Dear Editor,
Among the large number of Veratrum steroid alkaloids (Zhang et al., 2020), cyclopamine has been shown to have various inhibitory mechanisms against cancer, including inhibition of the hedgehog signaling pathway, improvement of tumor tissue microenvironment, inhibition of cell respiration, promotion of cell apoptosis, and reversal of tumor drug resistance (Liu et al., 2023). The biosynthetic pathway of verazine, a possible metabolic precursor of cyclopamine, has been discovered in Veratrum californicum (Augustin et al., 2015). However, the content of Veratrum steroid alkaloids in plants is very low (Supplemental Figure 2), so reliance on direct extraction from plants for commercial production is not feasible. A more effective strategy is to synthesize rare plant secondary metabolites using synthetic biology technology (Liu et al., 2022).
About 1 mg verazine was extracted from dried Veratrum nigrum rhizome for qualitative and quantitative analyses of verazine, and the nuclear magnetic resonance results are shown in Supplemental Tables 1 and 2. To synthesize the precursor molecule cholesterol required for verazine metabolism, the genes involved in the cholesterol biosynthetic pathway of Paris polyphylla were expressed in Nicotiana benthamiana (Yin et al., 2023). Using a high-yielding cholesterol synthesis chassis in N. benthamiana, four enzymes involved in the synthesis of verazine in V. californicum were transiently co-expressed. VcCYP90B27 (KJ869252) was confirmed to encode a cholesterol C-22 hydroxylase. However, the enzyme activity of VcCYP90B27 in N. benthamiana was not as high as expected, and the substrate utilization rate was less than 25% (Supplemental Figures 1A and 1B). Some key amino acid differences (Figure 1A) may have contributed to the low activity of VcCYP90B27 in N. benthamiana. As a result, the downstream enzymes VcCYP94N1 (KJ869255), VcGABAT (KJ869263), and VcCYP90G1 (KJ869260) may have had insufficient substrates for verazine synthesis.
Figure 1.
Heterologous and efficient synthesis strategy of verazine from N. benthamiana.
(A) Amino acid variations in CYP90B27 enzymes from different species.
(B) Tissue-specific expression patterns of the VnCYP90 family, the VnCYP94 family, and VnGABAT.
(C) Strategies for N-terminal signal peptide truncation of VcGABAT.
(D) Strategies for the biosynthesis of verazine.
(E‒H) Mass spectra of verazine and its intermediates.
(I‒M) Expression levels of VnCYP90B27, VnCYP94N2, VnCYP90G9, and VnGABAT and verazine contents in different organs (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).
(N–Q) Yield comparison of genes with the same function from different species in Nicotiana benthamiana by gas chromatography–mass spectrometry (GC–MS) and liquid chromatography–tandem MS (LC–MS/MS) analysis. C, cholesterol heterologous synthesis chassis (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).
(R) Effects of different N-terminal truncations of GABAT on verazine yield.
(S) Veratrum nigrum L (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).
(T) Subcellular localization of VcGABAT in Nicotiana benthamiana leaf cells observed by confocal laser scanning microscopy (Zeiss LSCM 800, Oberkochen, Germany). The sequence of EGFP is linked to the C terminus of VcGABAT. Bar = 20 μm.
Because of the low activity of VcCYP90B27 in N. benthamiana, it was necessary to find an alternative enzyme with high catalytic activity. We noted that V. nigrum, a traditional medicinal herb from China, also contains a wide variety of steroid alkaloids (Zhang et al., 2020). We sequenced the full-length transcriptome of V. nigrum and obtained four candidate genes, VnCYP90B27 (PP129547), VnCYP94N2 (PP129548), VnGABAT (PP129549), and VnCYP90G9 (PP129550), on the basis of their predicted amino acid sequence similarity to orthologs in V. californicum. These candidate genes had similar tissue expression specificities (Figure 1B) and the same accumulation patterns as verazine in V. nigrum (Figures 1I‒1M). Agrobacteria harboring these four genes were infiltrated into N. benthamiana leaves to test their functions. The substitution of VnCYP90B27 for VcCYP90B27 resulted in an approximately 20-fold increase in 22(R)-hydroxycholesterol content (Figure 1N). With the addition of VnCYP90B27, VcCYP94N1 was able to oxygenate the C-22 position of 22(R)-hydroxycholesterol, and the product 22,26-dihydroxycholesterol was observed by gas chromatography–mass spectrometry. Liquid chromatography–tandem mass spectrometry revealed that VcGABAT replaces the C-22 aldehyde group of 22,26-dihydroxycholesterol with an amino group, yielding 22-hydroxy-26-aminocholesterol. VcCYP90G1 then oxidizes the C-22 hydroxyl group of 22-hydroxy-26-aminocholesterol into a carbonyl group, which undergoes a ring formation reaction with the C-26 amino group to synthesize verazine (Figures 1E–1H).
