Skip to main content
NCBI home page
Plant Communications logoLink to Plant Communications
. 2025 Jun 24;6(9):101430. doi: 10.1016/j.xplc.2025.101430

The complete synthetic pathway of echinacoside from Cistanche deserticola and its de novo biosynthesis in yeast

Yali Ban 1, Jixuan Jiang 1, Hongwang Yang 1, Haiyang Jia 2, Yaru Pang 2, Xu Cheng 2, Jianbin Yan 3, Qinggang Liao 3, Chun Li 2,4,∗, Bo Lv 2,∗∗, Yongjun Feng 1,∗∗∗
PMCID: PMC12447445  PMID: 40566684

Abstract

Echinacoside (ECH), a representative phenylethanol glycoside, exhibits diverse pharmacological properties and is used in the treatment of neurodegenerative disorders (e.g., Parkinson’s and Alzheimer’s diseases), ischemic brain injury, and cancer. The growing therapeutic demand for ECH has highlighted the need for scalable production. However, conventional methods face major limitations: chemical synthesis is hindered by the compound’s structural complexity, and the yield of ECH extracted from plants is naturally low due to the host-dependent growth of Cistanche deserticola (C. deserticola), a parasitic desert plant. To establish a sustainable microbial production platform, we first deciphered the biosynthetic pathway of ECH in C. deserticola by integrating metabolomics analyses of plant tissues and callus cultures. This enabled the identification of key precursors, enzymatic steps, and regulatory mechanisms. Leveraging this knowledge, we engineered the pathway in Saccharomyces cerevisiae, achieving de novo ECH biosynthesis at a titer of 7.52 ± 1.42 mg/l. This study lays the foundation for industrial-scale ECH production and deepens our understanding of bioactive compound biosynthesis in parasitic plants, offering insights for future pathway engineering efforts.

Key words: Echinacoside, Cistanche deserticola, differential gene analysis, metabolomics, microbial biomanufacturing

Echinacoside (ECH), a prototypical phenylethanol glycoside, exhibits a wide range of pharmacological activities. However, its chemical synthesis is hindered by complex structure, and plant extraction is constrained by the low natural yield from Cistanche, a desert-dwelling parasitic plant. To overcome these limitations, this study successfully elucidates the biosynthetic pathway of ECH in Cistanche deserticola and achieves its de novo synthesis in Saccharomyces cerevisiae.

Introduction

Echinacoside (ECH), a compound derived from the valuable traditional medicinal plants of the Cistanche genus in the Orobanchaceae family, is well known as a typical phenylethanol glycoside compound (PhG) with broad medicinal applications, including antioxidant, neuroprotective, cognition-enhancing, liver-protective, anti-inflammatory, and anti-tumor properties (Chen et al., 2019; Zhang and Hao, 2020; Chuang et al., 2022; Wang et al., 2023b). ECH promotes the expression of osteogenesis-related proteins and the osteogenic genes RUNX2 and OCN, and improves levels of the angiogenesis-related proteins MMP-2 and VEGF to inhibit osteoclast formation, reducing the risk of fractures in the elderly (Yi et al., 2024). Clinical evidence shows that ECH ameliorates brain inflammation in Parkinson’s disease by regulating microglia-mediated nucleotide-binding oligomerization domain-like receptor protein 3 (NLRP3), caspase-1, and interleukin-1β inflammatory signaling pathways (Gao et al., 2020). Additionally, ECH is widely used in healthcare products, food, and cosmetics (Wu et al., 2023), leading to a significant increase in global demand and drivinggrowth for the ECH industry.

However, Cistanche species are obligate parasitic plants and depend on the root exudates of their hosts, Haloxylon ammodendron and Tamarix chinensis, which are distributed in arid and semi-arid areas and rely heavily on limited natural resources or semi-artificial cultivation. Unfortunately, ECH is a structurally complex molecule with a hydroxytyrosol scaffold, a branched caffeic acid at the C-4′ glucose motif position, a rhamnose modification at the C-3′ position, and a glucose modification at the C-6′ position. The complex chemical structure of ECH contains 1,2-trans β- and α-glycosidic linkages, which are usually constructed under the neighboring group participation effect of the C-2 acyl-protecting function (Mulani et al., 2014). Environmental challenges and the presence of stereoisomeric mixtures hinder scalable and sustainable production (He et al., 2018). Recently, the advancement of synthetic biology has significantly improved the capacity to construct cellular factories, enabling the de novo synthesis of natural products like the adjuvant QS saponins (Martin et al., 2024), flavonoid epimedoside in Saccharomyces cerevisiae (An et al., 2023), and the paclitaxel precursor baccatin III in microbial cell factories (Jiang et al., 2024).

Despite the considerable interest in the pharmaceutical potential of ECH, the ECH biosynthetic pathway remains unknown. Fortunately, the key intermediate acteoside has a similar biosynthetic pathway in autotroph plants, such as Sesamum indicum, Olea europaea, and Rehmannia glutinosa (Yang et al., 2021, 2023), and according to the bioretrograde evolution hypothesis, the biosynthesis of ECH can be divided into four critical processes: (1) the supply of the precursors p-coumaric acid, caffeic acid, and salidroside; (2) the synthesis of acteoside by BAHD-acyltransferases (HCTs), followed by rhamnosylation at the C-3′ position by rhamnosyltransferase; (3) hydroxylation of the intermediate by the cytochrome P450 (CYP) enzyme; and (4) the final step of glucose glycosylation on acteoside to form ECH. Meanwhile, comparative metabolomics and transcriptomics have been used to explore the phenylpropane and phenylalanine biosynthetic pathways; however, the expression levels of candidate genes involved in HCTs and glycosylation of ECH (Hou et al., 2022), cistanoside A, and isoacteoside were inconsistent with the trends seen in the PhGs pathway, which shows a significant negative correlation with the content of ECH. Completely non-photosynthetic parasitic Cistanche species lack typical roots and stomata, exhibit extensive host interaction, and evolved nutritional dependence, horizontal gene transfer, and signal exchange (Zheng et al., 2014; Fan et al., 2023). In addition, further elucidation of the genes involved in the host-dependent biosynthesis of ECH, including the photo-induced adaptation of lignin production, is needed.

Here, we report the complete biosynthetic pathway of acteoside and ECH and establish their microbial biomanufacturing in S. cerevisiae, providing new insights into the replacement of plant-based supply of PhGs with industrial fermentation at scale. To accomplish this, we chose Cistanche deserticola, which has the highest content of ECH, as the focus of this study. First, we quantified the content of relevant PhGs in C. deserticola and sequenced the transcriptomes at different growth phases to assess correlations between expression levels of candidate genes involved in ECH production. Subsequently, due to the complexity of cultivating C. deserticola, we verified its metabolic profile by combining it with callus culture under different precursor feeding conditions and identified statistically significant candidate genes by weighted gene co-expression network analysis (WGCNA). Furthermore, the C. deserticola-derived candidate genes were expressed and characterized by biological methods. Finally, we artificially reconstituted a biosynthetic pathway for the de novo production of acteoside and ECH in engineered yeast.

