Transcriptomics integrated with targeted metabolomics reveals endogenous hormone changes in tuberous root expansion of Pueraria

Abstract

Background

Pueraria is a widely cultivated medicinal and edible homologous plant in Asia, and its tuberous roots are commonly used in the food, nutraceutical, and pharmaceutical industries. “Gange No. 5” is a local variety of Pueraria montana var. thomsonii (Bentham) M.R. Almeida (PMT) in Jiangxi Province, China. After optimizing its cultivation technique, we shortened the cultivation cycle of this variety from two years to one year, suggesting that the regulatory mechanism of the endogenous hormone system during tuberous root expansion may have changed significantly. In this study, we focused on the molecular mechanisms of endogenous hormones in promoting tuberous root expansion during one-year cultivation of “Gange No. 5”.

Results

The mid-late expansion period (S4) is critical for the rapid swelling of “Gange No. 5” tuberous roots during annual cultivation. At S4, the number of cells increased dramatically and their volume enlarged rapidly in the tuberous roots, the fresh weight of a single root quickly increased, and the contents of multiple nutrients (total protein, total phenol, isoflavones) and medicinal components (puerarin, puerarin apigenin, and soy sapogenin) were at their peak values. We compared the transcriptomes and metabolomes of S1 (the pre-expansion period), S4, and S6 (the final expansion period), and screened 42 differentially accumulated hormone metabolites and 1,402 differentially expressed genes (DEGs) associated with hormone biosynthesis, metabolism, and signaling. Most Auxin, cytokinins (CKs), jasmonic acids (JAs), salicylic acid (SA), melatonin (MLT), and ethylene (ETH), reached their maximum levels at S1 and then gradually decreased; however, abscisic acid (ABA) appeared in S6, indicating that most of the endogenous hormones may play a key role in regulating the initiation of tuberous root expansion, while ABA mainly promotes tuberous root maturation. Notably, multiple key genes of the ‘Tryptophan metabolism’ pathway (ko00380) were significantly differentially expressed, and COBRA1, COBRA2, YUCCA10, IAA13, IAA16, IAA20, IAA27, VAN3, ACAA2, and ARF were also identified to be significantly correlated with the expansion of “Gange No. 5” tuberous roots.

Conclusions

Our study has revealed how endogenous hormone regulation affects the expansion of “Gange No. 5” tuberous roots. These findings offer a theoretical foundation for improving the yield of PMT tuberous roots.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-024-11010-w.

Introduction

To date, there are approximately 35 species of Pueraria DC. Worldwide, primarily distributed in subtropical and temperate regions such as China, Russia, Japan, Korea, India, Vietnam, Malaysia [1]. China is the center of origin for Pueraria DC., with about 12 species [2]. Among these, Pueraria montana var. lobata (Willdenow.) Maesen & S. M. Almeida ex Sanjappa et Predeep (PML) and Pueraria montana var. thomsonii (Bentham) M. R. Almeida (PMT) are the two most significant species. They are utilized in medicine, food, and animal feed and hold high economic value [3, 4].

The yield of PMT is directly related to both the number of tuberous roots and their degree of thickening. Similar to sweet potatoes, the morphology of the underground parts of PMT can be categorized into three types: fibrous roots, faggot roots, and tuberous roots [5, 6]. Among these, the fibrous root is the primary organ for water and nutrient absorption; faggot roots also absorb some water and nutrients but eventually consume more nutrients than they absorb as they grow; tuberous roots serve as the carbohydrate storage organ and are the primary focus of harvesting in current artificial cultivation [7]. Tuberous roots expand and grow in a fusiform shape, while faggots thicken uniformly [5]. Some researchers have proposed that faggots thicken to form tuberous roots [8], similar to sweet potatoes, whereas others suggest that faggots become highly lignified and cease to expand into tuberous roots [9].