The yields of eight enzymes involved in the biosynthesis of verazine from V. nigrum and V. californicum were compared in different combinations (Figures 1N–1Q). Because PpCYP90B27 has been confirmed to be a cholesterol C-22 hydroxylase, like VnCYP90B27 and VcCYP90B27 (Yin et al., 2023), there were three candidate genes for CYP90B27 derived from V. californicum, V. nigrum, and P. polyphylla. VnCYP90B27 and PpCYP90B27 showed the same level of activity, and the production of VnCYP90B27 was slightly higher (Figure 1N). The second enzyme for comparison was CYP94N1, and the candidate enzymes were VnCYP94N2 and VcCYP94N1. The yield of VnCYP94N2 was slightly higher, about 1.4 times that of VcCYP94N1 (Figure 1O). The catalytic product yields of VcGABAT and VnGABAT were compared, and that of VnGABAT was significantly lower, approximately one-eighth that of VcGABAT (Figure 1P). Moreover, after addition of VnCYP90G9 on this basis, the amount of verazine synthesized was only 0.03 μg/g DW (Figure 1Q). On the basis of the aforementioned combinations, replacement of VnGABAT with VcGABAT significantly increased the downstream production of verazine. Overall, the combination of VnCYP90B27, VnCYP94N2, VcGABAT, and VcCYP90G1 produced the highest verazine yield of 0.53 μg/g dry weight (DW) (Figure 1Q).
The three CYP450s involved in the metabolism of cholesterol to verazine are located on the endoplasmic reticulum (Supplemental Figure 1C). However, another key rate-limiting enzyme, GABAT, is localized in the chloroplast (Figure 1T). We speculated that differences in subcellular localization may affect the efficiency of GABAT in utilization of 22-dihydroxycholesterol-26-ol. Because GABAT is localized in chloroplasts and on the cytoplasmic membrane, we predicted that truncating its N-terminal signal peptide would increase its opportunity to bind with the substrate, thereby improving product yield. The strategy for truncating GABAT is shown in Figure 1C. Use of a truncated GABAT missing 15 amino acids, t15-VcGABAT, increased the verazine yield to 5.11 μg/g DW, about 10 times higher than that obtained with untruncated VcGABAT (Figure 1R). The metabolic pathway of verazine involves a metabolic branch: CYP94N1 catalyzes the formation of two isomers from 22(R)-hydroxycholesterol produced by CYP90B27, namely 22(R),26[25(R)]-dihydroxycholesterol and 22(R),26[25(S)]-dihydroxycholesterol. Only 22(R),26[25(S)]-dihydroxycholesterol can be used to synthesize verazine. At the same time, CYP90G1 oxidizes the C-22 hydroxyl group of 22(R)-hydroxy-26[25(S)]-aminosterol, forming 22-ketohydroxycholesterol-26[25(R)]. This pathway not only leads to the loss of pathway intermediates but also wastes the catalytic activity of the enzyme. One potential strategy for further increasing the verazine yield is therefore to optimize the product specificity of VnCYP94N2 and the substrate selectivity of VcCYP90G1. Introduction of glutamate decarboxylase into the synthetic chassis to increase levels of γ-aminobutyric acid, the amino donor for GABAT, could be an alternative approach. In addition, the catalytic efficiency of enzymes is influenced by differences in amino acids in the active center. Further experiments have shown that saturation mutagenesis at key amino acid sites in the active center can produce highly efficient enzymes. Enzyme engineering through directed mutagenesis is based on this principle (Chen et al., 2019; Jamil et al., 2022). Different species of Veratrum may have differences at multiple amino acid sites in these key enzymes (CYP90B27 and GABAT) that may lead to variations in their catalytic efficiency.
