S2 in Supporting information). Thus, PpGT3, PpGT5, PpGT8 and PpGT11 were selected for further studies.

  The candidate genes (Tables S3 and S4 in Supporting information) were cloned into the pET-28a(+) vectors using primers in Table S5 (Supporting information) and the proteins were expressed in BL21(DE3) strains of E. coli. The enzymes were purified by Ni-NTA affinity chromatography and ion-exchange chromatography, and then analyzed by SDS-PAGE (Figs. S3 and S4 in Supporting information). The functions were characterized using paris saponin Ⅴ (5), paris saponin Ⅵ (6), dioscin (7), pennogenin-3-O-chacotrioside (8) as substrates and UDP-rhamnose (UDP-Rha) as sugar donor. The reaction mixtures were analyzed by liquid chromatography coupled with mass spectrometry (LC/MS).

  LC/MS analysis indicated that PpGT11 (PpRhaGT1) could catalyze the 4′-O-rhamnosylation of 5 and 6 to generate 7 and 8, respectively (Fig. 2, Figs. S5–S9 in Supporting information). PpRhaGT1 exhibited maximum activity at 30 ℃, pH 8.0 (50 mmol/L Tris–HCl) and was independent of divalent metal ions (Fig. S10 in Supporting information). The K_m values for the 4′-rhamnosylation of 5 and 6 were 9.46 and 11.25 µmol/L, respectively. The k_cat for the 4′-rhamnosylation of 5 and 6 were 0.003 s⁻¹ (Fig. S11 in Supporting information).

  Figure 2

  

  Figure 2.  Functional characterization of PpRhaGT1. (A) Glycosylation of paris saponin Ⅴ (5) and paris saponin Ⅵ (6) catalyzed by recombinant PpRhaGT1 using UDP-Rha as sugar donor. (B) Total ion chromatograms (TICs) of the enzymatic reaction mixture and the reference standards. (C) LC/MS analysis of the glycosylated products 7 and 8.

  DownLoad: Full-Size Img PowerPoint

  PpGT5 (PpRhaGT2) exhibited potent activities of 4′′-O-rhamnosylation. It catalyzed the synthesis of paris saponin Ⅱ (PS Ⅱ, 9) and Ⅶ (PS Ⅶ, 10) using dioscin (7) and pennogenin-3-O-chacotrioside (8) as substrates, respectively (Fig. 3, Figs. S12–S16 in Supporting information). PpRhaGT2 showed maximum activity at 37 ℃, pH 8.0 (50 mmol/L NaH₂PO₄Na₂HPO₄) or pH 9.0 (50 mmol/L citric acid-sodium citrate), and was independent of divalent metal ions (Fig. S17 in Supporting information). The K_m for the 4′′-rhamnosylation of 7 and 8 were 24.78 and 5.24 µmol/L and the k_cat for the 4′′-rhamnosylation of 7 and 8 were 0.00021 and 0.00007 s⁻¹, respectively (Fig. S18 in Supporting information).

  Figure 3

  

  Figure 3.  Functional characterization of PpRhaGT2. (A) Glycosylation of dioscin (7) and pennogenin 3-O-chacotrioside (8) catalyzed by recombinant PpRhaGT2 using UDP-Rha as sugar donor. (B) TICs of the enzymatic reaction mixture and the reference standards. (C) LC/MS analysis of the glycosylated products 9 and 10.

  DownLoad: Full-Size Img PowerPoint

  Then, nine sugar donors, including UDP-rhamnose (UDP-Rha), UDP-xylose (UDP-Xyl), UDP-glucose (UDP-Glc), UDP-galactose (UDP-Gal), UDP-glucosamine (UDP-GlcN), UDP galactosamine (UDP-GalN), UDP-arabinose (UDP-Ara), UDP-glucuronic acid (UDP-GlcA) and UDP-galacturonic acid (UDP-GalA), were used to probe the sugar donor specificity of PpRhaGT1 and PpRhaGT2 using 6 and 8 as substrates, respectively. The results indicated that PpRhaGT1 and PpRhaGT2 possessed strict sugar donor selectivity towards UDP-Rha (Figs. S19 and S20 in Supporting information). Only a weak product was detected in the reaction of PpRhaGT2 using UDP-xylose as sugar donor. In addition, PpRhaGT1 showed high substrate specificity. Six steroidal and triterpene di-glycosides were tested, and PpRhaGT1 could utilize none of them as substrate (Fig. S21 in Supporting information).

