Abstract

Background

   Dendrobium orchids (Dendrobium spp.) are valuable medicinal and
   attractive ornamental plants. Due to their limited wild resources, the
   size of the Dendrobium spp. population required for market demand
   primarily depends on artificial cultivation. However, the nutritional
   and therapeutic value of natural products may differ as growth
   conditions change. In this study, we profiled metabolites from wild and
   cultivated Dendrobium flexicaule (D. flexicaule) to explore the
   variations and interrelationships among bioactive components.

Results

   A total of 840 annotated metabolites were discovered, 231 of which
   differed significantly between wild and cultivated D. flexicaule. A
   comparative investigation found that the types and amounts of
   metabolites, particularly flavonoids, lipids, amino acids and their
   derivatives, varied between wild and cultivated D. flexicaule. Using
   metabolite correlation analysis, a series of differentially abundant
   metabolites were found to be significantly correlated with
   phytohormones such as abscisic acid (ABA), salicylic acid (SA), and
   zeatins, indicating that plant hormones play a role in the accumulation
   of specific metabolites. Furthermore, many distinct metabolites were
   identified as key active ingredients of traditional Chinese medicines.
   Additionally, 78 components were discovered to be active pharmaceutical
   substances against various diseases, probably contributing to the
   diverse medical values of wild and cultivated D. flexicaule.

Conclusions

   Overall, comprehensively analyzed the metabolic profiles of wild and
   cultivated D. flexicaule in this study, serving as a theoretical and
   material foundation for quality control, health efficacy, and
   industrial development.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12870-025-06054-3.

   Keywords: Dendrobium flexicaule, Metabolome, Differential metabolites,
   Metabolite correlation analysis, Phytohormone

Introduction

   Dendrobium orchids (Dendrobium spp.), a well-known traditional Chinese
   medicinal plant, is highly valuable owing to its ecological and
   economic significance [[40]1]. However, due to environmental concerns
   and peculiar growth conditions, Dendrobium spp. grown in wild forests
   are endangered. Cultivated Dendrobium spp. have become popular due to
   their commercial demand, prompting the development of artificial
   cultivation methods. It is widely accepted that natural metabolites
   such as polysaccharide, flavones, and phenolic acids are the primary
   resources that determine the nutritional quality and consumer appeal of
   Dendrobium spp., and all of these metabolites vary with the growth
   environment (e.g. cultivation substrate, light intensity, and
   temperature) [[41]2–[42]4]. Therefore, it is crucial to distinguish
   between wild and cultivated Dendrobium spp. Furthermore, the quality of
   cultivated Dendrobium spp. should be considered in terms of yield and
   effective ingredients.

   Insights into the metabolic profiles of Dendrobium spp. grown under
   different environmental conditions will provide valuable information
   for cultivating and efficiently using distinct Dendrobium spp. Several
   recent studies have evaluated the quality of distinct Dendrobium spp.
   produced in different regions, under different substrates, and under
   stressful conditions using metabolite comparison analysis. For
   instance, more valuable metabolites, including polysaccharides,
   polyphenols, and flavonoids, have been detected in Dendrobium
   officinale (D. officinale) grown in forests than in plants grown in
   greenhouses [[43]5]. Similarly, significant differences in metabolite
   content and composition were observed in Dendrobium huoshanense (D.
   huoshanense) after stone planting in forests and greenhouses [[44]6].
   Variable metabolites have been detected in D. officinale plants grown
   under different epiphytic cultivation conditions (e.g., tree or stone)
   or on different substrates within the same growth region [[45]5,
   [46]7].

   Additionally, research has been conducted on how metabolites are
   altered in Dendrobium spp. under severe growth conditions in recent
   years. For example, the accumulation of polysaccharides, flavonoids,
   and anthocyanins were apparent under high-intensity light (HL) or UV-B
   light treatments [[47]8, [48]9], whereas only anthocyanin accumulation
   increased under red-blue light [[49]10]. Dendrobium spp. are
   drought-tolerant and increased accumulation of metabolites (such as
   flavonoids and osmolytes) under drought stress in Dendrobium
   moniliforme has also been reported [[50]11, [51]12]. Dendrobium spp.
   growth is sensitive to heat stress, and significantly increased levels
   of metabolites, including chlorophyll, carotenoids, glutathione (GSH),
   and flavonoids, increase antioxidant activity for survival under heat
   conditions [[52]13]. Collectively, these studies strongly suggest that
   different growth conditions strongly influence the content and
   compositions of natural metabolites in Dendrobium spp.