In summary, we completed the heterologous biosynthesis of verazine in N. benthamiana. We compared the activities of enzymes from different sources and identified the optimal enzyme combination. These genes make about 100 times as much verazine in N. benthamiana than in Camelina sativa seed (Augustin et al., 2017). Heterologous biosynthesis of verazine provides new insight into the production mode of steroid alkaloids, as well as substrate support for the resolution of the verazine-to-cyclopamine metabolic pathway.
Funding
This work was supported by the National Key Research and Development Program of China (2023YFA0915800) and the Jilin Provincial Agricultural Science & Technology Innovation Project (CXGC202105GH).
Author contributions
P.M. and Z.X. conceived and designed the research and revised the manuscript; C.K. performed the research and wrote the paper; R.M. and W.S. offered constructive suggestions; and X.H., X.Y., Jia Liu, D.H., and Jingling Liu helped perform the research.
Acknowledgments
The authors acknowledge professor Qinlong Zhu, College of Life Science, South China Agricultural University, for providing the fluorescent fusion protein vector toolbox. The authors acknowledge the technical support of Dr. Shengnan Tan from the Analysis and Test Center, Northeast Forestry University. No conflict of interest is declared.
Published: February 1, 2024
Footnotes
Published by the Plant Communications Shanghai Editorial Office in association with Cell Press, an imprint of Elsevier Inc., on behalf of CSPB and CEMPS, CAS.
Supplemental information is available at Plant Communications Online.
Contributor Information
Zheyong Xue, Email: xuezhy@126.com.
Pengda Ma, Email: mapengda@163.com.
Supplemental information
References
- Augustin M.M., Ruzicka D.R., Shukla A.K., Augustin J.M., Starks C.M., O'Neil-Johnson M., McKain M.R., Evans B.S., Barrett M.D., Smithson A., et al. Elucidating steroid alkaloid biosynthesis in Veratrum californicum: production of verazine in Sf9 cells. Plant J. 2015;82:991–1003. doi: 10.1111/tpj.12871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Augustin M.M., Shukla A.K., Starks C.M., O’Neil-Johnson M., Han L., Holland C.K., Kutchan T.M. Biosynthesis of Veratrum californicum specialty chemicals in Camelina sativa seed. Plant Biotechnology Reports. 2017;11:29–41. [Google Scholar]
- Chen X., Shukal S., Zhang C. Integrating Enzyme and Metabolic Engineering Tools for Enhanced α-Ionone Production. J. Agric. Food Chem. 2019;67:13451–13459. doi: 10.1021/acs.jafc.9b00860. [DOI] [PubMed] [Google Scholar]
- Jamil O.K., Cravens A., Payne J.T., Kim C.Y., Smolke C.D. Biosynthesis of tetrahydropapaverine and semisynthesis of papaverine in yeast. Proc. Natl. Acad. Sci. USA. 2022 doi: 10.1073/pnas.2205848119. 119:e2205848119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu B., Kou C., Ren Y., Li Q., Wang T., Ma R., Sun W., Xue Z., Ma P. Advances of Veratrum nigrum L. Steroid Alkaloids. Industrial Crops and Products. 2023;191 [Google Scholar]
- Liu X., Zhang P., Zhao Q., Huang A.C. Making small molecules in plants: A chassis for synthetic biology-based production of plant natural products. J. Integr. Plant Biol. 2022;65:417–443. doi: 10.1111/jipb.13330. [DOI] [PubMed] [Google Scholar]
- Yin X., Liu J., Kou C., Lu J., Zhang H., Song W., Li Y., Xue Z., Hua X. Deciphering the network of cholesterol biosynthesis in Paris polyphylla laid a base for efficient diosgenin production in plant chassis. Metab. Eng. 2023;76:232–246. doi: 10.1016/j.ymben.2023.02.009. [DOI] [PubMed] [Google Scholar]
- Zhang M., Gao L., Xu S., Huang W., Li X., Ye Y. Advances in studies on steroidal alkaloids and their pharmacological activities in genus Veratrum. China J. Chin. Mater. Med. 2020;45:5129–5142. doi: 10.19540/j.cnki.cjcmm.20200612.201. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.