  The above results indicated that PpRhaGT1 played a critical role in the biosynthesis of PS Ⅱ and PS Ⅶ. There are two possible glycosylation orders in the biosynthesis of tri-glycosides (Fig. 4A). Both 3 and 5 were used as substrates, and the results revealed that PpRhaGT1 preferred to utilize 5, and only generated a very minor product with 3 (Fig. 4B, Fig. S22 in Supporting information). Thus, the biosynthesis of PS Ⅱ (9) and PS Ⅶ (10) may follow the order of 3-O-glucosylation, 2′-O-rhamnosylation, 4′-O-rhamnosylation, and 4′′-O-rhamnosylation (Pathway 2).

  Figure 4

  

  Figure 4.  Substrate selectivity of PpRhaGT1. (A) Two putative biosynthetic pathways of dioscin (7) and pennogenin 3-O-chacotrioside (8). (B) TICs of the enzymatic reaction mixture and the reference standards. (C) Distance between the oxygen atom and anomeric carbon in 50 ns molecular dynamics simulation. (D) Structural analysis of PS Ⅴ (5) with surrounding amino acids. (E) The frequency of hydrogen bonds formed by 2′-O-rhamnosyl group and surrounding amino acids in MD.

  DownLoad: Full-Size Img PowerPoint

  To further explore the mechanisms of substrate selectivity of PpRhaGT1, the protein model was constructed through Alphafold 2 [23,24]. UDP-Rha was docked into the sugar donor binding pocket based on the crystal structure of UGT89C1/UDP-Rha [25,26]. In addition, PS Ⅴ (5) and trillin (3) were docked into the substrate binding pocket using autodock. Molecular dynamics (MD) of PpRhaGT1/trillin and PpRhaGT1/PS Ⅴ were conducted. MD indicated the distance (around 5 Å) between the 4′-O of PS Ⅴ (5) and C-1 of UDP-Rha remained stable in the simulation time. However, the distance between the 4′-O of trillin (3) and C-1 of UDP-Rha increased from 5 Å to 8 Å (Fig. 4C). These results were consistent with the experiments, and indicated the significance of 2′-O-rhamnosyl group for the catalytic reaction. In 5, 3′′-O and 4′′-O of 2′-O-rhamnosyl group formed three hydrogen bonds with E383 and S382, and these hydrogen bonds remain stable in MD (Fig. 4, Fig. 4). Then we constructed the S382A and E383A mutants, and both of them showed weak activities towards 5 and 6 (Fig. S23 in Supporting information). These results revealed the substrate preference of PpRhaGT1 was due to the interactions between 5 and surrounding amino acids particularly S382 and E383.

  Subsequently, we constructed the semisynthetic pathways of PS Ⅱ (9) and PS Ⅶ (10) in Nicotiana benthamiana through Agrobacterium-mediated transient expression. Pgm, GalU and EpRhS were used to produce UDP-Glc and UDP-Rha [27]. Glycosyltransferase genes were divided into different groups and infiltrated into the leaves of tobacco. After 4 days of infiltration, diosgenin (1) and pennogenin (2) were fed as substrates (Fig. 5A). Three days later, the leaves were extracted and analyzed through LC/MS. The results were consistent with the in vitro reactions. In groups 1 and 3, two tri-glycosides dioscin (7) and pennogenin-3-O-chacotrioside (8) were synthesized. More importantly, paris saponin Ⅱ (9) and paris saponin Ⅶ (10) were synthesized in group 2 and group 4 through expression of all four glycosyltransferases (Figs. 5B–D).

  Figure 5

  

  Figure 5.  Semi-biosynthesis of dioscin (7), pennogenin 3-O-chacotrioside (8), paris saponin Ⅱ (9) and paris saponin Ⅶ (10) in tobacco. (A) The experimental procedure. (B) The engineered biosynthetic pathway of 7, 8, 9 and 10 in Nicotiana benthamiana. (C) Semi-biosynthesis of 7, 8, 9 and 10. (D) LC/MS analysis of the glycosylated products. The structures of 7, 8, 9 and 10 were identified by comparing with reference standards.

