Integrated Metabolomic and Transcriptomic Analysis of Phenylpropanoid Biosynthesis in Silphium perfoliatum

Abstract

Silphium perfoliatum is a promising economic plant rich in bioactive secondary metabolites, yet the molecular regulation of phenylpropanoid biosynthesis across development remains unclear. To elucidate the regulatory networks underlying these metabolic processes, we integrated metabolomic and transcriptomic analyses across six developmental stages, from cotyledon to flowering. LC–MS/MS identified 1964 metabolites, with phenylpropanoids representing the largest class (601 compounds). Differential accumulation analysis showed pronounced temporal dynamics in phenylpropanoid levels, especially chlorogenic acid and its derivatives, with many compounds peaking at the flowering stage. In parallel, RNA-seq revealed 31,624 differentially expressed genes (DEGs). Functional enrichment highlighted phenylpropanoid and flavonoid biosynthetic pathways as major metabolic hubs. Correlation analysis indicated that PAL, 4CL, HCT, F3H, FLS, and F3′H expression was tightly coordinated with the accumulation of phenolic acids and flavonoids, suggesting these gene encoded enzymes may represent rate-limiting steps. Furthermore, weighted gene co-expression network analysis (WGCNA) identified a “blue” module strongly associated with phenylpropanoid accumulation and significantly enriched in pathway-related genes. Together, these results provide a comprehensive regulatory framework for phenylpropanoid biosynthesis in S. perfoliatum and offer valuable genetic targets for metabolic engineering and molecular breeding to enhance bioactive compound production.

1. Introduction

Silphium perfoliatum L., commonly known as cup plant, is a perennial herb of the Asteraceae family native to North America. Historically a prairie species characterized by a robust fibrous root system, stand longevity, and high aboveground biomass, it has recently gained considerable attention as a multifunctional crop in temperate regions [1]. In Europe, where it was introduced in recent decades, S. perfoliatum is increasingly regarded as a sustainable alternative to conventional annual crops—such as silage maize—particularly on marginal or degraded lands, due to its marked tolerance to drought and frost and modest input requirements [2]. While primarily studied as a competitive feedstock for biogas and solid biofuel production, its utility extends to forage, ornamental applications, and ecological services. Its prolonged flowering period and structurally complex canopy support diverse pollinator and arthropod communities, thereby enhancing biodiversity within intensive agricultural landscapes [3,4,5]. Beyond these agronomic and ecological attributes, there is growing interest in the pharmacological potential of S. perfoliatum. The plant is rich in bioactive secondary metabolites, including phenolic acids [6], flavonoids [7], terpenoids [8], and polysaccharides [9], which have demonstrated hepatoprotective [10,11], hypoglycemic [12], antibacterial [13], and anti-inflammatory activities [14]. Despite these promising traits, the domestication status of S. perfoliatum remains rudimentary; its genetic base is narrow, the mechanisms underlying its environmental adaptability are poorly understood, and the biosynthetic pathways of its key active ingredients have not yet been detailed.

Among the diverse metabolic processes in plants, phenylpropanoid metabolism stands out as a critical pathway for secondary metabolite synthesis [15]. This pathway is responsible for producing compounds such as lignin, flavonoids, anthocyanins, and organic acids, which are essential for regulating adaptive growth [16]. In medicinal plants, phenylpropanoid metabolism is particularly significant, serving as the origin for a vast array of pharmacologically active constituents with a phenylpropanoid backbone, including flavonoids and various phenolic compounds [17]. Furthermore, specific metabolites derived from this pathway can be secreted as root exudates into the rhizosphere, where they modulate the microbial community structure, thereby influencing plant growth and resistance to biotic and abiotic stresses [15]. In S. perfoliatum, numerous phenylpropanoid derivatives—such as chlorogenic acid analogues and flavonoids—have been isolated [12]. However, current research has predominantly focused on chemical separation and pharmacological screening, while studies on the genetic basis of phenylpropanoid biosynthesis and the regulatory networks governing their accumulation remain scarce.

