3.1 Transcriptomic Identification of DEGs in the Comparisons between Different Tissuess of B. purpurasceus
Based on the transcriptomic analysis of root, stem and leaf of B. purpurasceus, a total of 65.66 Gb of clean reads were obtained from nine kernel samples; each sample contained ≥ 6 Gb of data with Q20 quality scores of ≥ 97.49% and Q30 quality scores of ≥ 93.03% (Table S2).
Under the parameters of “log2 fold change| ≥ 1” and “p ≤ 0.05”, 33994 DEGs (21416 and 12578 up-regulated and down-regulated, respectively) were identified by comparing root and leaf, 18132 DEGs (14252 and 3880 up-regulated and down-regulated, respectively) were identified by comparing root and stem, 24224 DEGs (14005 and 10219 up-regulated and down-regulated, respectively) were identified by comparing leaf and stem (Table S3 and Fig. 1A). Among them, there were 534 DEGs associated with secondary metabolism (Figure S2A). An overview of the expression profiles of all identified DEGs after filtration (|log2 fold change| ≥ 1, p<0.05, TPM ≥ 2) is shown in the heatmap in Fig. 1B. We compared the up-regulated and down-regulated DEGs from different comparisons using Venn diagram; including 12514 up-regulated and 1609 down-regulated DEGs in the root, 1782 up-regulated and 957 down-regulated DEGs in the stem, 10156 up-regulated and 6087 down-regulated DEGs in the leaf (Fig. 1C, D).
3.2 GO and KEGG Pathway Enrichment Analyses of DEGs in the Comparisons between Different Tissues of B. purpurasceus
To determine the function of the identified DEGs, we performed GO analysis, which shows GO term enrichment in three major functional categories: molecular function (MF), cellular component (CC), and biological process (BP) (Table S4_Sheet 1–3). The top 10 enriched GO terms from all DEGs across different comparisons are shown in Figure S3.
To further analyze the biological functions of DEGs, we performed KEGG enrichment analysis. The DEGs of root vs. stem comparison were mainly related to flavonoids biosynthesis, Pentose and glucuronate interconversions, Monoterpenoid biosynthesis and Phenylpropanoid biosynthesis (p<0.05) (Firure S4A; Table S5_Sheet1). The DEGs in the root vs. leaf comparison were mainly related to biosynthesis of cofactors, and others enriched were mostly related to nutrition metabolism, such as starch and sucrose metabolism and amino acid metabolism (p<0.05) (Firure S4B; Table S5_Sheet2). The DEGs in the leaf vs. stem comparison were mainly related to flavonoid biosynthesis, phenylpropanoid biosynthesis, glutathione metabolism, and other plant secondary metabolism (p<0.05) (Firure S4C; Table S5_Sheet3). A total of 139, 141 and 139 pathways were enriched in root vs. stem, root vs. leaf and leaf vs. stem comparisons. Flavonoid biosynthesis, Monoterpenoid biosynthesis and Phenylpropanoid biosynthesis pathways were enriched in root vs. stem and leaf vs. stem comparisons.
3.3 Metabolomic Identification of DAMs in the Comparisons between Different Tissues of B. purpurasceus
Based on the quasi-targeted metabolomics analysis of root, stem and leaf of B. purpurasceus. A total of 948 metabolites were identified, including amino acid and derivatives (180), flavonoids (161), carbohydrates and its derivatives (106), organic acid and its derivatives (97), nucleotide and its derivates (59), organoheterocyclic compounds (50), lipids (49), phenolic acids (45), henylpropanoids and polyketides (37), terpenoids (33), phenols and its derivatives (31), amines (22), alkaloids and derivatives (22), etc (Figure S5A). PLS-DA revealed significant biochemical changes in three groups (Figure S5B, C, D). This suggested that VIP analysis could be used to screen for DAMs.