  DownLoad: Full-Size Img PowerPoint

  In summary, two rhamnosyltransferases PpRhaGT1 and PpRhaGT2 involved in the biosynthesis of paris saponin Ⅱ and Ⅶ were identified from the medicinal plant Paris polyphylla var. yunnanensis. PpRhaGT1 catalyzes the 4′-O-rhamnosylation of paris saponin Ⅴ and Ⅵ to produce tri-glycosides. PpRhaGT2 further catalyzes the 4′′-O-rhamnosylation to synthesize paris saponin Ⅱ and Ⅶ. Both PpRhaGT1 and PpRhaGT2 exhibited strict sugar-donor selectivity for UDP-Rha. Molecular dynamics analysis indicated that interactions between 2′-O-rhamnosyl group and surrounding amino acids particularly S382 and E383 in PpRhaGT1 improved the binding stability of substrates. Furthermore, paris saponin Ⅱ and Ⅶ were synthesized in Nicotiana benthamiana through transient expression. This work elucidates the glycosylated pathways of paris saponin Ⅱ and Ⅶ, and provides biocatalysts for the synthesis of steroidal saponins.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Jia-Jing Zhou: Writing – review & editing, Writing – original draft, Validation, Software, Resources, Project administration, Methodology, Investigation, Data curation. Zi-Long Wang: Project administration, Methodology. Meng Zhang: Writing – review & editing, Visualization, Validation, Methodology. Yang-Oujie Bao: Methodology. Guo-Wei Chang: Methodology. Yun-Gang Tian: Methodology. Min Ye: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Data curation.

  Acknowledgments

  This work was supported by Yunnan Provincial Science and Technology Project at Southwest United Graduate School (No. 202302AP370006), and the National Key Research and Development Program of China (No. 2023YFA0914100).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110805.

  
  
• 1. [1]

     A.E. Tao, F.Y. Zhao, R.S. Li, J.F. Qian, C.L. Xia, Chin. Herb. Med. 51 (2020) 4809–4815.

  2. [2]

     Y.G. Ding, Y.L. Zhao, J. Zhang, et al., J. Ethnopharmacol. 278 (2021) 114293. doi: 10.1016/j.jep.2021.114293

  3. [3]

     Y. Wang, Y.J. Zhang, W.Y. Gao, L.L. Yan, China J. Chin. Mater. Med. 32 (2005) 1425–1428.

  4. [4]

     L. Guo, J. Su, B.W. Deng, et al., Hum. Reprod. 23 (2008) 964–971. doi: 10.1093/humrep/den001

  5. [5]

     X.J. Qin, D.J. Sun, W. Ni, et al., Steroids 77 (2012) 1242–1248. doi: 10.1016/j.steroids.2012.07.007

  6. [6]

     Y. Wang, W.Y. Gao, T.J. Zhang, Y.Q. Gao, Chin. Chem. Lett. 18 (2007) 548–550. doi: 10.1016/j.cclet.2007.03.011

  7. [7]

     J.S. Negi, V.K. Bisht, A.K. Bhandari, et al., Anticancer Agents Med. Chem. 14 (2014) 833–839. doi: 10.2174/1871520614666140611101040

  8. [8]

     F. Liu, L.N. Li, X.C. Tian, et al., J. Chem. 2021 (2021) 1442906.

  9. [9]

     J.Y. Zhang, Y.Z. Wang, Y.L. Zhao, et al., J. Asian Nat. Prod. Res. 13 (2011) 670–681. doi: 10.1080/10286020.2011.578247

  10. [10]

      D.S. Li, M.J. Yang, R.F. Mu, et al., Chin. Chem. Lett. 34 (2023) 107469. doi: 10.1016/j.cclet.2022.04.067

  11. [11]

      D.L. Chen, J.W. Li, X.D. Xu, et al., Chin. Chem. Lett. 35 (2024) 109451. doi: 10.1016/j.cclet.2023.109451

  12. [12]

      X. Yin, J. Liu, C.X. Kou, et al., Metab. Eng. 76 (2023) 232–246. doi: 10.1016/j.ymben.2023.02.009

  13. [13]

      X. Hua, W. Song, K.Z. Wang, et al., Commun. Biol. 5 (2022) 50. doi: 10.1038/s42003-022-03000-z

  14. [14]

      B. Christ, C. Xu, M. Xu, et al., Nat. Commun. 10 (2019) 3206. doi: 10.1038/s41467-019-11286-7

  15. [15]

      W. Song, C.C. Zhang, J.L. Wu, et al., ACS Synth. Biol. 11 (2022) 1669–1680. doi: 10.1021/acssynbio.2c00103

  16. [16]