Elucidating the molecular mechanisms underlying metabolite synthesis requires navigating complex, multi-target, and interconnected networks. To address this challenge, integrated multi-omics approaches—combining transcriptomics, proteomics, and metabolomics—have become indispensable tools in plant research [18,19]. These technologies have successfully identified key candidate genes and mapped biosynthetic pathways in various medicinal species, such as Camellia huana [20], Carthamus tinctorius [21], Hedera helix [22], and Prunus serrulata [23] among others. Despite the widespread application of multi-omics in resolving plant metabolic pathways, the specific biosynthesis of phenylpropanoids in S. perfoliatum has not been systematically characterized.

Phenylpropanoid metabolism serves as the central conduit for the synthesis of bioactive constituents in S. perfoliatum. Deciphering the temporal dynamics of metabolite accumulation and their governing physiological mechanisms is essential for stabilizing crop quality. By coupling transcriptomic profiling with metabolomics, this study dissects the regulatory architecture of phenylpropanoid biosynthesis across distinct developmental stages. We identified high-confidence candidate genes and correlated them with specific metabolite profiles to map the biosynthetic network. These data provide a mechanistic blueprint for understanding secondary metabolism in S. perfoliatum, offering a theoretical framework to enhance the species’ functional utility.

2. Materials and Methods

2.1. Plant Materials

S. perfoliatum used in this study was cultivated at the Germplasm Bank of Qinghai-Tibetan Plateau in Xining, Qinghai Province, China (101°37′37″ E, 36°45′5″ N, Altitude 2280 m). Fresh leaves were collected at six stages: cotyledon stage (S1), seedling stage (S2), growing stage (S3), rosette stage (S4), bolting stage (S5) and flowering stage (S6) (Figure S1). Samples representing the complete developmental timeline were harvested in biological triplicate for integrated metabolomic and transcriptomic analysis. Immediately following excision, tissues were snap-frozen in liquid nitrogen and stored at −80 °C to preserve RNA and metabolite integrity prior to extraction.

2.2. Metabolites Analysis

S. perfoliatum samples (100 mg ± 1 mg) were weighed into microcentrifuge tubes, followed by the addition of homogenization beads and 1000 μL of extraction solvent composed of methanol, acetonitrile, and water (2:2:1, v/v/v). The mixture was vortexed for 30 s, followed by three sequential cycles of homogenization (35 Hz, 4 min) and sonication in an ice-water bath (5 min). A 400 μL aliquot was then passed through a 0.22 μm filter plate, and the resulting filtrate was retained for instrumental analysis.

Chromatographic analysis was performed on a Vanquish Horizon UHPLC (Thermo Fisher Scientific, Waltham, MA, USA). Kinetex C18 column (2.1 mm × 50 mm, 2.6 μm; Phenomenex Inc., Guangzhou, China) was used with 0.01% acetic acid and isopropanol:acetonitrile (1:1, v/v) as mobile phase. MS/MS data acquisition was conducted on an Orbitrap Exploris 120 mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) with the following parameters: sheath and aux gas flow rate setting at 50 and 15 Arb; capillary temperature maintained at 320 °C; spray voltage setting at 3.8 kV in positive mode and −3.4 kV in negative mode. BiotreeDB (V3.0) and BT-Plant (V1.1) databases were used for metabolite identification. Differential Accumulation Metabolites (DAMs) were determined using rigorous statistical criteria: Variable Importance in Projection (VIP)> 1 combined with a p_adj < 0.05 and |log2 (FoldChange)| > 1.

2.3. RNA Sequencing and Data Analysis

Total RNA was extracted from leaf tissues using the TRNzol Total RNA Extraction Kit (Tiangen Biotech Co., Ltd., Beijing, China). cDNA library was constructed and sequenced on the BGI DNBSEQ-T7 high-throughput sequencing platform. Following next-generation sequencing, raw data were filtered to obtain clean reads for comparative transcriptomic analysis.