   Integrated metabolomic and transcriptomic combined analyses have been
   widely performed to reveal the metabolic molecular basis of metabolites
   difference and the regulatory mechanisms between wild and cultivated
   plants, including Ophiocordyceps sinensis (O. sinensis) [[53]14],
   Indian jujube (Ziziphus mauritiana) [[54]15], pea (Pisum sp.) [[55]16],
   chili peppers (Capsicum spp.) [[56]17], Chinese chives [[57]18], peach
   [[58]19], potato [[59]20, [60]21], tomato [[61]22–[62]24], watermelon
   [[63]25], Astragalus mongholicus [[64]26], tea [[65]27], and citrus
   [[66]28]. These studies provide insights into the molecular basis of
   metabolic variations between wild and cultivated plants including many
   crops and medical plants, and provide useful information for the future
   cultivation of them. Dendrobium flexicaule (D. flexicaule) is
   endangered according to the International Union for Conservation of
   Nature and Natural Resources (IUCN) and is a first-class National Key
   Protected Wild Plant in China. However, no studies have examined the
   quality of D. flexicaule based on the content and composition of its
   natural metabolites.

   Here we firstly profiled metabolites from wild and cultivated D.
   flexicaule using ultra-performance liquid chromatography-mass
   spectrometry (UPLC-MS/MS) to explore the variations and
   interrelationships among bioactive components in this study.
   Comparative analysis revealed that the flavonoid and phenolic acid
   contents significantly increased, while amino acid and lipid from
   glycerolipid and glycerol-phospholipid contents largely decreased in
   cultivated D. flexicaule compared to those in wild D. flexicaule.
   Pharmacological analysis of these differentially abundant metabolites
   from wild and cultivated D. flexicaule indicated differences in
   phenylpropanoid biosynthesis, amino acid biosynthesis or metabolism and
   diverse pharmacological activities. Collectively, these results might
   help understanding the differences between wild and cultivated D.
   flexicaule and provide valuable information for their cultivation and
   utilization.

Results

Metabolomic determination of D. flexicaule by UPLC-MS/MS

   Stems at different growth stages of D. flexicaule were collected from
   the wild forest and greenhouse to reduce the variation among stem
   samples. The samples were further analyzed using a UPLC-MS/MS system in
   positive and negative ion modes. In total, 840 metabolites were
   identified (Fig. [67]1A). These metabolites were classified and mainly
   divided into 16 classes: flavonoids, phenols, lipids, nucleotides and
   derivatives, alkaloids and derivatives, amino acids and derivatives,
   steroids and steroid derivatives, organoheterocyclic compounds, organic
   acids and derivatives, carbohydrates and carbohydrate conjugates,
   benzenoids, and others (Fig. [68]1B). The three most represented
   classes of metabolites were lipids, flavonoids, carbohydrates and
   carbohydrate conjugates containing 177, 91, and 83 metabolites,
   respectively.

Fig. 1.

   [69]Fig. 1
   [70]Open in a new tab

   Classification statistics of the identified metabolites from D.
   flexicaule grown in the wild forest and greenhouse. (A) Statistical
   diagram of identified differentially abundant metabolites. (B)
   Classification pie diagram of the identified differentially abundant
   metabolites at the superclass level

Profile difference analysis of metabolites from wild and cultivated D.
flexicaule

   An unsupervised pattern recognition principal component analysis (PCA)
   was conducted to assess the effects of different plant origins on the
   contents and compositions of D. flexicaule metabolites. As a result,
   the two principal components (PC1 and PC2) explained 25.23% and 20.44%
   of the variation, respectively (Fig. [71]2A). The stem samples from
   wild forest grown D. flexicaule (SWDf) and greenhouse cultivated D.
   flexicaule (SGDf) were separated in the direction of PC2, however,
   there samples (SGDf3 ~ 5) partially overlapped with those from SWDf in
   the direction of PC1. Overall, it showed that the samples from wild and
   cultivated D. flexicaule were distinguishable, and the six biological
   replicates from each variety formed a cluster, indicating that the
   experiment was reproducible and reliable.

Fig. 2.