Results

Combination of plant tissue and callus to analyze the biosynthetic pattern of ECH in C. deserticola

The synthetic ECH content in C. deserticola varies significantly across different growth stages (Qian et al., 2016). To facilitate comprehensive metabolomic and transcriptomic studies, we investigated 6 growth stages based on characteristic developmental phases, including succulent stem growth, 10 cm below ground, topsoil stage, 10 cm above ground, pregnant bud flowering, and seed-setting (Figure 1). The 10 cm below ground and topsoil stages exhibited the highest levels of both ECH and acteoside. During the succulent stem growth period, ECH and acteoside levels exhibited an increasing trend, likely reflecting favorable growth conditions of the host plants. As the plants transitioned from vegetative growth to reproductive phases, the concentrations of both ECH and acteoside decreased, with ECH reaching negligible levels during the pregnant bud flowering and seed-setting stages. During lignification, C. deserticola produces lignin glycosides in large quantities, in addition to compounds such as PhGs, iridoid and secoiridoid components, and sugars (Liu et al., 2018a; 2018b). Given the low salidroside content in C. deserticola, it was quantified by peak area rather than absolute concentration. Contrasting trends were observed for salidroside, which was more abundant during the 10 cm above ground, pregnant bud flowering, and seed-setting stages compared to the 10 cm below ground and topsoil stages (Figure 1). As for cistanoside A, a further downstream compound of ECH, its content in the six growth stages followed a pattern similar to that of ECH (Figure 1). Considering the chemical structures and concentrations of salidroside and ECH in C. deserticola, we hypothesize that salidroside may be an upstream precursor of ECH.

Figure 1.

Figure 1

Open in a new tab

Sampling strategy across growth stages of C. deserticola and quantification of four target compounds.

The first row illustrates six stages of artificial growth, from left to right: succulent stem growth period, 10 cm below ground, topsoil stage, 10 cm above ground, pregnant bud flowering period, and seed-setting stage. Each stage was further subdivided into 3–5 anatomical segments, represented by five boxes beneath each stage. All samples underwent metabolomic and transcriptomic analyses. Cp1–Cp6 represent the six growth stages, while CpnX1–CpnX5 represent the proximity of the tissue relative to the sucker; for example, Cp6X1 refers to the basal region of the stem and Cp6X5 refers to the inflorescence region at the seed-setting stage. The figure displays the levels of ECH, acteoside, salidroside (represented by peak area due to its low abundance), and castanoside across 66 samples, illustrating dynamic trends in compound concentrations throughout the course of growth.

To test this hypothesis and minimize the influence of the host on C. deserticola, we employed plant tissue culture techniques to develop calli from C. deserticola explants (Figure 2A). Analysis of ECH and acteoside contents based on dry weight revealed that callus-derived production was approximately 10-fold higher than that from plant samples. To enhance ECH and acteoside production in calli, different inducers were utilized, including 6-BA (6-Benzylaminopurine), phenylalanine, salidroside, and acteoside. Our analyses revealed that the addition of 6-BA favored ECH synthesis in calli. The addition of salidroside and phenylalanine also promoted ECH and acteoside synthesis, whereas the addition of acteoside alone enhanced ECH synthesis (Figure 2B and Supplemental Figure 2). Given that phenylalanine marks the onset of the aromatic amino acid metabolism pathway and that its addition boosted ECH yields, these results suggest that ECH synthesis may involve the downstream pathway from phenylalanine metabolism. Salidroside and acteoside are critical precursors for ECH synthesis. To identify the key enzymes involved in ECH synthesis, we performed transcriptomic analysis on these samples, aiming to fully elucidate the ECH biosynthesis pathway in C. deserticola.

Figure 2.

Figure 2

Open in a new tab

Cultivation of C. deserticola calli and the impact of different treatments on ECH and acteoside content.

(A) Morphology and culture status of C. deserticola calli at 20 days and 2 months.

(B) Changes in ECH and acteoside levels following treatment with 0.75 mg/l of 6-BA and supplementation with 0.1 mM phenylalanine, 0.05 mM salidroside, and 0.025 mM acteoside.

Statistical analyses were performed using IBM SPSS Statistics (version 21). Distribution comparisons among three groups were conducted using the Kruskal–Wallis test. Pairwise comparisons were performed using Wilcoxon’s signed rank test. Correlations among the three variables were evaluated using Spearman’s rho correlation coefficients. Graphs were generated using GraphPad Prism (version 7.02). ns, not significant. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001.

Identification of candidate genes and key metabolic pathways involved in ECH biosynthesis at the histological level

Genes involved in plant-specialized metabolic pathways are often coexpressed and may be physically colocalized, or “clustered,” within the genome (Reed et al., 2023). Coexpression analysis thus requires RNA sequencing data from various tissues and/or treatments. To systematically elucidate the ECH biosynthetic pathway in C. deserticola, we conducted transcriptomic analyses at different growth stages and in calli samples using the Illumina HiSeq 4000 platform. Trinity (Grabherr et al., 2011) and CD-HIT (Fu et al., 2012) analyses identified 50 973 to 73 572 UniGene sequences per plant sample, totaling 390 632 UniGene entries with an overall length of 743 022 718 bp and a GC content of approximately 41%. The highest UniGene count was observed in samples collected from 10 cm below ground, indicating active growth and metabolic activity for C. deserticola at this stage. In callus samples, UniGene counts ranged from 52 564 to 69 486, totaling 162 761 with a total length of 283 561 190 bp and a similar GC content of approximately 41%. UniGene abundance was the highest in 6-BA-treated tissues, highlighting the significant impact of different feeding conditions on tissue growth and metabolism. Comparison between the two tissue types revealed that the number of UniGene entries in plant samples was approximately 2.4 times higher than in calli samples, and their total length was 2.6 times greater. Analysis of lignin-related genes in C. deserticola samples revealed that the expression of candidate genes involved in p-hydroxyphenyl lignin, guaiacyl lignin, and syringyl lignin biosynthesis—such as Unigene 71614, CL12949, and 13355—increased rapidly during the pregnant bud flowering and seed-setting stages and was higher than in calli samples (Supplemental Figure 3).

To systematically analyze the transcriptome data, we employed WGCNA (Langfelder and Horvath, 2008). This method was applied to C. deserticola plant samples, calli samples, and a combined dataset of both. Co-expression network analysis identified 19 modules in plant samples, 42 modules in calli samples, and 19 modules in the combined dataset (Figure 3A). By assessing the correlation and significance between modules and traits, we identified 8 modules related to ECH and acteoside biosynthesis in plant samples. In calli samples, several modules were also related to ECH and acteoside biosynthesis, likely due to in vitro cultivation conditions inducing the expression of certain genes. However, when the two datasets were analyzed together, genes related to ECH and acteoside biosynthesis converged into a single co-expression module. This approach not only narrowed the scope of the gene search but also provided a favourable overlap of related genes in the biosynthesis pathway (Figure 3B). Analysis of trait correlations within this module clearly indicated the clustering of genes involved in ECH and acteoside biosynthesis, meeting the experimental requirements (Figure 3C). We then analyzed this gene cluster from Figure 3C using differential gene analysis to identify the candidate genes for the synthesis of ECH.

Figure 3.

Figure 3

Open in a new tab

Transcriptome analysis of C. deserticola plants and calli using WGCNA and differential expression analysis.

(A) Construction of co-expression networks using optimal thresholds, analyzing both sample types simultaneously.

(B) Identification of 19 gene modules through module–trait correlation analysis, analyzing both sample types simultaneously. The dark red module shows the strongest correlation with ECH and acteoside levels.

(C) Correlation between genes from the dark red module with acteoside (left) and ECH (right) levels.

(D) The aromatic amino acid metabolic pathway involved in ECH synthesis in C. deserticola, analyzed using experimental data and MetaboAnalyst software. Key genes in the dark red module were analyzed by differential expression analysis.

(E–G) Candidate enzymes for ECH biosynthesis: (E) possible acyltransferases; (F) possible CYP; (G) possible glycosyltransferases.