Functional annotation was performed using DIAMOND (v2.1.22) against the NR (Non-Redundant Protein Sequence Database), KOG/COG (Cluster of Orthologous Groups of Protein/Eukaryotic Orthologous Groups), KEGG (Kyoto Encyclopedia of Genes and Genomes), and SwissProt databases. The NT (Nucleotide Sequence Database) was annotated using Blastn (v2.14). Protein domains were identified using HMMER3 against the Pfam database. GO (Gene Ontology) annotation was conducted using Blast2GO (v6.0) based on NR annotation results. Transcript annotations shared across these seven databases were selected.

Read count data were normalized (using DESeq2, v1.50), followed by p-value calculation and False Discovery Rate (FDR) correction. Differentially Expressed Genes (DEGs) were filtered based on |log₂(FoldChange)| > 1 and p_adj < 0.05. GO enrichment analysis of DEGs was performed using GOseq, and KEGG pathway enrichment was analyzed using KOBAS (v3.0).

2.4. Quantitative Real-Time PCR

Ten DEGs were selected for validation, with the Actin 7 gene (ACT7) serving as the internal reference. Primers were designed by NCBI primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/ (accessed on 5 December 2025)) and synthesized by Sangon Biotech (Shanghai) Co., Ltd., Shanghai, China. The primer sequences were shown in Supplementary Table S1. Quantitative real-time PCR (qRT-PCR) reactions were conducted on a LightCycler 480 System (F. Hoffmann-La Roche Ltd., Grenzach-Wyhlen, Germany) with TB Green Premix Ex Taq II (Tli RNaseH Plus) kit (Takara Bio, Kusatsu, Shiga, Japan). Relative expression levels were calculated using the 2^−ΔΔCt method.

3. Results

3.1. Metabolites Profiling of S. perfoliatum in Six Developmental Stages

The temporal trajectory of metabolic alterations of S. perfoliatum across six distinct developmental phases was analyzed by LC-MS/MS. This comprehensive analysis yielded the identification of 1964 unique metabolites throughout the six investigated stages. The detected compounds spanned several major biochemical categories, including amino acids and their derivatives, fatty acids, carbohydrates, terpenoids, alkaloids and notably, phenylpropanoids. The phenylpropanoids class constituted the largest identified group, accounting for 601 individual compounds (Table S2).

To assess the robustness and quality of the generated data, principal component analysis (PCA) was performed. The PCA scatter plot demonstrated that the first two components (PC1 and PC2) accounted for 24.7% and 16.9% of the total metabolic variance, respectively. A vital outcome of the PCA was the clear segregation of the six sampling groups, exhibiting minimal intra-group variability (Figure 1A). This distinct separation confirms substantial metabolic differentiation correlating with the floral developmental timeline, thereby validating the dataset’s reliability for deeper investigation. Additionally, Hierarchical Clustering Analysis (HCA), represented in a corresponding heatmap, indicated that the most profound shifts in metabolite concentrations occurred between the S5 and S6 developmental stages (Figure 1B).

Comparative metabolomic profiling across the six growth stages yielded 1416 non-redundant DAMs. Terpenoids (464) and phenylpropanoids (458) represented the dominant chemical classes, followed by fatty acids (161) and alkaloids (152) (Table S3). Functional classification via KEGG enrichment analysis identified the top 20 metabolic routes associated with these developmental shifts (Figure S2). Core metabolic processes—specifically biosynthesis of various plant secondary metabolites (ko00999), and flavonoid biosynthesis (ko00941)—along with phenylpropanoid biosynthesis (ko00940), showed consistent enrichment across comparisons. Notably, pathways associated with phenylpropanoid biosynthesis were distinct features of the S6 versus S1 and S6 versus S2 contrasts.