   [72]Fig. 2
   [73]Open in a new tab

   Multivariate analysis of metabolomics data of D. flexicaule. (A)
   Principal component analysis (PCA) for D. flexicaule samples grown in
   the wild forest (SWDf) and green house (SGDf). PC1 represents the first
   principal component and PC2 represents the second principal component.
   (B) Score plots of partial least squares discriminant analysis (PLS-DA)
   of SWDf and SGDf samples. R2 is the interpretation rate of the model
   and Q2 the predictive ability of the model. The absolute values of R2
   and Q2 are more close to 1.0, and the model is more reliable

   Further partial least squares discriminant analysis (PLS-OA) was
   applied to determine the variables that significantly differed among
   these samples. The PLS-DA score plots showed that the validation
   metrics R2 and Q2 values of the models of SGDf vs. SWDf were 0.91 and
   − 0.80, respectively, indicating that the models were of high fitness
   and predictability (Fig. [74]2B).

Differential metabolite screening and enrichment analysis from wild and
cultivated D. flexicaule

   Variable importance in projection (VIP) values (VIP ≥ 1) combined with
   fold change (FC) values (|log[2] (FC)| ≥ 1) were applied as the
   criteria to screen the representative variable metabolites from the two
   samples (SGDf vs. SWDf). As a result, significantly different
   metabolites were detected, as shown in volcano plots (Fig. [75]3A).
   Compared with the SWDf samples, the contents of 231 metabolites were
   significantly different in the SGDf samples, of which 86 and 145
   metabolites were increased and decreased, respectively (Table [76]S1).
   Further annotation and comparison were conducted for 231 differentially
   abundant metabolites present in the two groups. The results revealed
   that the presence of 17 flavonoids, 23 lipids, 17 phenols, 10 alkaloids
   and their derivatives, 20 amino acids and their derivatives, and 11
   nucleotides and their derivatives (Fig. [77]3B, C).

Fig. 3.

   [78]Fig. 3
   [79]Open in a new tab

   Overview of metabolites analysis detected in wild and cultivated D.
   flexicaule. (A) Volcano maps of differentially abundant metabolites
   from wild (SWDf) and cultivated (SGDf) D. flexicaule. The numbers of
   identified differential metabolites were indicated. (B) Comparison of
   the numbers of differentially abundant metabolites from different
   classes. (C) KEGG enrichment maps of differentially abundant
   metabolites from SWDf and SGDf samples

   Differentially abundant metabolites from the two groups were further
   imported into MetaboAnalyst 5.0 ([80]https://www.metaboanalyst.ca/) for
   Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathway
   enrichment analysis. The enrichment bubble plots of the top 20 enriched
   metabolic pathways are shown in Fig. [81]3D. Among these pathways,
   amino acid biosynthesis or metabolism, lipid metabolism, carbon and
   nitrogen metabolism, and phenylpropanoid biosynthesis were the most
   enriched.

Biological function profiling of differentially abundant metabolites in D.
flexicaule

   We performed hierarchical clustering of classes using the expression
   levels of significantly differentially abundant metabolites to explore
   more information on the identified metabolites from wild and cultivated
   D. flexicaule. Generally, the contents of many amino acids and lipids
   were significantly altered in the SGDf compared with those in the SWDf.
   Among them, the content of 17 amino acids and their derivatives
   decreased, except for L-serine, L-cysteine, and threonine
   (Fig. [82]4A). In addition, the levels of 11 of the 23 lipids, mostly
   glycerolipids (Lysophosphatidylethanolamine (LysoPE) 14:0, LysoPE 16:0,
   LysoPE 18:0, LysoPE 18:1,
   1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (LysoPC) 15:1 and
   LysoPC 16:0) and glycerol-phospholipids (PC 16:1/14:1, 13-HOTrE and MAG
   (18:3)), were much lower in SGDf than in SWDf (Fig. [83]4B). In
   contrast, certain flavonoids and phenols were highly accumulated in
   SGDf (Fig. [84]4C). No differences were detected in the levels of
   representative Dendrobium alkaloids (data not shown). However, ten
   annotated alkaloids, including epigotrin (a constituent of the
   traditional Chinese herbal medicine Isatis indigotica), topotecan,
   vinorelbine tartrate, and N’-nitrosoanabasine (Fig. [85]4D), were
   identified, four of which were more abundant in SGDf than in SWDf.

Fig. 4.