Pathway analysis of the precursors and intermediates of ECH biosynthesis, conducted with MetaboAnalyst (Pang et al., 2021) (https://www.metaboanalyst.ca/), revealed that the key pathways involve phenylalanine and tyrosine biosynthesis (Supplemental Table 1). Key enzymes in these pathways were identified by integrating differential gene expression and heatmap analyses (Figure 3D; Supplemental Figure 9), which included pathways for aromatic amino acid synthesis, tyrosine metabolism, and phenylalanine metabolism. Because the aromatic amino acid synthesis pathway in C. deserticola involves numerous genes, accurately identifying specific candidates is challenging. However, the use of callus data helped to narrow the search for target genes, as their expression levels differed significantly from those in plant tissues, aiding the identification of pathway candidates. By integrating pathway analysis, chemical structure characterization, differential gene analysis, and WGCNA, we identified key genes potentially involved in downstream biosynthetic steps, including 11 acyltransferases (Figure 3E), 96 glycosyltransferases (Figure 3F), and 37 CYP enzymes (Figure 3G).

Mining key genes for osmanthuside A synthesis

In this study, we identified salidroside as a putative upstream precursor. Based on the chemical structure of osmanthuside A, p-coumaric acid and caffeic acid were proposed as possible intermediates following the removal of salidroside from osmanthuside A. And the completion of this reaction requires the participation of hydroxycinnamoyltransferase (Nomura et al., 2022) (Figure 4A). This enzyme catalyzes the transfer of p-coumaroyl and caffeoyl groups to acceptor molecules, using their corresponding coenzyme A (CoA) derivatives generated through the activity of 4-coumarate:CoA ligase (4CL). Therefore, we screened for the 4CL gene in C. deserticola and coupled it with the previously identified acyltransferases for this reaction. As a key enzyme for lignin formation in plants (Wang et al., 2024), 4CL’s structure and function are closely related to plant evolution (Lavhale et al., 2018). It primarily functions in the phenylalanine metabolic pathway, contributing to the formation of various phenylpropanoid derivatives (Alberstein et al., 2012). To mine 4CL genes from C. deserticola, we separately performed clustering and expression abundance analyses of genes annotated as 4CL in the two sample types and identified a total of 15 candidate genes (Figure 4B). Based on expression levels, we successfully cloned 5 genes: Unigene 134093, 232499, 2309, 128003, and 129456. The formation of caffeoyl-CoA from caffeic acid, which results in a distinct color change, allowed us to tentatively assess enzyme activity. In these assays, the reaction between Unigene 129456 and caffeic acid produced a clear color change and a significant shift in absorbance (Figure 4C). Key structural domains I (AMP-binding site) and II (putative catalytic site) were identified by protein sequence homology comparison with 4CLs from Striga hermonthica, Striga asiatica, S. indicum, and Salvia miltiorrhiza, consistent with previous reports (Tang et al., 2018) (Figure 4D). To better detect Unigene 129456's enzyme activity, coupling with hydroxycinnamoyltransferase is necessary for the assay.

Figure 4.

Figure 4

Open in a new tab

Screening and functional validation of key enzymes involved in osmanthuside A biosynthesis in C. deserticola.

(A) Proposed synthetic pathway for ECH based on chemical structure inference, highlighting the synthesis of osmanthuside A from salidroside and p-coumaric acid.

(B) Candidate 4CL genes identified by comparing the plant and calli transcriptomes.

(C) Heterologous expression and enzyme activity assay of UniGene 129456 (Cd4CL1). (1) Control group: no color change in the absence of enzyme. (2) Experimental group: reaction solution turned yellow after the addition of enzyme. Shown are full-wavelength scans of control and experimental groups. The decrease in absorbance at 320 nm (caffeic acid) and increase in absorbance at 346 nm (caffeoyl-CoA) in the experimental group indicate the production of CoA compounds.

(D) Protein sequence alignment showing key structural motifs (domains I and II) of 4CL.

(E) Mining of acyltransferases converting p-coumaroyl-CoA (p-CA.CoA) or caffeoyl-CoA (CA.CoA) and salidroside to osmanthuside A and syringalide A. Peaks 1 and 2 represent osmanthuside A and syringalide A, respectively.

(F) Expression abundance of UniGene 16544 in both sample types.

(G) The 4-coumarate:CoA ligase (Cd4CL1)-conjugated acyl reaction of UniGene 16544-CdHCT.

We identified 11 candidate hydroxycinnamoyltransferase genes in plant tissues based on functional annotation, including CL10082.Contig1, CL10082.Contig2, Unigene 16544 and 51157, CL15643.Contig48, CL1719.Contig2, and Unigene 22637, 44258, 54378, 127989, and 18287. Among these, CL10082 showed the highest transcript abundance in plant samples, reaching an FPKM of 200, and was selected as a target gene. However, the reaction using salidroside and p-coumaroyl CoA or caffeoyl-CoA as substrates did not yield detectable products. Subsequent testing of other candidates revealed that the enzyme reaction containing Unigene 16544, designated CdHCT, produced new peaks with salidroside and p-coumaroyl-CoA or caffeoyl-CoA, which were identified as osmanthuside A and syringalide A by comparison with standards (Figure 4E) and liquid chromatography–mass spectrometry (LC–MS) analysis (Supplemental Figure 12). Interestingly, the expression abundance (FPKM) of CdHCT in plant samples was extremely low, ranging from 0.01 to 0.85, whereas in callus samples, transcript levels ranged from 124.77 to 280 (Figure 4F). Thus, as a parasitic plant, whether C. deserticola receives precursors from its host H. ammodendron, thereby reducing the expression of its own hydroxycinnamoyltransferases, remains an open question. However, this gene was still detected in calli, suggesting that the in vitro culture conditions induced its expression. This finding highlights certain limitations in mining pathway genes from parasitic plants. Therefore, using calli for gene discovery can overcome the contradiction between low gene expression and high ECH yield. Meanwhile, through the mining of hydroxycinnamoyltransferase genes, we also successfully verified the function of the 4CL UniGgene 129456 (Cd4CL1) (Figure 4G). Through these studies, we identified key genes involved in the biosynthesis of ECH in C. deserticola, including Cd4CL1 and CdHCT. Phylogenetic analysis showed that CdHCT shared 88.17% similarity with an acyltransferase from Phtheirospermum japonicum (Supplemental Figure 13), while Cd4CL1 was most similar to a 4-coumarate ligase from S. asiatica and S. hermonthica, with 83.96% and 84.13% similarity, respectively (Supplemental Figure 14).

Identification of key genes for ECH biosynthesis

Based on the structural characteristics of ECH, the downstream biosynthetic pathway may involve osmanthuside A as a substrate. Initially, osmanthuside A is converted to osmanthuside B through a rhamnosylation reaction, followed by hydroxylation to form acteoside, and finally through one-step glucose glycosylation to produce ECH (Figure 5A). Analysis of glycosyltransferases (Figure 2G), combined with searching callus transcriptomes, identified four genes annotated as rhamnosyltransferases: UniGgene 133948, CL14078.Contig3, CL2720.Contig1, and CL2720.Contig2 (Figure 5B). These genes were primarily associated with secondary metabolism pathways, including the biosynthesis of anthocyanins, sterols, and terpenoids; however, only a few rhamnosyltransferases have been reported to function in phenylethanol glycoside biosynthesis. Differential gene analysis between the 6-BA–treated calli (with enhanced ECH production) and the control indicated that CL2720 was upregulated (Figure 5C). CL2720 (CdRHT) was cloned and heterologously expressed in Pichia pastoris GS115. Using salidroside and p-coumaroyl CoA (or caffeoyl CoA) as substrates and UDP-Rha (uridine diphosphate rhamnose) as the sugar donor—coupled with hydroxycinnamoyltransferase—CdRHT successfully converted osmanthuside A to osmanthuside B. We also synthesized syringalide A 3′-α-L-rhamnopyranoside using syringalide A as the substrate (Figure 5D), and confirmed the results by LC–MS (Supplemental Figure 15). The expression level of CdRHT in plant samples was highest during the seed-setting and succulent stem growth stages (Supplemental Figure 16). Gene annotation identified this enzyme as involved in anthocyanin biosynthesis, which may explain its elevated expression during these stages.