Generally, tuberous root expansion is strongly influenced by environmental factors such as light, temperature, soil, water, and fertilizer, all of which impact tuberous root development to varying degrees [10–12]. Additionally, the content and proportion of carbohydrates in the root system are crucial for tuberous root development. For example, starch is a major component of tuberous roots [13], sucrose is a key product and form of photosynthesis, and lignin synthesis competes with starch synthesis, showing a negative correlation with tuberous root formation [6]. Plant hormones are also critical in this process. Cytokinins, growth hormones, and abscisic acid play significant roles in regulating tuberous root expansion. For instance, ZR, DHZR, and tZR are important cytokinins that regulate tuberous root formation [14]. IAA (indole-3-acetic acid) is essential for attracting and transporting carbohydrates, and in sweet potatoes, IAA promotes tuber expansion by activating the tuber formation layer and increasing xylem cell proliferation [15]. Abscisic acid is linked to abnormal tuber formation layer activity and facilitates tuber expansion by promoting cell division in secondary meristematic tissues [16]. Conversely, GA₃ (gibberellic acid) inhibits tuber formation and expansion [17], while JA (jasmonic acid) and its derivatives promote tuber or tuberous root expansion [18]. Previously, most molecular-based studies on Pueraria have focused on puerarin biosynthesis and gene expression regulation. For instance, Wang et al. [19] sequenced the transcriptome of wild tuberous roots and leaves, predicting 47 potential structural genes for isoflavone synthesis, 22 uncharacterized transcriptomes (UTGs), and 45 methyltransferases. Hu et al. [20] conducted a comprehensive analysis of metabolite and transcriptome changes in the puerarin biosynthetic pathway during peak and dormant growth periods. In contrast, few studies have been reported on the molecular regulatory mechanisms of endogenous hormone changes during tuberous root expansion.

PMT cultivation in most areas of China is biennial, such as the traditional local variety “Gange No. 5” in Jiangxi Province, China, resulting in poor economic returns due to the long planting cycle and high labor costs, significantly limiting farmers’ motivation. Originally used as fodder for cattle and goats, and later for milling or as an edible vegetable, “Gange No. 5” has been cultivated locally as a biennial. Interestingly, after systematically improving key cultivation techniques for “Gange No. 5”, such as tuberous root pruning, vine pruning and fertilization increasing, we discovered that this variety has the potential to be cultivated annually. In this study, we focused on “Gange No. 5” to analyze the relationship between endogenous hormone changes and tuberous root expansion when using an annual production model. Our goal is to provide theoretical foundations for promoting the development of tuberous roots in this variety and ultimately achieving higher yields.

Results

Expansion process of tuberous roots

Figure 1A (a) shows the field growth of “Gange No. 5” plants in July. Figure 1A (b–c) illustrates the dynamic changes in tuberous root elongation and thickening from the initiation of expansion to harvest. We observed that the PMT variety “Gange No. 5” had the greatest increase in growth rate at stage S4 (the mid-late expansion period) when planted as an annual. Compared with stage S3 (the mid-expansion period), the fresh weight increased by 131.30%, the dry weight by 168.92%, the transverse diameter by 48.55%, and the longitudinal diameter by 14.08% at S4 (Fig. 1B and Table 1). At stage S6 (the final expansion period), the harvested fresh weight of individual tuberous roots reached up to 5.07 kg, which met the harvesting requirement of commercial kudzu.

Fig. 1.

Tuberous root phenotypes of PMT local variety “Gange No. 5” during the different expansion periods. A Plant phenotypes of PMT local variety “Gange No. 5” in the field in mid-July (a), and intact morphology (b) and cross-sectional morphology (c) of tuberous root at different expansion periods. B Statistical analysis of traits related to tuberous roots at different expansion periods. Data are presented as mean ± SD (n = 3). Student’s t-test was used to generate P-values, different letters indicate P < 0.05

Table 1.

Morphological indexes in the period of “Gange No.5” tuber formation

Treatment	Fresh weight (kg)	Dry weight (kg)	Transverse diameter (cm)	Longitudinal diameter (cm)
S1	0.29 ± 0.01f	0.18 ± 0.01f	3.50 ± 0.08f	39.30 ± 0.82f
S2	0.76 ± 0.02e	0.47 ± 0.02e	5.80 ± 0.12e	46.82 ± 0.90e
S3	1.15 ± 0.07d	0.74 ± 0.05 cd	7.25 ± 0.06d	50.86 ± 1.25d
S4	2.66 ± 0.09c	1.99 ± 0.06c	10.77 ± 0.07c	58.02 ± 0.92c
S5	3.31 ± 0.08b	2.51 ± 0.10b	11.85 ± 0.12b	64.02 ± 1.01b
S6	5.07 ± 0.06a	3.96 ± 0.15a	13.07 ± 0.27a	65.60 ± 0.55a

Different lowercase letters indicate significant differences (p< 0.05)

Based on this, we concluded that “Gange No. 5” a local PMT variety, can be effectively used for annual production.