   [86]Fig. 4
   [87]Open in a new tab

   Heatmaps analysis of the representative differentially abundant
   metabolites from wild (SWDf) and cultivated (SGDf) D. flexicaule: (A)
   amino acids and derivatives, (B) glycerolipids and
   glycerol-phospholipids, (C) components of flavonoids (circle), phenols
   (square), terpenoids (triangle) and alkaloids (star). Different colors
   in the heat map represent the relative content of the differential
   metabolite. The value obtained after normalization reflects the
   relative content (red represents high content and blue represents low
   content). (D) Quantification of the representative differentially
   abundant metabolites from (C)

   The contents and types of total flavonoids that identified in SGDf and
   SWDf were further investigated. The result showed that total of 47
   flavonoids were identified (Table S2) and the flavonoids content of
   SGDf was significantly higher than that of SWDf (Fig. [88]5A). These
   flavonoids were composed of 5 flavanols, 8 flavanones, 11 flavones, 16
   flavonols, and 7 isoflavones (Fig. [89]5B). In consistent, the contents
   for most types of flavonoids were higher in SGDf than that in SWDf
   except for flavanones (Fig. [90]5C). Among these flavonoids, two
   flavonols of kaempferol and quercetin were significantly accumulated in
   SGDf (Fig. [91]5D). In contrast, the precursors of kaempferol including
   dihydrokaempferol and naringenin from flavonol biosynthesis pathway
   were obviously lower in SWDf than that in SGDf (Fig. [92]5E). These
   results indicated the difference of metabolic flux between SGDf and
   SWDf.

Fig. 5.

   [93]Fig. 5
   [94]Open in a new tab

   Comparison analysis of total flavonoids detected in wild and cultivated
   D. flexicaule. (A) Quantitative analysis of the total flavonoids
   content in stems of wild and cultivated D. flexicaule. Data bars
   represent the mean ± standard deviation (SD) of six biological
   replicates (n = 6). Different letters indicate significant differences
   (p < 0.05) according to Student’s T test. (B) Heatmaps analysis of
   different types of flavonoids detected in wild and cultivated D.
   flexicaule. The value obtained after normalization reflects the
   relative content (red represents high content and blue represents low
   content). (C-D) Quantitative analysis of the different types of
   flavonoids (C) and representative ones (D) detected in wild and
   cultivated D. flexicaule. (E) The proposed pathway associated with
   flavonoid biosynthesis in Dendrobium. Key enzyme gene abbreviation:
   CHS, chalcone synthase; CHI, chalcone isomerase; FNS, flavone synthase;
   FLS, flavonol synthase; F3H, flavonoid 3-hydroxylase; F3’H, flavonoid
   3’-hydroxylase. The FLS is highlighted in responsible for conversion
   from dihydrokaempferol to kaempferol

   In addition to universal metabolites, some phytohormones and their
   derivatives have also been identified, such as abscisic acid (ABA),
   salicylic acid (SA) O-glucoside, zeatin and its derivatives, and the
   precursor of ethylene, i.e., 1-aminocyclohexane carboxylic acid (ACC).
   Compared with the SWDf samples, the levels of ABA and SA-O-glucoside
   were significantly higher, whereas the content of zeatin and its
   derivatives, and ACC were significantly lower in the SGDf samples
   (Fig. [95]6A).

Fig. 6.

   [96]Fig. 6
   [97]Open in a new tab

   Differentially abundant metabolites correlation analysis between the
   representative metabolites and phytohormones. (A) Heatmap analysis of
   phytohormones from wild and cultivated D. flexicaule. Different colors
   in the heat map represent the relative content of the differential
   metabolite. The value obtained after normalization reflects the
   relative content (red represents high content and blue represents low
   content). (B-E) Correlation between alkaloids (B), terpenoids (C),
   flavones (D), phenols (E), and phytohormones, respectively. Metabolites
   closely related to different hormones were indicated with square (ABA),
   triangle (SA), and cross (ACC), respectively. Different colors in the
   charts represent the relative intensity of correlation (red represents
   positive correlation and blue represents negative correlation)

   Furthermore, differentially abundant metabolite correlation analysis
   was conducted to analyze the possible correlation between phytohormones
   and accumulation of active components in D. flexicaule. The results
   revealed that seven alkaloids, four terpenoids, five flavonoids and
   nine phenols were significantly correlated with ABA, as shown in the
   correlation heatmaps (Fig. [98]6B-E). In contrast, only six
   differentially abundant metabolites from the four groups were closely
   related to SA, and all were more abundant in the SGDf than in the SWDf.
   Notably, only one of the alkaloids, topotecan, showed a significant
   positive correlation with the same trend of change for both ABA and SA
   (Fig. [99]6B). Additionally, six differentially abundant metabolites
   from each group were closely related to ACC, and all were also tightly
   correlated with zeatins, implying the correlation among the
   phytohormones and other differentially abundant metabolites.