Figure 5.

Figure 5

Open in a new tab

Screening and validation of key enzymes involved in ECH biosynthesis in C. deserticola.

(A) Proposed biosynthetic pathway for ECH based on chemical structure inference, using osmanthuside A as a substrate.

(B) Identification of a rhamnosyltransferase responsible for generating osmanthuside B from osmanthuside A through integration of WGCNA and functional annotation.

(C) Expression level of CL2720-CdRHT in 6-BA-fed callus samples.

(D) Mining of rhamnosyltransferases converting osmanthuside A to osmanthuside B. Peaks 3–6 represent caffeoyl-CoA, p-coumaroyl-CoA, syringalide A 3′-α-L-rhamnopyranoside, and osmanthuside B, respectively.

(E) Candidate CYPs identified for the conversion of osmanthuside B to acteoside.

(F) Expression level of UniGene 120362-CdP450 in 6-BA- and phenylalanine-fed callus samples.

(G) Expression level of UniGene 46335-CdUGT in phenylalanine- and salidroside-fed callus samples.

(H) Mining of glycosyltransferases for the production of ECH from acteoside. Peaks 7 and 8 represent acteoside and ECH, respectively.

(I) Homology analysis of UniGgene 46335-CdUGT.

To identify CYP enzymes involved in acteoside generation, we analyzed their expression during the succulent stem growth period, which corresponds to the phase of highest product accumulation (Figure 2F). We selected 12 genes with expression levels above 100 for cloning, including Unigene 210323 and 120362; CL5844.Contig1; CL9774.Contig2; CL14563.Contig10; CL12571.Contig10; CL7529.Contig10; Unigene 63408, 132605, and 221206; CL2428.Contig1; and CL878.Contig10. These genes were introduced into osmanthuside B-producing S. cerevisiae for product assays. Compared to the control group, S. cerevisiae expressing UniGene 120362 (CdP450) produced a new peak corresponding to the position of the standard for acteoside (Figure 5E), and its m/z was confirmed by LC–MS to match acteoside (Supplemental Figure 15). Additionally, the expression abundance of CdP450 in calli samples treated with 6-BA and phenylalanine was significantly higher than control group (Figure 5F).

To identify glycosyltransferases involved in converting acteoside to ECH, we analyzed calli samples that exhibited significantly enhanced ECH levels, particularly those treated with phenylalanine and salidroside. Only nine genes showed increased expression when treated with phenylalanine and salidroside, with UniGgene 46335 exhibiting the most significant increase (Figure 5G). These nine genes (UniGgene 112, 18142, 139862, 46335; CL5126.Contig360; CL12449.Contig7; CL10552.Contig3; CL7758.Contig1; and CL30671.Contig1) were cloned, heterologously expressed, and tested in vitro using acteoside and UDP-glucose as the sugar donor. Among them, only UniGgene 46335 (CdUGT) converted acteoside to ECH (Figure 5H), as confirmed by LC–MS and NMR (Supplemental Figures 15 and 17). Although CdUGT showed low expression across various growth stages, it was more abundant during the succulent stem growth period (Supplemental Figure 18). The successful identification of this key enzyme marks a significant advancement in understanding ECH biosynthesis in C. deserticola.

We then analyzed the protein sequence homology of CdRHT, CdP450, and CdUGT identified in this study. The rhamnosyltransferase most closely related to CdRHT was from P. japonicum, showing 91.45% sequence similarity (Supplemental Figure 19). The CYP enzyme with the highest similarity to CdP450 was p-coumarate-3-hydroxylase from R. glutinosa, with 87.65% similarity (Supplemental Figure 20). The glycosyltransferase CdUGT, involved in converting acteoside to ECH, shared the highest similarity (64.88%) with scopoletin glucosyltransferase from P. japonicum (Figure 5I). The finding that the genes involved in ECH synthesis in C. deserticola have high homology to those from R. glutinosa and P. japonicum may indicate a relatively close evolutionary relationship among these species.

Through these studies, we successfully identified key genes involved in ECH biosynthesis in C. deserticola and performed homology analysis of the corresponding enzymes. To investigate key enzyme–substrate interactions, we predicted the protein structures of four key enzymes using AlphaFold2 and analyzed their binding with corresponding compounds using AutoDock Vina (Supplemental Figures 22–25). The prediction of structures and binding modes/affinities provides a solid foundation for future studies on enzyme catalysis mechanisms.

Construction of a microbial cell factory for the synthesis of ECH

ECH, a complex PhG, is challenging to synthesize chemically in large quantities. Artificially grown C. deserticola contains low levels of ECH, and its extraction is resource intensive. Although callus yield has improved slightly, harsh culture conditions still limit large-scale synthesis. Building on known metabolic pathways (Yang et al., 2023) and the ECH biosynthetic pathway and key genes identified in this study, we aimed to reconstruct the ECH synthetic pathway in microbial systems for biosynthesis purposes (Figure 6A).

Figure 6.

Figure 6

Open in a new tab

Construction of a microbial cell factory for ECH biosynthesisusing S. cerevisiae as a chassis organism.

(A) Schematic diagram of the ECH biosynthetic pathway reconstructed in S. cerevisiae.

(B) Production of salidroside, p-coumaric acid, and osmanthuside B using sd-1 as the starting strain.

(C) Construction of gene expression cassettes encoding key enzymes.

(D) Biosynthesis of ECH and acteoside in strain sg-1 upon supplementation with exogenous p-coumaric acid.

(E) HPLC chromatograms of key products of ECH biosynthesis. Peaks 6–9 represent osmanthuside B, acteoside, p-coumaric acid, and ECH, respectively.

(F) Yields of de novo synthesized ECH and acteoside in the engineered sg-2 strain.

S. cerevisiae, known for its high yield of salidroside and p-coumaric acid—both precursors in ECH biosynthesis—was selected as the optimal chassis organism (Liu et al., 2021; Chen et al., 2022). The CEN.PK2-1C derivative strain sd-1 (Supplemental Table 3), capable of producing salidroside, was chosen as the starting strain. To enhance p-coumaric acid production, we introduced and amplified the CoTAL gene at the TRP2 and NTS loci (Li et al., 2020), knocked out the competitive PHA2 pathway, and inserted multiple copies of ARO4K229L and ARO7G141S. This modified strain, sd-2, produced salidroside and p-coumaric acid at concentrations of 400.57 ± 2.31 mg/l and 16.17 ± 0.21 mg/l, respectively, as determined by high-performance liquid chromatography (HPLC) (Figure 6B).