Under the parameters of “VIP > 1, |log2 fold change| ≥ 1” and “p < 0.05”, a number of (DAMs) were identified in the different comparisons, including 125 up-regulated and 138 down-regulated DAMs in the root vs. stem comparison, 161 up-regulated and 206 down-regulated DAMs in the root vs. leaf comparison, 192 up-regulated and 189 down-regulated DAMs in the leaf vs. stem comparison (Fig. 2A, Table S6). Among them, there were 351 differentially accumulated secondary metabolites (Figure S2B). An overview of the metabolite profiles of the three tissues is shown in Fig. 2B. We compared the up-regulated and down-regulated DAMs from different comparisons using a Venn diagram; including 37 up-regulated and 17 down-regulated differentially accumulated secondary metabolites (DASMs) in the root, 31 up-regulated and 14 down-regulated DASMs in the stem, 53 up-regulated and 50 down-regulated DASMs in the leaf (Fig. 2C, D, Figure S6).
3.4 KEGG Pathway Enrichment of DAMs in the Comparisons between Different Tissues of B. purpurasceus
Based on KEGG annotation, the DAMs in different comparisons were enriched in many pathways, including biosynthesis of secondary metabolites, flavonoid biosynthesis, phenylpropanoid biosynthesis, and amino acid metabolism (Table S7). The top 20 KEGG terms are shown in Figure S7, of which the most significantly enriched terms in the root vs. stem comparison were flavonoid biosynthesis and phenylpropanoid biosynthesis (Figure S7A). In the root vs. leaf and leaf vs. stem, biosynthesis of secondary metabolites was the most enriched term (Figure S7B, C).
3.5 Weighed gene co-expression network analysis
In order to explore phenylpropanoid and flavonoid biosynthesis related genes, we performed WGCNA analysis on all DEGs and seven modules were identified (Figure S8A). As shown in Fig. 3A, the correlation analysis between modules and traits revealed that ME-A and ME-F modules demonstrated significant correlations with the biosynthesis of the majority of phenylpropanoid metabolites. The ME-A module (contains 334 genes) exhibited a positive correlation with a total of 25 metabolites, while the ME-F module (contains 920 genes) displayed a negative correlation with the same set of 25 metabolites. Among them, 746 genes exhibited significant correlations with coumarins (Fig. 3B; Table S8_Sheet3). 921 genes in ME-A and ME-F exhibited significant correlations with the synthesis of flavone and flavonol (Figure S9; Table S8_Sheet4).
KEGG analysis was carried out to annotate the gene function in these two modules (Figure S8B, C). Genes in ME-A module were mainly enriched in phenylpropanoid and flavonoid biosynthesis pathways. Genes in ME-F module were mainly enriched in flavonoid biosynthesis pathways. The structural and regulatory genes play important roles in phenylpropanoid and flavonoid synthesis (Fig. 6, 7). In the ME-A module, four structural genes (CCR, CAD, COMT, and CCoAOMT) involved in the phenylpropanoid biosynthesis pathway and one structural gene (PGT1) associated with the flavonoid biosynthesis pathway were identified. Conversely, the ME-F module exhibited the presence of two structural genes (PAL and 4CL) involved in the phenylpropanoid biosynthesis pathway, along with seven structural genes (CHS, CHI, F3H, DFR, LAR, ANS, and ANR) associated with the flavonoid biosynthesis pathway. The presence of numerous structural genes associated with phenylpropanoid and flavonoid biosynthesis in the two modules provided further evidence to suggest that the coexpression genes in these modules were involved in the accumulation of phenylpropanoids and flavonoids in B. purpurasceus. It was found that the genes in the two modules were also significantly enriched for plant hormone signal transduction pathways.
Transcription factors play crucial roles in regulating genes involved in various aspects of plant growth and development, encompassing secondary metabolism. In this study, the transcription factors in ME-A and ME-F modules include four bHLHs, ten MYBs, one NAC, seven WRKYs, four LBDs, nine ZINC-FINGERs, ten AP2/ERFs, two MADSs and five bZIPs. Among them, DN643_c0_g1 (GLABRA3, a bHLH transcription factor) might involve in anthocyanin biosynthesis. Additionally, DN1014_c1_g4 (MYBD), DN1283_c0_g1 (MYB7), DN13711_c0_g1 (MYB4) and DN4322_c0_g2 (MYB111) were related to flavonol and anthocyanin biosynthesis.