Network pharmacological analysis of differentially abundant metabolites
between wild and cultivated D. flexicaule

   Differentially abundant metabolites were further analyzed using network
   pharmacology to explore the differences in the pharmacological activity
   of wild and cultivated D. flexicaule. A total of 78 differentially
   abundant metabolites (45 decreased and 33 increased in SGDf vs. SWDf)
   with potential drug properties were screened (Fig. [100]7A), and 863
   gene targets were matched using public databases (SwissTargetPrediction
   and TCMSP: Traditional Chinese Medicine Systems Pharmacology Database
   and Analysis Platform) (Table S3). The results showed that 398 genes
   were common targets of all metabolites, whereas 263 and 202 genes were
   specifically targeted by increased and decreased differentially
   abundant metabolites, respectively, in the SGDf and SWDf comparisons
   (Fig. [101]7B). A compound-target network was then constructed using
   Cytoscape ([102]https://cytoscape.org). This network contained 938
   nodes and 2840 edges, suggesting that the physiological activity of D.
   flexicaule resulted from the cooperative action of numerous active
   components and targets (Fig. [103]7B).

Fig. 7.

   [104]Fig. 7
   [105]Open in a new tab

   Network pharmacological analysis of differentially abundant metabolites
   from wild (SWDf) and cultivated (SGDf) D. flexicaule. (A) Heatmap
   analysis of the potential medical metabolites from wild and cultivated
   D. flexicaule. (B) Chemical composition-target network diagram. The
   numbers outside and inside the cycles indicate for the number of genes
   and differential metabolites, respectively. (C-D) Connection network of
   differentially abundant metabolites and related genes. Red triangles
   and green ellipses indicate the increased and decreased potential
   medical metabolites from SGDf compared to SWDf, respectively. Gene
   targets are indicated with rectangles indicate the key targets with
   component ID of Uniprot. (E-F) KEGG enrichment analysis for the
   increased (E) and decreased (F) potential medical metabolites. (G-H)
   Disease enrichment bubble chart of the increased (G) and decreased (H)
   potential medical metabolites. The size of dots represents the number
   of metabolites enriched in the pathways. The color of the dot
   represents the p-value, and the deeper the red of the dot, the stronger
   the enrichment effects

   After further analysis of this network using CytoNCA, the degree,
   betweenness, closeness, eigenvector, local average connectivity-based
   method, and network of nodes were obtained. As a result, 39 and 36 key
   genes were selected from the targets of 31 upregulated and
   downregulated 28 metabolites, respectively, in SGDf (Fig. [106]7C, D).
   KEGG enrichment analysis indicated that the increased differentially
   abundant metabolites in the SGDf group were related to the
   renin-angiotensin system, lysine degradation, and thyroid and prostate
   cancer (Fig. [107]7E), whereas the decreased differentially abundant
   metabolites were associated with amino acid metabolism, starch and
   sucrose metabolism, galactose metabolism and other glycan degradation
   processes (Fig. [108]7F).

   Additionally, the disease enrichment analysis of these potential
   targets from increased differentially abundant metabolites in SGDf
   showed that the top disease categories were mainly associated with
   respiratory diseases (pulmonary edema and embolism and pneumonia) and
   cardiovascular diseases (ischemia, coronary artery disease, and
   thrombosis) (Fig. [109]7G). In contrast, the decreased levels of
   differentially abundant metabolites in the SGDf group were mainly
   related to neurological diseases (GM2 gangliosidosis, mental
   depression, sphingolipidosis, alcohol dependence and use disorders)
   (Fig. [110]7H). This suggested diverse physiological activities in wild
   and cultivated D. flexicaule.