Phenylpropanoids are both as the abundant metabolites and predominant DAMs in the six developmental stages of S. perfoliatum. To delineate the temporal variation of phenylpropanoids metabolites throughout the S. perfoliatum developmental process, the 601 annotated phenylpropanoids compounds were subjected to Fuzzy C-Means Clustering utilizing the R package Mfuzz (v2.66.0). This analysis partitioned the expression patterns of these metabolites into nine distinct clustering modules (Figure 2A). Notably, the metabolites assigned to Cluster 6 displayed a positive correlation with developmental stage, reaching peak abundance during the S6 stage. Across the 15 pairwise comparative groups, 31–252 differential accumulation phenylpropanoids (DAPs) were detected, respectively (Figure 2B). A progressive trend in differential regulation was observed: the count of significantly upregulated metabolites gradually increased with development, yielding the highest number in the S6 vs. S2 comparison and the lowest in S5 vs. S4. Conversely, the S2 vs. S1 comparison contained the maximum number of downregulated DAPs, while the S6 vs. S2 transition exhibited the minimum downregulation. In total, 458 non-redundant DAPs were identified across the six developmental stages, including 147 flavonoids, 65 phenylpropanoids (C6-C3), 58 coumarins, 44 phenolic acids (C6-C1), 43 lignans, and 30 isoflavonoids (Table S4).

3.2. Transcriptomic Analysis in Six Developmental Stages of S. perfoliatum

To investigate the dynamic transcriptomic reprogramming that occurs throughout the six distinct growth stages of S. perfoliatum, high-throughput RNA sequencing (RNA-seq) was conducted on samples representing each developmental phase. Across the six developmental stages, we identified a total of 31,624 non-redundant DEGs, indicating extensive transcriptional restructuring during ontogeny.

PCA was used to evaluate global expression patterns and revealed that the first two components (PC1 and PC2) accounted for 31.7% and 19.9% of the total variance, respectively, suggesting distinct transcriptomic signatures underlying stage-specific biological processes (Figure 3A). HCA visualized as a heatmap further demonstrated that the most dramatic shifts in either gene expression or metabolite-associated profiles occurred during the transition corresponding to the S4 stage, highlighting this as a potentially critical developmental transition point (Figure 3B).

Pairwise differential expression analysis between growth stages uncovered varying numbers of upregulated and downregulated genes (Figure 4A). For example, comparisons such as S4 versus S1 and S5 versus S2 exhibited particularly high DEGs counts, underscoring substantial regulatory changes between early and later phases. Across all stage comparisons, 1855 DEGs were shared universally, representing core genes whose expression fluctuates throughout development. Meanwhile, stage-specific DEGs sets were also observed: 1628, 978, 966, 918, 917, and 830 DEGs were uniquely associated with S4, S5, S2, S3, S6, and S1, respectively (Figure 4B). These unique and shared patterns reveal both conserved regulatory programs and stage-specific transcriptional features, reflecting the complex and dynamic nature of developmental gene expression landscapes.

Time-series transcriptome clustering revealed 12 distinct expression modules spanning the six developmental stages (S1–S6), indicating extensive stage-dependent transcriptional reprogramming (Figure 5). Most clusters (Cluster 1, 3, 4, 6, 8, 10, 12) displayed pronounced transient induction at intermediate stages, characterized by sharp peaks followed by rapid declines, suggesting activation of stage-specific regulatory programs during key developmental transitions. In contrast, Cluster 5 exhibited progressive upregulation toward late stages, consistent with the gradual establishment of mature tissue functions. Other modules show high expression in the early stage and then inhibition, which indicates a shift from the early developmental process to the later metabolic and structural pathways. Notably, clusters with abrupt changes around mid-to-late stages highlight potential transcriptional switches that may coincide with major physiological remodeling. Collectively, these clustered temporal patterns delineate coordinated gene sets that likely underpin developmental progression and provide candidate modules for downstream functional enrichment and regulatory network analyses.

To elucidate the functional implications of the transcriptomic variations, GO and KEGG pathway enrichment analyses were conducted on the identified DEGs. Across the 15 pairwise comparisons, DEGs were stratified into the three canonical GO domains. Within the Molecular Function (MF) domain, “catalytic activity” and “binding” emerged as the predominant subclasses, with protein binding (GO:0005515) exhibiting the highest degree of enrichment. In the Biological Process (BP) domain, which was primarily characterized by “metabolic process” and “cellular process,” the most significant terms included protein phosphorylation (GO:0006468), oxidation-reduction process (GO:0055114), and DNA-templated regulation of transcription (GO:0006355). Concurrently, the Cellular Component (CC) domain was defined largely by “cellular anatomical entity” and “protein-containing complex,” with specific enrichment concentrated in integral component of membrane (GO:0016021) and membrane (GO:0016020) (Figure 6A).