The top 20 hub genes associated with phenylpropanoid and flavonoid biosynthesis, which exhibiting the highest connectivity values in the ME-A and ME-F modules, were subjected to gene correlation analysis with Cytoscape software (Fig. 4A, B). Two genes related to lignin biosynthesis (DN1716_c0_g1, KNAT7 and DN1657_c0_g1, CRL2) in the ME-A module. A gene related to coumarin synthesis (DN44914_c0_g1, COSY) and a MYB transcription factor (DN13711_c0_g1, MYB4) related to flavonoid biosynthesis in the ME-F module. Other genes might also be involved in phenylpropanoid and flavonoid biosynthesis, which needs further verification.
3.6 Phenylalanine Biosynthesis in Different Tissues of B. purpurasceus
At least 25% of photosynthetic products in plants are stored in phenylpropanoids (e.g., lignans, flavonoids) derived from phenylalanine, and multiple classes of secondary metabolites associated with this pathway were identified in B. purpurasceus associated with this pathway. We therefore analyzed the key DEGs and DAMs involved in the phenylalanine biosynthetic pathway. We identified 19 candidate enzyme genes for phenylalanine biosynthesis among all DEGs, including AcoC (chorismate synthase), CM (chorismate mutase), PAT (prephenate aminotransferase), and ADT (arogenate dehydratase), whose expression levels were mostly up-regulated in leaf (Table S9_Sheet1). The accumulation of phenylalanine was higher in stem and the expression level of ADT was relatively high in the stem (Fig. 5B). This pathway is also the upstream pathway for the synthesis of several alkaloids, and Fig. 5A shows the accumulation pattern of 22 alkaloids identified in B. purpurasceus (Fig. 5A).
3.7 Phenylpropanoid Biosynthesis in Different Tissues of B. purpurasceus
Comparative analysis of the transcriptome and metabolome of different organs of B. purpurasceus, DEGs and DAMs in different comparisons were significantly enriched in the phenylpropanoid biosynthesis pathway. We analyzed the major DEGs and DAMs involved in this pathway. We identified 89 candidate enzyme genes related to phenylpropanoid biosynthesis (Table S9_Sheet2). Figure 6 shows that PAL (phenylalanine ammonia-lyase), 4CL (4-coumarate–CoA ligase), and CAD (cinnamyl-alcohol dehydrogenase) genes were up-regulated in leaves, E1.11.1.7 (peroxidase) gene were mainly up-regulated in stems and roots, and CYP73A (trans-cinnamate 4-monooxygenase) was highly expressed in all tissues. A total of 12 metabolites were enriched in the phenylpropanoid biosynthesis pathway. The accumulated levels of p-coumaraldehyde, p-coumaryl alcohol, coniferyl alcohol were highest in the leaf. The accumulated levels of phenylalanine and caffeyl aldehyde were highest in the stem. The accumulated levels of p-coumaric acid, ferulic acid, coniferyl aldehyde and coniferin were highest in the root. (Fig. 6, Table S10_Sheet2)
3.8 Flavonoid Biosynthesis in Different Tissues of B. purpurasceus
DEGs and DAMs in all comparisons were significantly enriched in the flavonoid biosynthesis pathway. We identified 22 candidate enzyme genes for flavonoid biosynthesis, mostly up-regulated in stems, including E5.5.16 (chalcone isomerase), CHS (chalcone synthase), CYP73A (trans-cinnamate 4-monooxygenase), ANS (anthocyanidin synthase), F3H (naringenin 3-dioxygenase), F3’H (flavonoid 3'-monooxygenase) and DFR (dihydroflavonol 4-reductase) (Fig. 7, Table S9_Sheet3). Nineteen flavonoid metabolites were identified in flavonoid biosynthesis pathway, mostly up-regulated in root, such as epicatechin and catechol. Naringenin chalcone and naringenin were up-regulated 215.3-fold and 8.8-fold in the root vs. leaf, up-regulated 23.8-fold and 24-fold in the root vs. stem. (Fig. 7, Table S10_Sheet3)