We next aimed to construct the biosynthetic pathway for osmanthuside B in strain sd-2, as osmanthuside B is a key intermediate in the biosynthesis of many phenylethanol glycosides, including acteoside and ECH. Among candidate enzymes reported to generate osmanthuside A from salidroside and p-coumaric acid, the S. indicum-derived acyltransferase SiAT1 exhibited the highest activity. Thus, we introduced Pt4CL1 and SiAT1 at the GAL80 locus. Because S. cerevisiae cannot convert UDP-glucose to UDP-rhamnose, VvRhm (Pei et al., 2018) and CdRHT were introduced at the MET17 locus. The resulting strain was designated sd-3. HPLC analysis showed an osmanthuside B yield of 3.66 ± 1.05 mg/l (Figure 6B). To synthesize ECH, we simultaneously introduced CdP450 and MtCPR at the HO locus and CdUGT at the δ locus in sd-3, generating strain sg-1. The gene expression cassette was constructed as shown in Figure 6C. Initial attempts failed to produce ECH, although salidroside was still detected, suggesting a limitation of p-coumaric acid availability in the pathway. After supplementing with 600 mg/l of p-coumaric acid, the production of acteoside in strain sg-1 reached 2.25 ± 0.25 mg/l, and ECH production was 5.42 ± 1.34 mg/l (Figure 6D).

To achieve de novo synthesis of ECH, we leveraged the robust exocytosis capabilities of S. cerevisiae BY4742. In this project, BY4742 was engineered to reconstruct the p-coumaric acid synthesis pathway according to a previous report (Liu et al., 2019). The resulting strain was fused with sg-1 to create strain sg-2. When fermented with 2% glucose, sg-2 produced acteoside at 4.19 ± 0.14 mg/l and ECH at 7.52 ± 1.42 mg/l (Figure 6E and 6F). We successfully achieved de novo synthesis of ECH in S. cerevisiae, demonstrating the feasibility of producing ECH using a microbial cell factory.

Discussion

C. deserticola, a traditional Chinese medicinal plant, exhibits a wide range of pharmacological activities and holds significant research and pharmaceutical value. The plant’s pharmacological properties primarily arise from its diverse natural products, with ECH as its signature compound. ECH is extensively used in medical treatments, health care products, and cosmetics, presenting substantial market potential (Liu et al., 2018a; 2018b). However, its complex structure has hindered large-scale chemical synthesis, while plant extraction is constrained by resource limitations. Therefore, developing novel synthesis methods is crucial for the industrial production of ECH. This study addresses this challenge by reconstructing the ECH biosynthetic pathway.

For the project, we selected C. deserticola (Huang et al., 2023), a plant known for its high ECH content, as our research model. Metabolomic and transcriptomic analyses were conducted at various plant growth phases to identify key precursors and genes involved in ECH biosynthesis. Our metabolomics studies revealed that salidroside levels were inversely related to ECH levels, suggesting its potential role as a precursor in ECH synthesis. To further investigate this hypothesis and minimize host interference, we established cell suspension cultures supplemented with salidroside, phenylalanine, or acteoside. These treatments significantly enhanced ECH content. This dual-metabolomics approach proved effective for our study and may also be suitable for studying other parasitic plants (Fan et al., 2023).

Correlation networks are increasingly applied in bioinformatics, particularly for gene expression analysis. WGCNA, a systems biology method that delineates correlation patterns among genes across microarray samples (Gao et al., 2023; Sun et al., 2024), is predominantly used to identify disease-related genes but is now increasingly applied in the identification of key plant genes (Haase et al., 2021; Xu et al., 2022; Wang et al., 2023a). Transcriptome analysis of C. deserticola across various growth stages and under different calli treatments identified 390 632 UniGene entries—too many for differential expression analysis alone to effectively identify candidates for ECH biosynthesis. In this study, we combined WGCNA and differential expression analysis to successfully narrow the list of candidate genes. By integrating data from both plant tissue and calli samples, we were able to superimpose key genes, significantly reducing the complexity of gene mining. This approach provided a solid foundation for successful identification of ECH biosynthesis genes.

This study leveraged a systems biology approach by systematically analyzing the metabolomes and transcriptomes of C. deserticola across two culture systems to elucidate the synthetic pathway of ECH. In parallel, we expanded the enzyme libraries for glycosyltransferases, 4-coumarate ligases, CYPs, and hydroxycinnamoyltransferases. Notably, the glycosyltransferase involved in ECH biosynthesis is a newly identified enzyme, marking its first application in ECH production. This study also deepened our understanding of key enzymes through ligand-binding analysis. Interestingly, acyltransferase genes showed low expression levels in plant tissues but high expression levels in calli, providing insights into gene regulation specificity in parasitic plants. In addition to characterizing the ECH pathway, this work also yielded several PhGs, providing a robust foundation for future study of their pharmacological activities.

The advancement of synthetic biology offers a sustainable strategy for producing plant-derived natural products. A well-defined biosynthetic pathway is essential for advancing synthetic biology research. Therefore, elucidating the complete biosynthetic pathway of ECH not only enhances our understanding of plant secondary metabolism but also holds practical importance for guiding the industrial production of ECH. In this study, we successfully achieved de novo synthesis of ECH in S. cerevisiae, demonstrating the feasibility of microbial production and laying a solid foundation for the industrial production of ECH.

Through metabolomic studies on C. deserticola plants at various growth stages and under different callus culture conditions, we identified the key precursors required for ECH biosynthesis. Transcriptomic analyses of both plant and callus samples, coupled with WGCNA and comparative histology, enabled the identification of candidate genes essential for ECH biosynthesis. Using biological assays, we effectively screened these candidate genes, identified those crucial for ECH synthesis, and examined their ligand-receptor interactions. This work expanded both the enzyme library and the catalog of PhGs. To facilitate industrial production, we experimented with synthesizing ECH in a microbial cell factory and successfully accomplished de novo synthesis in S. cerevisiae.

Methods

Materials

Analytical-grade substances, including acteoside, ECH, osmanthuside A, osmanthuside B, salidroside, p-coumaric acid, and caffeic acid, were obtained from Phytostandard Pure Biotechnology. Plant materials were collected from Wolin Ecological Park, Qitai County, Changji Hui Autonomous Prefecture, Xinjiang, China. The initial strain used for microbial cell factory construction, designated sd-1, was generously provided by Prof. Yunzi Luo’s laboratory at Tianjin University. Primers were purchased from GENEWIZ Biotech and are detailed in Supplemental Table 2. Engineered strains were developed using the CRISPR–Cas system as listed in Supplemental Table 3. Plasmids for CRISPR–Cas9 were assembled using the Golden Gate method and generously provided by Prof. Sheng Yang. Guide RNAs were designed via CRISPRdirect (http://crispr.dbcls.jp/).

Sampling and processing of C. deserticola (artificially cultivated)

Based on the growth stages of C. deserticola, six sampling points were selected: (1) succulent stem growth period, (2) 10 cm below ground, (3) topsoil stage, (4) 10 cm above ground, (5) pregnant bud flowering period, and (6) seed-setting stage. All samples were collected from the same plot, free of pests, disease, and mechanical damage. At each stage, 8–9 roots were harvested, and 3 representative plants were selected for transcriptome sequencing. Depending on the degree of lignification, each plant was segmented into 2–5 sections and cut into approximately 1 cm3 pieces. Tissues were immediately submerged in liquid nitrogen, transferred to dry ice, and stored at −80°C. A portion of each sample was dried in a Petri dish at 60 °C until constant weight, ground, and passed through a 60-mesh screen. Exactly 1.0 g of powdered sample was transferred to a 30 ml volumetric flask, fixed with 75% methanol, and mass was confirmed in a centrifuge tube. After standing for 30 min, samples were ultrasonicated for 40 min (power 250 W, frequency 35 kHz), cooled, reweighed, and adjusted to original mass with 75% methanol. After shaking and standing, the supernatant was filtered through a 0.22 μm organic membrane and stored for analysis.