Changes in nutritional compositions of tuberous roots

To understand the variation in tuberous roots during one-year cultivation, we investigated several nutrient quality characteristics at six expansion stages. As shown in Fig. 2, the protein content fluctuated relatively smoothly throughout the expansion period, with higher levels observed at stages S1 (the pre-expansion period) and S3 and the lowest at S5 (the late expansion period), and these differences were statistically significant (P < 0.05) (Fig. 2A). The trends in soluble sugar and starch contents were similar, with minimal changes from S1 to S3, a significant increase at S4, and stabilization at their maximum levels at S6 (Fig. 2B, C). This indicates that substantial accumulation of soluble sugars and starches in the tuberous roots begins at S4. The maximum values of VC (vitamin C) and cellulose content were recorded at S1, after which both began to decrease (Fig. 2D, E). In contrast, the reducing sugar content initially declined but then increased, reaching its peak at S6 (Fig. 2F).

Fig. 2.

Dynamic changes in nutrient composition during different periods of tuberous root expansion. A Total protein. B Soluble sugars. C Starch. D VC. E Cellulose. F Reducing sugars. G Total phenolics. H Isoflavones. Data are presented as mean ± SD (n = 3). Student’s t-test was used to generate P-values, different letters indicate P < 0.05

Overall, the phenotypic changes of tuberous roots were prominent at stage S4 (Fig. 1A), and the increases in their fresh weight, dry weight, and transverse and longitudinal diameters reached the maximum; in addition, the increases in soluble sugars and starches were also large, and total phenolics and isoflavonoids reached their peaks at S4 (Fig. 2G, H). Thus, we speculate that the S4 stage may be a critical period for tuberous root expansion of “Gange No. 5” cultivated as an annual.

Changes in medicinal compositions of tuberous roots

The primary medicinal compounds in the tuberous roots of PMT are flavonoids and isoflavones. Clinically, flavonoids are often used for their antioxidant and anti-inflammatory properties, and isoflavones for cardiovascular disease. Here, we analyzed the changes in these medicinal compositions across six growth periods. Of these, the content of 3'-hydroxygeraniniol initially decreased gradually from S1 to S4, increased rapidly at S5, peaking at 0.17 μg/μL, and then decreased to 0.13 μg/μL in S6 (Fig. 3A). This indicates that S5 is the main period for the accumulation of 3'-hydroxygeraniniol. The trends in puerarin and puerarin apigenin contents were similar: both showed a continuous and significant decrease from S1 to S3, followed by a substantial increase at S4, reaching peak values of 29.97 μg/μL and 2.05 μg/μL, respectively. These levels then stabilized until harvest (Fig. 3B, C). In addition, soybean glycoside content fluctuated considerably throughout the growth period. It decreased rapidly from S1 to S2 (the early expansion period), followed by a slow increase from S2, reaching a peak value of 0.77 μg/μL at harvest (S6) (Fig. 3D). Similarly, adenosine levels reached a low point at S4, while soybean sapogenin levels peaked (Fig. 3E, F). All levels eventually stabilized. Finally, genistein content, though less variable, also peaked at harvest (Fig. 3G).

Fig. 3.

Dynamic changes in medicinal composition during different periods of tuberous root expansion. A 3'-hydroxygeranin. B Puerarin. C Puerarin apoigenin. D Soybean glycosides. E Adenosine. F Soybean sapogenin. G Genistein. Data are presented as mean ± SD (n = 3). Student’s t-test was used to generate P-values, different letters indicate P < 0.05

In summary, among the seven medicinal components, the peaks of puerarin, puerarin apigenin, and soybean sapogenin all appeared in stage S4, suggesting that S4 may also be a critical period for the accumulation of medicinal compositions.

Observation of ultrastructural characteristics

Given that stage S4 was identified as the peak of tuberous root expansion, accompanied by significant changes in nutritional and medicinal characteristics, we chose to make ultrastructural observations on tuberous roots from three expansion stages, S1, S4, and S6.