      Y. Chen, Q. Yan, Y. Ji, et al., Acta Pharm. Sin. B 13 (2023) 4638–4654. doi: 10.1016/j.apsb.2023.05.033

  17. [17]

      Y.J. Xu, T.H. Xu, J.Y. Yang, et al., Chin. Chem. Lett. 21 (2010) 580–583. doi: 10.1016/j.cclet.2010.01.021

  18. [18]

      L. Tang, X.J. Wu, B.M. Feng, et al., Chin. Chem. Lett. 22 (2011) 835–838. doi: 10.1016/j.cclet.2011.01.014

  19. [19]

      S. Irmisch, S. Jo, C.R. Roach, et al., Plant Cell 30 (2018) 1864–1886. doi: 10.1105/tpc.18.00406

  20. [20]

      X. Jian, Q.Y. Sun, W.T. Xu, et al., Angew. Chem. Int. Ed. 63 (2024) e202409867. doi: 10.1002/anie.202409867

  21. [21]

      J. Li, I. Mosongo, H. Li, et al., Front. Plant Sci. 12 (2021) 713036. doi: 10.3389/fpls.2021.713036

  22. [22]

      H.T. Wang, Z.L. Wang, K. Chen, et al., Nat. Commun. 14 (2023) 6658. doi: 10.1038/s41467-023-42393-1

  23. [23]

      J. Jumper, R. Evans, A. Pritzel, et al., Nature 596 (2021) 583–589. doi: 10.1038/s41586-021-03819-2

  24. [24]

      P. Cramer, Nat. Struct. Mol. Biol. 28 (2021) 704–705. doi: 10.1038/s41594-021-00650-1

  25. [25]

      O. Trott, A.J. Olson, J. Comput. Chem. 31 (2010) 455–461. doi: 10.1002/jcc.21334

  26. [26]

      G.N. Zong, S. Fei, X. Liu, et al., Plant J. 99 (2019) 257–269. doi: 10.1111/tpj.14321

  27. [27]

      K.P. Feng, R.D. Chen, K.B. Xie, et al., Org. Biomol. Chem. 16 (2018) 452–458. doi: 10.1039/C7OB02763J

• 

  Figure 1  Bioinformatic analyses of candidate rhamnosyltransferase genes from Paris polyphylla var. yunnanensis. (A) Putative biosynthetic pathway of paris saponin Ⅱ (9) and paris saponin Ⅶ (10). (B) Expression levels of candidate genes in the transcriptome of different parts of the plant. (C) Cluster analysis of five candidate genes with complete reading frames.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Functional characterization of PpRhaGT1. (A) Glycosylation of paris saponin Ⅴ (5) and paris saponin Ⅵ (6) catalyzed by recombinant PpRhaGT1 using UDP-Rha as sugar donor. (B) Total ion chromatograms (TICs) of the enzymatic reaction mixture and the reference standards. (C) LC/MS analysis of the glycosylated products 7 and 8.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Functional characterization of PpRhaGT2. (A) Glycosylation of dioscin (7) and pennogenin 3-O-chacotrioside (8) catalyzed by recombinant PpRhaGT2 using UDP-Rha as sugar donor. (B) TICs of the enzymatic reaction mixture and the reference standards. (C) LC/MS analysis of the glycosylated products 9 and 10.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Substrate selectivity of PpRhaGT1. (A) Two putative biosynthetic pathways of dioscin (7) and pennogenin 3-O-chacotrioside (8). (B) TICs of the enzymatic reaction mixture and the reference standards. (C) Distance between the oxygen atom and anomeric carbon in 50 ns molecular dynamics simulation. (D) Structural analysis of PS Ⅴ (5) with surrounding amino acids. (E) The frequency of hydrogen bonds formed by 2′-O-rhamnosyl group and surrounding amino acids in MD.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Semi-biosynthesis of dioscin (7), pennogenin 3-O-chacotrioside (8), paris saponin Ⅱ (9) and paris saponin Ⅶ (10) in tobacco. (A) The experimental procedure. (B) The engineered biosynthetic pathway of 7, 8, 9 and 10 in Nicotiana benthamiana. (C) Semi-biosynthesis of 7, 8, 9 and 10. (D) LC/MS analysis of the glycosylated products. The structures of 7, 8, 9 and 10 were identified by comparing with reference standards.

  下载: 全尺寸图片 幻灯片
•