Subsequent KEGG pathway analysis stratified the top 50 enriched pathways into five primary functional hierarchies (Figure 6B). The metabolism cluster figured prominently, featuring significant enrichment in glycolysis/gluconeogenesis (K00850), starch and sucrose metabolism (K01194), photosynthesis (K02717), oxidative phosphorylation (K02138), and phenylpropanoid biosynthesis (ko00940). In the realm of environmental information processing, the plant hormone signal transduction (ko04075) and MAPK signaling (ko04016) pathways were notably upregulated. Furthermore, enrichment was observed in endocytosis (K12184) and motor proteins (K11498) under cellular processes, ribosome (K03115) under genetic information processing, and plant–pathogen interaction (K15397) under organismal systems. Notably, as phenylalanine acts as an essential precursor for phenylpropanoid synthesis—yielding downstream metabolites such as phenolic acid, lignins and flavonoids—the differential regulation of genes within the phenylpropanoid biosynthetic pathway highlights their potential influence on the developmental physiology of S. perfoliatum.

3.3. Expression of Genes Related to the Biosynthetic Pathway of Phenylpropanoids in S. perfoliatum

KEGG pathway enrichment analysis highlighted the “Phenylpropanoid biosynthesis” (map00940) and “Flavonoid biosynthesis” (map00941) pathways as focal points of metabolic activity (Figure 7A). A total of 20 DEGs associated with these pathways were identified, comprising transcripts encoding PAL (1 gene), 4CL (2 genes), HCT (1 gene), C3′H (2 genes), CCR (1 gene), CCoAOMT (2 genes), CHS (2 genes), CHI (1 gene), F3H (1 gene), FLS (2 genes), F3′H (2 genes), CAD (1 gene), and C4H (2 genes).

Transcriptional profiling across six developmental stages revealed distinct spatiotemporal patterns in gene expression relative to metabolite accumulation (Figure 7B). The majority of identified DEGs exhibited peak expression during the early developmental phases (S1 and S2), followed by a marked downregulation in subsequent stages. This expression profile suggests that the initial growth period is characterized by high metabolic flux and rapid metabolite accumulation. However, C3′H, CCR, and F3′H displayed divergent expression patterns compared to the general trend.

To elucidate the regulatory network, mantel tests were conducted, revealing significant correlations between the expression of these DEGs and the accumulation of representative phenolic acids and flavonoids. Specifically, the content of phenolic acids—including caffeic acid, chlorogenic acid, and coumaroylquinic acid—showed a strong positive correlation with PAL, 4CL, HCT, and CCR expression, suggesting these transcripts act as key determinants of phenolic acid variance across developmental stages. Concurrently, major flavonoids such as luteolin, kaempferol, rutin, and quercetin were highly correlated with CHI, F3H, FLS, and F3′H, emphasizing the critical role of these genes in flavonoid skeleton biosynthesis.

Notably, chlorogenic acid and its derivatives were identified as the predominant phenylpropanoids in S. perfoliatum. The biosynthesis of chlorogenic acid typically proceeds via the HCT-mediated conversion of p-coumaroyl-CoA to p-coumaroylquinic acid, which subsequently undergoes hydroxylation by C3′H to form chlorogenic acid. In this study, chlorogenic acid content was closely associated with the expression levels of HCT and C3′H. The expression of both HCT and C3′H increased rapidly beginning at the S2 stage, accompanied by a corresponding rapid accumulation of chlorogenic acid (Figure 7B,C). This evidence supports the hypothesis that HCT and C3′H function as key rate-limiting genes governing the differential accumulation of chlorogenic acid throughout plant development.