As shown in Fig. 4, during the S1 period, the tuberous root ducts were regular in shape and well-developed, with lumen diameters significantly larger than those of the thin-walled cells in the xylem. The amyloplasts in the thin-walled cells were few in number and sparsely distributed, primarily located around the vascular formation layer after the primary phloem. At the S4 period, there was an increase in the proportion of secondary xylem as the tuber expanded. The number of ducts was more differentiated compared to S1. Continuous cell division in the vascular formation layer led to rapid thickening of the root. The thin-walled cells, originally round, became irregular ellipsoids as they compressed each other, and amyloplasts were densely packed within the cells. Additionally, starting from the middle column of the pith, secondary xylem, newly formed xylem, and amyloplasts exhibited a radial, ordered, dense, and randomly sparse distribution pattern. By S6, the number of cells had gradually decreased, but the ducts remained regular, well-developed, and widely distributed. At this stage, starch was primarily accumulated in the thin-walled cells of the secondary xylem, with some starch granules also found in the cells of the secondary phloem.

Fig. 4.

Cell biological observations on cross- and vertical sections of tuberous roots in three expansion periods: S1, S4, and S6. The scale is 2000 μm and 200 μm, respectively. Px: primary xylem; Sx: secondary xylem; SPh: secondary phloem; PPh: primary phloem; VC: vascular cambium; SG: starch grain; WF: wood fiber

Taken together, we speculate that stage S4 may also be an active period of cell division and volume expansion within the tuberous roots.

Transcriptome analysis during tuberous root expansion

We performed high-throughput Illumina HiSeq sequencing to complete the transcriptome analysis of nine groups of samples from three expansion periods of tuberous roots: S1, S4, and S6. As detailed in Table 2, we obtained 59.52 Gb of high-quality sequence data, with each sample group providing over 5 Gb of high-quality bases. The Q30 values for each sample group exceeded 94%, indicating an overall sequencing error rate of 0.04%, which reflects high accuracy. The GC ratio ranged from 45 to 46%, suggesting stable instrument performance during sequencing.

Table 2.

Results of sequencing data

Sample	Raw reads	Clean reads	Clean base (G)	Error rate (%)	Q20 (%)	Q30 (%)	GC content (%)
S1-1	42697372	38382592	5.76	0.02	98.36	94.95	46.00
S1-2	42697372	45006200	6.75	0.02	98.37	94.88	45.76
S1-3	55797016	53519110	8.03	0.02	98.3	94.76	45.61
S4-1	45595696	43261724	6.49	0.02	98.34	94.9	45.81
S4-2	45513388	43362620	6.5	0.02	98.25	94.65	45.48
S4-3	47027464	44724630	6.71	0.02	98.38	94.98	45.54
S6-1	44630938	42293656	6.34	0.02	98.35	94.84	44.73
S6-2	45134908	42789960	6.42	0.02	98.29	94.69	45.00

Principal component analysis (PCA) of these transcripts revealed low variability among the three biological replicates for each stage. Principal component 1 and principal component 2 accounted for 40.04% and 25.31% of the variation in these samples, respectively, resulting in clear separation among the three expansion stages (Fig. 5A). These results confirm that the sequencing data are of high quality and suitable for subsequent analysis.

Fig. 5.

Transcriptome analysis of tuberous roots in three expansion periods: S1, S4, and S6. A Scores scatter plot of the tuberous root transcriptome at S1, S4, and S6 as determined by PCA. B Comparison of DEGs between S1_vs_S4, S1_vs_S6, and S4_vs_S6; the red column represents the up-regulated genes and the blue column represents the down-regulated genes. C Venn diagram of regulated genes between S1_vs_S4, S1_vs_S6, and S4_vs_S6. D KEGG enrichment analysis among S1_vs_S4 (a), S1_vs_S6 (b), and S4_vs_S6 (c). E Expression heatmap of plant hormone-related DEGs (partial display), including Auxin (a), ABA (b), CKs (c), GAs (d), SA (e), JAs (f), ETH (g), and MLT (h)

We identified 17,582 differentially expressed genes (DEGs) across the three developmental periods (Table S1). Specifically, we detected 9,950 DEGs in the S1_vs_S4 comparison, with 5,609 down-regulated and 4,341 up-regulated genes (Table S2). In the S1_vs_S6 comparison, there were 13,432 DEGs, including 7,945 down-regulated and 5,487 up-regulated genes (Table S3). For the S4_vs_S6 comparison, 9,635 DEGs were found, with 5,575 down-regulated and 4,060 up-regulated genes (Table S4). A total of 2,295 DEGs were shared among all three comparisons (S1_vs_S4_vs_S6) (Fig. 5B, C).