Discussion

   Dendrobium spp. are a well-known source of traditional Chinese medicine
   and nourishing foods. However, most Dendrobium spp. on the market are
   artificially grown, resulting in variable quality. Effective natural
   metabolites are generally used to assess the quality of Dendrobium. A
   comparative analysis revealed significant differences in the content of
   many ingredients (polysaccharides, flavonoids, nucleosides, bibenzyls,
   lignans and volatile compounds) wild and cultivated D. huoshanense
   [[111]29]. Thus, analyzing the full metabolite profiles of distinct
   Dendrobium spp. is critical to comprehensively evaluate their quality,
   which may be used to guide medicinal plants. Hence, we conducted the
   comparative metabolomics analysis to explore the metabolites and assess
   the quality of wild and cultivated D. flexicaule based on its natural
   metabolite content.

   Plants produce a wide range of specialized metabolites for optimal
   growth and development in response to constantly shifting environmental
   conditions [[112]30]. The composition and content of plant metabolites
   appear to be regulated by the growth environment [[113]31]. In this
   study, the two cultivation regions (wild-forest and greenhouse) had
   distinct physical properties (e.g., water content and light intensity),
   which may have had significantly affected D. flexicaule growth and
   metabolism. As expected, the clear separation between the two groups in
   the PCA and PLS-DA data revealed that the metabolism of D. flexicaule
   obtained in the two growth regions differed significantly
   (Fig. [114]2). However, no substantial change in the accumulation of
   polysaccharides was detected in D. flexicaule grown in the greenhouse
   (data not shown). Previous studies have revealed that HL or UV-B
   treatment stimulates the accumulation of polysaccharides [[115]7,
   [116]8], whereas red-blue LED light does not affect the polysaccharide
   content of D. officinale seedlings [[117]10]. These findings implied
   that specific light or light intense is necessary for polysaccharide
   accumulation.

   Conversely, the levels of metabolites such as flavonoids, lipids, and
   amino acids, in Dendrobium are frequently altered when growth
   conditions change. For example, flavonoids are the most variable
   metabolites in D. officinale grown on various substrates [[118]7].
   Another comparative investigation revealed that the metabolite
   compositions of D. officinale samples from different plant origins were
   similar; however, their contents were significantly different, and
   distinct metabolites were primarily involved in flavone and flavonol
   production [[119]32].

   In this study, D. flexicaule grown in greenhouse were found enriched of
   the total content and most types of flavonoids expect for flavanoes
   when compared with that of wild-forest-grown D. flexicaule. The enzyme
   flavonol synthase (FLS) has been proved in controlling the metabolic
   balance through conversion of dihydroflavonol from the anthocyanin
   pathway into flavonols [[120]33, [121]34]. And transgenic plants
   expressing the FLS1 homolog from D. officinale DoFLS1 exhibited
   increase in flavonol content and decrease in anthocyanin content
   compare to wild-type plants [[122]35]. Moreover, three FLS genes were
   found involved in flavonol accumulation, such as kaempferol,
   eriodictyol, and quercetin [[123]36]. FLS homologs likely contribute to
   the biosynthesis of kaempferol likely to occur through
   dihydrokaempferol. Thus, higher accumulation of kaempferol in SGDf is
   likely due to high levels of certain FLS homologs, which remains to be
   further investigated.

   Interestingly, the contents of phytohormones, including ABA, SA,
   zeatins, and ACC, were clearly different between wild and cultivated D.
   flexicaule in our study (Fig. [124]6). Plant hormones are small
   molecules derived from primary metabolites, and their homeostasis and
   signaling networks generally modulate the biosynthesis of secondary
   metabolites [[125]37, [126]38]. For instance, the plant hormone
   jasmonate (JA) promotes the production of important secondary
   metabolites in some plants, including phenolic acid biosynthesis in
   Salvia miltiorrhiza [[127]39], phenolic compound biosynthesis in tomato
   [[128]40], terpene biosynthesis in strawberry [[129]41], and the
   accumulation of alkaloids in D. officinale [[130]42]. Flavonoid
   accumulation is also associated with brassinosteroid (BR) and auxin
   metabolism in D. officinale [[131]43]. ABA is an important phytohormone
   that induces drought resistance and influences the expression of genes
   involved in flavonoid production during drought stress [[132]44,
   [133]45]. Similarly, exogenous SA treatment significantly increased
   total flavonoid accumulation [[134]46–[135]48]. Considering the
   positive correlation between flavonoids, ABA, and SA in D. flexicaule,
   we propose that the accumulation of these specific flavonoids results
   from increased ABA or SA.