Culture of C. deserticola calli

Fresh young fleshy stem buds were rinsed under running water for 24 h, stored at 7 °C for at least 5 h, then rinsed in distilled water. Explants were sterilized with 75% ethanol for 30 s at an ultra-clean bench, soaked in 10% NaClO for 5 min, and rinsed 3–5 times in sterile water. The treated C. deserticola explants were cut into 0.5 cm3 pieces and inoculated onto callus induction medium containing 0.5 mg/l indoleacetic acid, 2 mg/l agonist, and 3 mg/l vitamin C with 4–6 explants per vial. The medium was based on Murashige and Skoog medium, supplemented with 30 g/l sucrose, 4.2 g/l agar, 500 mg/l amicase, and 0.6 g/l activated charcoal, adjusted to pH 5.6–6.0, and sterilized at 115 °C for 20 min. Callus formation was observed after approximately 10 days of incubation in the dark at 25 °C.

Treatment of C. deserticola calli

Calli were cultured for 20 days prior to treatment. The primary treatment compounds included 0.05 mM salidroside, 0.025 mM acteoside, 0.1 mM phenylalanine, and 0.75 mg/l 6-BA, prepared as a master stock. Calli were transferred to treatment medium containing 0.5 mg/l indoleacetic acid, 2 mg/l agonist, and 3 mg/l vitamin C. Treatment reagents were added at varying concentrations, and cultures were incubated for 7 days, and samples were collected every 24 h over 7 days. Samples were dried at 60 °C to constant weight, after which 10 mg of each sample was extracted with 1 ml of 75% methanol and processed as previously described. The medium used was 1/2 Murashige and Skoog supplemented with 20 g/l sucrose, 500 mg/l amicase, and 0.6 g/l activated charcoal. The pH was adjusted to 5.6–6.0 prior to sterilization at 115 °C for 20 min.

Transcriptome sequencing and analysis of C. deserticola plants and calli

Plant and callus samples showing significant differences in ECH or acteoside content—either across different batches of artificially cultivated plants or between treatment and control callus groups—were selected for transcriptome sequencing following HPLC analysis. Samples stored at −80 °C were used for total RNA extraction, mRNA library construction, cDNA sequencing, and de novo assembly. Transcriptome sequencing was performed by the Beijing Genomics Institute. Differential gene expression analysis and gene annotation were conducted using assembled transcripts. Gene functions were inferred through integration with KEGG, GO, and other databases. Open reading frames were predicted using TransDecoder and aligned against the SwissProt and Pfam databases. Five additional databases (NCBI’s Nt and Nr, GO, KOG, and KEGG) were used for transcript annotation. Software tools such as CPC, txCdsPredict, and CNCI, along with the Pfam database, were used to predict coding and non-coding sequences. Candidate gene sequences identified through these analyses were used to design primers for target gene cloning and cDNA library construction.

Cloning, heterologous expression, and metabolic pathway construction of selected genes

To optimize the expression of plant-derived eukaryotic genes, multiple host systems were employed. P. pastoris GS115 was used for the expression of glycosyltransferases; Escherichia coli BL21(DE3) was used for the expression of glycosyltransferases, hydroxycinnamoyltransferase, and 4-coumarate:CoA ligase; and S. cerevisiae CEN.PK2-1C was used for both heterologous enzyme expression and metabolic pathway reconstruction. E. coli DH5α was used for plasmid construction, including recombinant expression cassettes and CRISPR–Cas9 systems, and cultured in Luria–Bertani medium at 37 °C with shaking at 200 rpm. For heterologous expression in P. pastoris, the target genes were cloned into the pGAP-Zα shuttle vector, treated with BlnI, and electroporated into GS115 (Zhang et al., 2023). The resulting transformants were cultured in buffered YPD medium (Yeast Extract Peptone Dextrose Medium) at 30 °C and 200 rpm until reaching an optical density (OD600) of 2.0–5.0, after which the cells were harvested for further analysis. The pET-28a vector was used for expression in E. coli BL21(DE3). For pathway construction in S. cerevisiae, the pESC shuttle plasmid was used to introduce target genes into CEN.PK2-1C via chemical transformation (Gietz and Schiestl, 2007). Engineered yeast strains were screened using SD dropout medium supplemented with appropriate dropout nutrients (His/−Leu−Trp−Ura).

Microbial product isolation and detection

Validated yeast strains were pre-cultured in 5 ml YPD medium at 30 °C and 200 rpm for 30 h, then transferred to 20 mL fresh YPD medium at a 5% (v/v) inoculation ratio. After 3 days of fermentation in P. pastoris GS115, 1 ml of yeast culture was mixed with beads and lysed using a freeze grinder at 70 Hz for 60 s with 20-s intervals, repeated 5 times at 4 °C. A mixture containing 2 mM acteoside, 4 mM UDP-glucose, and 200 μl of lysed yeast product was incubated for 10 h. The reaction was terminated with 50% methanol and analyzed by HPLC. New products were further identified using HPLC–MS to determine m/z values, followed by 1H- and 13C-NMR spectroscopy for structural confirmation.

Compound–protein interaction studies

Protein 3D structures were predicted using the AlphaFold deep learning-based modeling platform (Pinheiro et al., 2021; Editorial, 2023). Before molecular docking, 3D structures of acteoside, osmanthuside A, and osmanthuside B were optimized using data from PubChem (Kim et al., 2016). Interactions between the compounds and proteins were analyzed via semi-flexible docking using AutoDock Tools. The binding pocket was defined based on predicted catalytic residues, and the docking poses were assessed using POCASA.

Analytical methods

Quantitative analysis of PhGs was performed using an HPLC system equipped with a C18 column (5 μm, 250 × 4.6 mm; Shimadzu, Japan). The flow rate was maintained at 0.8 ml/min, with an injection volume of 5 μl. Gradient elution was performed using 5% acetonitrile for the first 5 min, increasing to 13.5% at 5–8 min, 30% at 18–43 min, 100% at 43–54 min, and re-equilibration with 5% methanol at 54–60 min.

Data availability

All data supporting the conclusions of this study are included in the main text and/or supplementary materials. Additional data related to this paper may be obtained upon request from the corresponding authors. Sequences functionally characterized in this study have been deposited in the GenBank database under the following accession numbers: CdHCT, PV641602; Cd4CL1, PV641603; CdP450, PV641605; CdRHT, PV641604; and CdUGT, PV641606. Raw sequencing data for transcriptome analysis have been deposited in the National Genomics Data Center (NGDC) under accession number PRJCA041348.

Funding

This research was financially supported by National Natural Science Foundation of China (22478031).

Acknowledgments

We are very grateful to Mr. Zongcheng Geng and Mr. Yingwen Jiang from Xinjiang Wolin Ecology for their guidance and assistance during the sampling of C. deserticola. We also thank Prof. Mingjia Yu from Beijing Institute of Technology for providing the experimental platform for tissue culture, and Prof. Yunzi Luo from Tianjin University for providing the chassis cells for salidroside production. There are no conflicts of interest among the institutions and individuals involved.

Author contributions

Y.F., B.L., and C.L., funding acquisition, project administration, guidance, conceptualization, writing – review & editing; B.L., software and visualization; Y.B., investigation, methodology, software, resources, writing – original draft, and writing – review & editing; J.J., formal analysis and investigation; H.Y., investigation, resources, and software; H.J., writing – review & editing; Y.P. and X.C., resources and data curation; J.Y. and Q.L., data curation and resources.

Published: June 24, 2025

Footnotes

Supplemental information is available at Plant Communications Online.

Contributor Information

Chun Li, Email: lichun@tsinghua.edu.cn.

Bo Lv, Email: lv-b@bit.edu.cn.