3.4. Gene Co-Expression Network Analysis

To systematically deconvolute the regulatory networks governing phenylpropanoid accumulation, we performed Weighted Gene Co-expression Network Analysis (WGCNA) utilizing the 14,683 genes retained after filtration. This systems biology approach constructs a scale-free network to identify clusters of highly correlated genes. We first calculated the topological overlap matrix (TOM) to measure the interconnectedness of gene pairs, which served as the basis for hierarchical clustering. In the resulting dendrogram, branches correspond to distinct co-expression modules (Figure 8A). Genes that did not show significant co-expression patterns were assigned to the “grey” module, representing background noise or unassigned transcripts.

The dynamic tree cut algorithm identified 21 distinct modules of varying sizes, ranging from the extensive “turquoise” module (4412 genes) to the compact “royalblue” module (39 genes). To link transcriptional patterns with metabolic phenotypes, we analyzed the correlation between module eigengenes (MEs)—the representative expression profiles for each module—and the abundance of 16 targeted phenylpropanoids (Figure 8B). This module–trait relationship analysis highlighted seven modules with significant biological relevance. Specifically, the “tan,” “cyan,” “pink,” and “blue” modules (along with the “grey” set) displayed strong positive correlations with the majority of phenylpropanoid metabolites, indicating their potential role in upregulating biosynthesis. Conversely, the “royalblue” and “turquoise” modules exhibited negative correlations. To further parse the biological function of these trait-associated clusters, we mapped the constituent genes to the KEGG database. Notably, enrichment analysis revealed that the “blue” module is significantly enriched in pathways critical to phenylpropanoid biosynthesis (map00940) (Figure S3), suggesting this module serves as a core regulatory hub coordinating the enzymatic machinery required for metabolite production.

3.5. Quantitative Real-Time PCR Validation

To independently evaluate the robustness of the RNA-seq–based expression estimates, ten DEGs annotated to the phenylpropanoid biosynthetic pathway were randomly selected for validation using RT–qPCR. RT–qPCR was performed on RNA extracted from the same six developmental stages analyzed by transcriptome sequencing. Across all tested targets, the RT–qPCR profiles largely recapitulated the RNA-seq-inferred expression patterns, demonstrating strong cross-platform concordance and supporting the reliability of the transcriptomic dataset (Figure S4).

4. Discussion

Phenylpropanoids arise predominantly from the shikimate pathway, which channels carbon from glycolysis and the pentose phosphate pathway into aromatic metabolism. As one of the most chemically diverse and functionally consequential classes of plant secondary metabolites, phenylpropanoids underpin key aspects of plant structural reinforcement and stress adaptation through multiple downstream branches, notably those leading to lignins and flavonoids [15]. Their structural repertoire spans simple hydroxycinnamates (e.g., cinnamic, p-coumaric, ferulic, caffeic, and sinapic acids), broader pools of phenolic acids (including hydroxycinnamic and hydroxybenzoic acids), and extensive flavonoid subclasses (flavones, flavonols, anthocyanins, and isoflavonoids), in addition to lignans, coumarins, and stilbenoids [17]. Functionally, phenylpropanoids act as potent antioxidants that mitigate oxidative damage from abiotic stress, while also reinforcing cell walls through lignification to withstand biotic attacks. Consequently, the activation of this pathway is a critical adaptive strategy for survival. Furthermore, research underscores that external factors significantly modulate phenylpropanoid accumulation. For instance, Sytar et al. [24,25] demonstrated that chemical treatments enhance antioxidant levels in buckwheat, while Li et al. [26] revealed that geographical origin dictates the phenolic profiles and anti-inflammatory activity of dandelions, highlighting the pivotal role of environmental conditions in regulating secondary metabolism.