   Additionally, exogenous ABA significantly alters the abundance of most
   lipid metabolites in tea [[136]44]. Consistently, we found that the
   contents of several glycerophospholipids, glycerolipids, and fatty
   acids were significantly decreased in cultivated D. flexicaule.
   Previous studies have shown that the carbon flux is redirected toward
   flavonoid metabolism when lipid synthesis is inhibited [[137]49].
   Therefore, elevated ABA levels may significantly reduce the content of
   lipid metabolites while increasing the abundance of flavonoids. We also
   identified 28 metabolites with potential drug properties significantly
   positively correlated with ABA or SA (Fig. [138]S1). Most metabolites
   were specifically associated with ABA and SA. These findings suggested
   that ABA and SA contribute differently to the accumulation of various
   metabolites. Hence, it is possible to apply ABA or SA to promote the
   production of specific important metabolites in cultivated Dendrobium
   spp.

   The composition and content of metabolites in plants directly affect
   their quality and activity [[139]50]. The concentration of umami
   compounds in tomato differs determines the nutraceutical feature, while
   modern commercial tomato varieties are substantially less flavorful
   with significantly lower amounts of many of these important flavor
   chemicals than older varieties [[140]23, [141]24]. Comparative analysis
   of phenolic compound and flavonoids accumulation uncovers the unique
   flavor and quality from different tea cultivars and citrus species,
   respectively [[142]27, [143]28]. Comparative metabolite profiling the
   content of major bioactive compounds showed total different outcomes
   between wild and cultivated licorice, a famous Chinese medicinal
   material [[144]51]. Notably, D. flexicaule contains many metabolites
   that are medicinal compounds. Thus, it is vital to study the
   fluctuations in the metabolites and pharmacological activities of
   medicinal plants in response to changes in their growing environment.
   Pharmacological activity analysis of D. flexicaule revealed that wild
   plants might contribute more to neurological diseases, whereas
   cultivated plants likely increase the capacity for resistance to
   respiratory and cardiovascular diseases (Fig. [145]7E-H).

   Collectively, these findings suggested that the diverse metabolites of
   D. flexicaule from different origins are implicated in various diseases
   and that the influence of herbal origin should be considered more
   carefully when treating special types of diseases in clinical practice.
   We believe that the cultivation approach for Dendrobium spp. is
   crucial. Furthermore, we propose that the quality criteria of
   Dendrobium spp. include not only the polysaccharide content but also
   the content of other active compounds. Overall, our study sheds new
   light on how metabolites are altered in wild and cultivated D.
   flexicaule. This study also contributes to the establishment of quality
   control standards and serves as a reference for the efficient use of
   raw resources for development and utilization.

Conclusion

   In conclusion, our study sheds light on the metabolite profiles of wild
   and cultivated D. flexicaule. We observed substantial differences in
   the abundance of metabolites in the flavone, amino acid, and lipid
   groups. Importantly, these metabolites were strongly associated with
   the dynamics of numerous phytohormones. Our findings highlight the
   importance of hormones in the metabolism of natural compounds. We also
   demonstrated the diverse physiological functions of wild and
   domesticated D. flexicaule. Thus, we emphasize the necessity of
   cultivating Dendrobium growth conditions for quality control while also
   providing guidance on raw resource utilization.

Materials and methods

Plant materials and metabolite extraction

   Cultivated D. flexicaule samples were collected from an opensided
   greenhouse under the natural conditions (temperature range of 18 ° to
   32 °C, 13–14 h of natural daylight) in the National Orchid Conservation
   Center of China (NOCCC) in Guangdong Province, and wild D. flexicaule
   samples grown on the tree were collected from the Shennongjia Nature
   Reserve in Hubei Province, China.

   The method for metabolite extraction was modified from previous study
   [[146]52]. Generally, the collected stems of each sample were ground
   with liquid nitrogen, and the homogenate was resuspended in prechilled
   80% methanol and 0.1% formic acid by vortexing. The samples were
   incubated on ice for 5 min and then sonicated at 40 Hz for 20 min. The
   samples were subsequently centrifuged at 15,000 × g and 4 °C for
   20 min. Finally, the supernatant was injected into the UPLC- MS/MS
   system (Thermo Fisher, Germany) for analysis.