Yongjun Feng, Email: fengyj@bit.edu.cn.

Supplemental information

Document S1. Supplemental Figures 1–28 and Supplemental Tables 1–3
mmc1.pdf (5.7MB, pdf)
Document S2. Article plus supplemental information
mmc2.pdf (32MB, pdf)

References

  1. Alberstein M., Eisenstein M., Abeliovich H. Removing allosteric feedback inhibition of tomato 4-coumarate: CoA ligase by directed evolution. Plant J. 2012;69:57–69. doi: 10.1111/j.1365-313x.2011.04770.x. [DOI] [PubMed] [Google Scholar]
  2. An T., Lin G., Liu Y., Qin L., Xu Y., Feng X., Li C. De novo biosynthesis of anticarcinogenic icariin in engineered yeast. Metab. Eng. 2023;80:207–215. doi: 10.1016/j.ymben.2023.10.003. [DOI] [PubMed] [Google Scholar]
  3. Chen C., Xia B., Tang L., Wu W., Tang J., Liang Y., Yang H., Zhang Z., Lu Y., Chen G., et al. Echinacoside protects against MPTP/MPP+-induced neurotoxicity via regulating autophagy pathway mediated by Sirt1. Metab. Brain Dis. 2019;34:203–212. doi: 10.1007/s11011-018-0330-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen R., Gao J., Yu W., Chen X., Zhai X., Chen Y., Zhang L., Zhou Y.J. Engineering cofactor supply and recycling to drive phenolic acid biosynthesis in yeast. Nat. Chem. Biol. 2022;18:520–529. doi: 10.1038/s41589-022-01014-6. [DOI] [PubMed] [Google Scholar]
  5. Chuang H.W., Wang T.Y., Huang C.C., Wei I.H. Echinacoside exhibits antidepressant-like effects through AMPAR-Akt/ERK-mTOR pathway stimulation and BDNF expression in mice. Chin. Med. 2022;17:9. doi: 10.1186/s13020-021-00549-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Editorial AlphaFold and beyond. Nat. Methods. 2023;20:163. doi: 10.1038/s41592-023-01790-6. [DOI] [PubMed] [Google Scholar]
  7. Fan Y., Zhao Q., Duan H., Bi S., Hao X., Xu R., Bai R., Yu R., Lu W., Bao T., Wuriyanghan H. Large-scale mRNA transfer between Haloxylon ammodendron (Chenopodiaceae) and herbaceous root holoparasite Cistanche deserticola (Orobanchaceae) iScience. 2023;26 doi: 10.1016/j.isci.2022.105880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Fu L., Niu B., Zhu Z., Wu S., Li W. CD-HIT: Accelerated for clustering the next-generation sequencing data. Bioinformatics. 2012;28:3150–3152. doi: 10.1093/bioinformatics/bts565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gao M.R., Wang M., Jia Y.Y., Tian D.D., Liu A., Wang W.J., Yang L., Chen J.Y., Yang Q., Liu R., Wu Y.M. Echinacoside protects dopaminergic neurons by inhibiting NLRP3/Caspase-1/IL-1β signaling pathway in MPTP-induced Parkinson’s disease model. Brain Res. Bull. 2020;164:55–64. doi: 10.1016/j.brainresbull.2020.08.015. [DOI] [PubMed] [Google Scholar]
  10. Gao X.M., Zhou X.H., Jia M.W., Wang X.Z., Liu D. Identification of key genes in sepsis by WGCNA. Prev. Med. 2023;172 doi: 10.1016/j.ypmed.2023.107540. [DOI] [PubMed] [Google Scholar]
  11. Gietz R.D., Schiestl R.H. Large-scale high-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method. Nat. Protoc. 2007;2:38–41. doi: 10.1038/nprot.2007.15. [DOI] [PubMed] [Google Scholar]
  12. Grabherr M.G., Haas B.J., Yassour M., Levin J.Z., Thompson D.A., Amit I., Adiconis X., Fan L., Raychowdhury R., Zeng Q., et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 2011;29:644–652. doi: 10.1038/nbt.1883. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Haase F., Gloss B.S., Tam P.P.L., Gold W.A. WGCNA identifies translational and proteasome-ubiquitin dysfunction in Rett syndrome. Int. J. Mol. Sci. 2021;22:9954. doi: 10.3390/ijms22189954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. He F., Chen L., Liu Q., Wang X., Li J., Yu J. Preparative separation of phenylethanoid and secoiridoid glycosides from Ligustri lucidi fructus by high-speed counter-current chromatography coupled with ultrahigh pressure extraction. Molecules. 2018;23:3353. doi: 10.3390/molecules23123353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hou L., Li G., Chen Q., Zhao J., Pan J., Lin R., Zhu X., Wang P., Wang X. De novo full length transcriptome analysis and gene expression profiling to identify genes involved in phenylethanol glycosides biosynthesis in Cistanche tubulosa. BMC Genom. 2022;23:698. doi: 10.1186/s12864-022-08921-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Huang J., Peng Y., Ji Z., Guo L., Hao J., Zhang Z., Zhao D. Drying characteristics of Cistanche deserticola with different pretreatment and degradation mechanism of phenylethanoid glycosides during drying. Ind. Crops Prod. 2023;204 doi: 10.1016/j.indcrop.2023.117339. [DOI] [Google Scholar]
  17. Jiang B., Gao L., Wang H., Sun Y., Zhang X., Ke H., Liu S., Ma P., Liao Q., Wang Y., et al. Characterization and heterologous reconstitution of Taxus biosynthetic enzymes leading to baccatin III. Science. 2024;383:622–629. doi: 10.1126/science.adj3484. [DOI] [PubMed] [Google Scholar]
  18. Kim S., Thiessen P.A., Bolton E.E., Chen J., Fu G., Gindulyte A., Han L., He J., He S., Shoemaker B.A., et al. PubChem substance and compound databases. Nucleic Acids Res. 2016;44:D1202–D1213. doi: 10.1093/nar/gkv951. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Langfelder P., Horvath S. WGCNA: An R package for weighted correlation network analysis. BMC Bioinf. 2008;9:559. doi: 10.1186/1471-2105-9-559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lavhale S.G., Kalunke R.M., Giri A.P. Structural, functional and evolutionary diversity of 4-coumarate-CoA ligase in plants. Planta. 2018;248:1063–1078. doi: 10.1007/s00425-018-2965-z. [DOI] [PubMed] [Google Scholar]
  21. Li Y., Mao J., Song X., Wu Y., Cai M., Wang H., Liu Q., Zhang X., Bai Y., Xu H., Qiao M. Optimization of the l-tyrosine metabolic pathway in Saccharomyces cerevisiae by analyzing p-coumaric acid production. 3 Biotech. 2020;10:258. doi: 10.1007/s13205-020-02223-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Liu J., Yang L., Dong Y., Zhang B., Ma X. Echinacoside, an inestimable natural product in treatment of neurological and other disorders. Molecules. 2018;23:1213. doi: 10.3390/molecules23051213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Liu W., Can Y., Song Q., Zhen J., Zhao Y., Tu P., Li J., Song Y. Chemical characterization for flowers and lignified stems of Cistanche deserticola. China J. Chin. Mater. Med. 2018;43:3708–3714. doi: 10.19540/j.cnki.cjcmm.20180612.001. [DOI] [PubMed] [Google Scholar]