Consistent with its economic and medicinal relevance, S. perfoliatum exhibits a substantial reservoir of these metabolites. Metabolomic profiling across six developmental stages has catalogued 603 phenylpropanoids, revealing pronounced ontogenic dynamics rather than a static chemical inventory. Cluster analyses delineated discrete temporal trajectories: 86 metabolites—dominated by phenolic acids (65.1%) and flavonoids (38.4%)—accumulated progressively with development and reached maximal abundance at stage S6, whereas 64 metabolites—again enriched in phenolic acids (43.8%) and flavonoids (35.9%)—declined over time and peaked at the juvenile S1 stage. Chemically, the phenolic-acid fraction is characterized by chlorogenic acid and closely related congeners, while flavonoid flux is directed largely toward quercetin, kaempferol, and luteolin, predominantly occurring as glycosides. The high abundance of these core constituents provides a plausible material basis for the reported bioactivity of S. perfoliatum; pharmacological evidence has linked such phenylpropanoid profiles to anti-inflammatory, antibacterial, and cholestatic (bile-regulatory) effects, supporting its application in phytomedicine [10,11,12,13].

In parallel with metabolite variation, transcriptomic analysis identified 20 DEGs associated with phenylpropanoid biosynthesis, including core pathway genes (PAL, C4H, 4CL, HCT, C3′H, CCR, and CCoAOMT) and downstream flavonoid genes (CHS, CHI, F3H, FLS, F3′H), as well as CAD and C4H. Most of these genes exhibited their highest expression at early developmental stages (S1–S2), consistent with coordinated transcriptional activation during juvenile growth. Mantel tests further supported gene–metabolite coupling, showing significant positive associations between phenolic-acid abundance and the expression of PAL, 4CL, HCT, and CCR. This pattern accords with prior observations that PAL expression tracks chlorogenic acid accumulation in species such as Nicotiana tabacum [27], Ipomoea batatas [28], and Solanum tuberosum [29], and that C4H and 4CL operate as regulatory nodes influencing phenolic acids and lignin in crops including lettuce, pear and tea [30,31,32]. Experimental evidence also suggests that 4CL overexpression, or compensatory 4CL induction following C4H suppression, can enhance chlorogenic acid biosynthesis [33]. HCT, which is required for shikimate- and quinate-ester formation and thus for flux redirection, similarly shows functional leverage: HCT silencing restricts growth and lignification in tobacco, whereas overexpression can increase chlorogenic acid substantially, in agreement with the strong correlations observed here [34,35]. For flavonoids, accumulation was positively correlated with CHI, F3H, FLS, and F3′H transcript levels [36]. CHS is widely regarded as a principal gatekeeper of flux into flavonoid metabolism and a determinant of intraspecific variation [32], while F3H catalyzes the conversion of naringenin to dihydrokaempferol, a step central to anthocyanin-associated pigmentation; gain- and loss-of-function studies across multiple taxa link F3H activity to increases in anthocyanins [37,38,39]. Structural diversification is further governed by FLS and F3′H: FLS channels intermediates toward flavonol formation, whereas F3′H introduces B-ring hydroxylation required for cyanidin and quercetin derivatives. Consistent with reports in Vitis vinifera [40] and Camellia sinensis [41], elevated FLS expression is often associated with enhanced flavonol accumulation and constrained anthocyanin synthesis, while genetic disruption of F3′H (e.g., via CRISPR-Cas9 or mutant analyses) reduces di-hydroxylated flavonoids and shifts metabolism toward mono-hydroxylated products such as pelargonidin or kaempferol derivatives, frequently diminishing total pigment levels [42,43,44,45]. Collectively, these observations support the hypothesis that the elevated expression of key structural genes contributes materially to phenylpropanoid biosynthesis and metabolite accumulation in S. perfoliatum; nevertheless, functional validation and elucidation of upstream regulatory mechanisms remain necessary.

5. Conclusions

The present study establishes a robust multi-omics framework, integrating high-throughput transcriptomic sequencing with metabolomic profiling, to dissect the complex genetic architecture underlying secondary metabolism in Silphium perfoliatum. By systematically correlating gene expression patterns with metabolite accumulation profiles, we have successfully reconstructed the regulatory network governing phenylpropanoid biosynthesis. This approach allowed for the precise identification of candidate structural genes and, crucially, the upstream transcription factors that orchestrate the flux of carbon through the shikimate and phenylpropanoid pathways. Collectively, these findings provide a foundational genetic blueprint, offering essential insights into the molecular mechanisms that drive both the synthesis of bioactive compounds and the species’ environmental adaptability.