UPLC-MS/MS analysis

   UPLC-MS/MS analyses were performed using a Vanquish UHPLC system
   coupled with an Orbitrap Q ExactiveTM HF mass spectrometer (Thermo
   Fisher, Germany) at Novogene Bioinformatics Technology Co., Ltd.
   (Beijing, China). The samples were injected onto a Hypesil Gold column
   (100 × 2.1 mm, 1.9 μm) using a 17-min linear gradient at a flow rate of
   0.2 mL/min. The eluents for the positive polarity mode were eluent A
   (0.1% formic acid in water) and eluent B (methanol). The eluents for
   the negative polarity mode were eluent A (5 mM ammonium acetate, pH
   9.0) and eluent B (methanol). The solvent gradient was set as follows:
   1.5 min, 2% B; 12.0 min, 2-100% B; 14.0 min, 100% B; 14.1 min, 100-2%
   B; 17 min, 2% B. A QExactiveTM HF mass spectrometer was operated in
   positive/negative polarity mode with a spray voltage of 3.2 kV,
   capillary temperature of 320 °C, sheath gas flow rate of 40 arb and aux
   gas flow rate of 10 arb.

Data processing and metabolite identification

   The raw data files generated by UPLC-MS/MS were processed using
   Compound Discoverer 3.1 (CD3.1, Thermo Fisher) to perform peak
   alignment, peak picking, and quantitation for each metabolite. The main
   parameters were set as follows: retention time tolerance, 0.2 min;
   actual mass tolerance, 5 ppm; signal intensity tolerance, 30%;
   signal/noise ratio, 3; and minimum intensity, 100,000. After that, the
   peak intensities were normalized to the total spectral intensity. The
   normalized data were used to predict the molecular formula based on
   additive ions, molecular ion peaks and fragment ions. Then, the peaks
   were matched with the mzCloud ([147]https://www.mzcloud.org/), mzVault
   and MassList databases to obtain accurate qualitative and relative
   quantitative results.

Data analysis

   These metabolites were annotated using the KEGG database
   ([148]https://www.genome.jp/kegg/pathway.html), HMDB
   ([149]https://hmdb.ca/metabolites) and LIPIDMaps database
   ([150]http://www.lipidmaps.org/). Principal component analysis (PCA)
   and partial least squares discriminant analysis (PLS-DA) were performed
   using the free online platform OmicShare developed by Gene Denovo
   ([151]https://www.omicshare.com) [[152]53]. Statistical significance (P
   value) was calculated with univariate analysis (t test). The
   metabolites with VIP > 1 and P value < 0.05 and fold change (FC) ≥ 2 or
   FC ≤ 0.5 were considered to be differentially abundant metabolites.
   Volcano plots were used to filter metabolites of interest based on the
   log2(FC) and -log10(p value) values of the metabolites. For clustering
   heatmaps, the data were normalized using z scores of the intensity
   areas of differentially abundant metabolites and were plotted by
   TBTools II [[153]54]. The functions of these metabolites and metabolic
   pathways were studied using the KEGG database. The metabolic pathway
   enrichment of differentially abundant metabolites was performed based
   on the online OmicShare website.

Mining for potential targets of differentially abundant metabolites

   The Swiss Target Prediction database [[154]55] and Traditional Chinese
   Medicine Systems Pharmacology Database and Analysis Platform (TCMSP)
   [[155]56] were used to predict the potential targets of the metabolites
   in Dendrobium. The targets with probability values greater than 0 and
   the biological sources of the targets were set to “Homo sapiens” in the
   Swiss Target Prediction database. According to the TCMSP database, the
   oral bioavailability (OB) and drug-like properties (DL) were set to
   over 30% and 0.18%, respectively. All targets were normalized to the
   UniProt database ([156]https://www.uniprot.org), and duplicate values
   were removed. KEGG pathway and disease enrichment analyses were
   performed via the Search Tool for the Retrieval of Interacting
   Genes/Proteins (STRING) database [[157]57], with a false discovery rate
   (FDR) < 0.05 as the screening criterion, and the genes were sorted by
   their false discovery rate (FDR) values from smallest to largest.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [158]Supplementary Material 1^ (6.7MB, tif)
   [159]Supplementary Material 2^ (26.3KB, xlsx)
   [160]Supplementary Material 3^ (78.7KB, xlsx)
   [161]Supplementary Material 4^ (19.2KB, xlsx)

Acknowledgements