  24. Liu Q., Yu T., Li X., Chen Y., Campbell K., Nielsen J., Chen Y. Rewiring carbon metabolism in yeast for high level production of aromatic chemicals. Nat. Commun. 2019;10:4976. doi: 10.1038/S41467-019-12961-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Liu H., Tian Y., Zhou Y., Kan Y., Wu T., Xiao W., Luo Y. Multi-modular engineering of Saccharomyces cerevisiae for high-titre production of tyrosol and salidroside. Microb. Biotechnol. 2021;14:2605–2616. doi: 10.1111/1751-7915.13667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Martin L.B.B., Kikuchi S., Rejzek M., Owen C., Reed J., Orme A., Misra R.C., El-Demerdash A., Hill L., Hodgson H., et al. Complete biosynthesis of the potent vaccine adjuvant QS-21. Nat. Chem. Biol. 2024;20:493–502. doi: 10.1038/s41589-023-01538-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mulani S.K., Guh J.H., Mong K.K.T. A general synthetic strategy and the anti-proliferation properties on prostate cancer cell lines for natural phenylethanoid glycosides. Org. Biomol. Chem. 2014;12:2926–2937. doi: 10.1039/c3ob42503g. [DOI] [PubMed] [Google Scholar]
  28. Nomura T., Yoneda A., Kato Y. BAHD acyltransferase induced by histone deacetylase inhibitor catalyzes 3-O-hydroxycinnamoylquinic acid formation in bamboo cells. Plant J. 2022;112:1266–1280. doi: 10.1111/tpj.16013. [DOI] [PubMed] [Google Scholar]
  29. Pang Z., Chong J., Zhou G., De Lima Morais D.A., Chang L., Barrette M., Gauthier C., Jacques P.É., Li S., Xia J. MetaboAnalyst 5.0: Narrowing the gap between raw spectra and functional insights. Nucleic Acids Res. 2021;49:W388–W396. doi: 10.1093/nar/gkab382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Pei J., Chen A., Sun Q., Zhao L., Cao F., Tang F. Construction of a novel UDP-rhamnose regeneration system by a two-enzyme reaction system and application in glycosylation of flavonoid. Biochem. Eng. J. 2018;139:33–42. doi: 10.1016/j.bej.2018.08.007. [DOI] [Google Scholar]
  31. Pinheiro F., Santos J., Ventura S. AlphaFold and the amyloid landscape. J. Mol. Biol. 2021;433 doi: 10.1016/j.jmb.2021.167059. [DOI] [PubMed] [Google Scholar]
  32. Qian H., Yu F., Geng Z., Xie J., Xie D., Qi Z., Ma Z. Comparative study of the content of four phenylethanoid glycosides in Cistanche deserticola Y.C.Ma from different seasons. Chin. J. Pharm. Ana. 2016;36:1971–1976. doi: 10.16155/j.0254-1793.2016.11.11. [DOI] [Google Scholar]
  33. Reed J., Orme A., El-Demerdash A., Owen C., Martin L.B.B., Misra R.C., Kikuchi S., Rejzek M., Martin A.C., Harkess A., et al. Elucidation of the pathway for biosynthesis of saponin adjuvants from the soapbark tree. Science. 2023;379:1252–1264. doi: 10.1126/science.adf3727. [DOI] [PubMed] [Google Scholar]
  34. Sun Q., Wang Z., Xiu H., He N., Liu M., Yin L. Identification of candidate biomarkers for GBM based on WGCNA. Sci. Rep. 2024;14 doi: 10.1038/s41598-024-61515-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Tang Y.H., Liu F., Mao K.Q., Xing H.C., Chen J.R., Guo Q.Q. Cloning and characterization of the key 4-coumarate CoA ligase genes in Boehmeria nivea. South Afr. J. Bot. 2018;116:123–130. doi: 10.1016/j.sajb.2018.02.398. [DOI] [Google Scholar]
  36. Wang R., Wang Y., Yao W., Ge W., Jiang T., Zhou B. Transcriptome sequencing and WGCNA reveal key genes in response to leaf blight in poplar. Int. J. Mol. Sci. 2023;24 doi: 10.3390/ijms241210047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Wang W., Jiang S., Zhao Y., Zhu G. Echinacoside: A promising active natural products and pharmacological agents. Pharmacol. Res. 2023;197 doi: 10.1016/j.phrs.2023.106951. [DOI] [PubMed] [Google Scholar]
  38. Wang M., Jia L., Liu Y., Zhang C., Chen Y., Lin Y., Huang Y., Lai Z. Genome-wide identification, evolutionary analysis of 4CL gene family and expression analysis in banana. Food Agric. Immunol. 2024;35 doi: 10.1080/09540105.2023.2287958. [DOI] [Google Scholar]
  39. Wu G., Wang X., Wang J., Wang J., Zhang L., Zhang H., Wang P. Progress of Cistanche deserticola extract and its whitening and anti-aging mechanism in cosmetics. Technique express. 2023;8:1006–7264. doi: 10.3969/j.issn.1006-7264.2023.08.007. [DOI] [Google Scholar]
  40. Xu L., Cao M., Wang Q., Xu J., Liu C., Ullah N., Li J., Hou Z., Liang Z., Zhou W., Liu A. Insights into the plateau adaptation of Salvia castanea by comparative genomic and WGCNA analyses. J. Adv. Res. 2022;42:221–235. doi: 10.1016/j.jare.2022.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Yang Y., Wu Y., Zhuang Y., Liu T. Discovery of glycosyltransferases involved in the biosynthesis of ligupurpuroside B. Org. Lett. 2021;23:7851–7854. doi: 10.1021/acs.orglett.1c02873. [DOI] [PubMed] [Google Scholar]
  42. Yang Y., Xi D., Wu Y., Liu T. Complete biosynthesis of the phenylethanoid glycoside verbascoside. Plant Commun. 2023;4 doi: 10.1016/j.xplc.2023.100592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Yi Q., Sun M., Jiang G., Liang P., Chang Q., Yang R. Echinacoside promotes osteogenesis and angiogenesis and inhibits osteoclast formation. Eur. J. Clin. Invest. 2024;54 doi: 10.1111/eci.14198. [DOI] [PubMed] [Google Scholar]
  44. Zhang X., Hao Y. Beneficial effects of echinacoside on diabetic cardiomyopathy in diabetic Db/Db mice. Drug Des. Devel. Ther. 2020;14:5575–5587. doi: 10.2147/DDDT.S276972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zhang C., Du J., Tang X., Ma D., Qin L., Zhang A., Jiang N. Electrotransformation optimization of plasmid pGAPZαA-CecMd3cs into Pichia pastoris GS115 with response surface methodology. Electron. J. Biotechnol. 2023;61:54–60. doi: 10.1016/j.ejbt.2022.11.002. [DOI] [Google Scholar]
  46. Zheng S., Jiang X., Wu L., Wang Z., Huang L. Chemical and genetic discrimination of cistanches herba based on UPLC-QTOF/MS and DNA barcoding. PLoS One. 2014;9 doi: 10.1016/j.isci.2022.105880. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Supplemental Figures 1–28 and Supplemental Tables 1–3
mmc1.pdf (5.7MB, pdf)
Document S2. Article plus supplemental information
mmc2.pdf (32MB, pdf)

Data Availability Statement

All data supporting the conclusions of this study are included in the main text and/or supplementary materials. Additional data related to this paper may be obtained upon request from the corresponding authors. Sequences functionally characterized in this study have been deposited in the GenBank database under the following accession numbers: CdHCT, PV641602; Cd4CL1, PV641603; CdP450, PV641605; CdRHT, PV641604; and CdUGT, PV641606. Raw sequencing data for transcriptome analysis have been deposited in the National Genomics Data Center (NGDC) under accession number PRJCA041348.

Articles from Plant Communications are provided here courtesy of Elsevier

RESOURCES