Functional enrichment analysis of these DEGs, based on the KEGG ortholog database, revealed 20 pathways categorized for each comparison: for S1_vs_S4, the top three pathways were the metabolic pathway, biosynthesis of secondary metabolites, and biosynthesis of cofactors; for S1_vs_S6, the top three were the metabolic pathway, biosynthesis of secondary metabolites, and pyruvate metabolism; for S4_vs_S6, the top three were the metabolic pathway, starch and sucrose metabolism, and amino sugar and nucleotide sugar metabolism. Additionally, DEGs were significantly enriched in pathways related to the biosynthesis of nucleotide sugars, cyanoamino acid metabolism, biosynthesis of various plant secondary metabolites, degradation of flavonoids, and O-antigen nucleotide sugar biosynthesis (Fig. 5D). We also identified 1,402 DEGs involved in endogenous hormone biosynthesis, catabolism, and signal transduction during tuberous root expansion. These included 107 genes related to GAs, 6 to MLT, 131 to CKs, 482 to ABA, 186 to JAs, 334 to ETH, 412 to Auxin, and 384 to SA (Table S5).

Based on this, we concluded that tuberous root expansion was closely related to the changes in endogenous hormones, particularly on ABA, Auxin, and SA (Fig. 5E).

Metabolomic analysis of tuberous root hormone targeting

According to previous studies on tuberous root expansion and our transcriptome results, we further analyzed the hormone-targeted metabolome of tuberous roots at different expansion periods. The identification of metabolites at the three expansion periods revealed a total of 42 major differential metabolites, including 10 Auxins, 15 Cytokinins, 8 Jasmonic Acids, 4 Salicylic Acids, 2 Abscisic Acids, 1 Gibberellin, 1 Melatonin, and 1 Ethylene, which were enriched in 14 KEGG pathways. Among these, 12-OH-JA, 2MeScZR, BAP, DHZR, H2JA, IAA, ICAld, IPR, IPRMP, JA, JA-ILE, JA-Phe, JA-Val, MEIAA, SA, tZR, and tZRMP showed a significant gradient decrease with the expansion phase, only DZ, Phe, and tZOG showed a gradient of increasing, while the highest levels of ABA were found in stage S6 (Fig. 6A, B).

Fig. 6.

Metabolome analysis of tuberous roots in three expansion periods: S1, S4, and S6. A Comparison of the levels of 42 endogenous hormones in three expansion periods of tuberous roots. B Veen analysis of the metabolic pathways of the different endogenous hormone contents between S1_vs_S4, S1_vs_S6, and S4_vs_S6. C Comparison of the endogenous hormone content of tuberous roots in three expansion periods: S1, S4, and S6. D Comparison of the correlations between the main endogenous hormones. E KEGG enrichment analysis between S1_vs_S4, S1_vs_S6, and S4_vs_S6. F K-means cluster plot for differential metabolites in at least one of the three period comparison groups

Next, we investigated the correlations between the eight endogenous hormones. The results showed that SA had the highest correlation, followed by Auxin, but the change trends of the two were exactly opposite (Fig. 6C). In addition, CKs, JAs, SA, and MLT were positively correlated with each other, and CKs were significantly correlated with MLT (r > 0.9). Auxin, ABA, and ETH were also positively correlated with each other, with Auxin being significantly correlated with ETH (r > 0.9). GAs was positively correlated with ABA, but CKs, SA, and MLT were significantly negatively correlated with ABA (r < -0.90). JAs were also significantly negatively correlated with GAs (r < -0.90) (Fig. 6D).

To further understand the potential biological functions of tuberous root differential metabolites over expansion time, we compared all differentially expressed metabolites in the KEGG database. The results showed that the S1_vs_S4 and S4_vs_S6 comparison groups were significantly enriched in the Tryptophan pathway. Of these, one differentially expressed metabolite was significantly up-regulated and three differentially expressed metabolites were significantly down-regulated in the S1_vs_S4 group (P-value = 0.2308), while two differentially expressed metabolites were significantly up-regulated and two differentially expressed metabolites were significantly down-regulated in the S4_vs_S6 group (P-value = 0.2782) (Fig. 6E and Table 3). Furthermore, we present the K-means clustering heatmap of the metabolites that differed in at least one of the comparison groups across the three periods. The trend analysis showed that the 36 metabolites were divided into a total of six clusters. In cluster 1, the expression trend displayed a small change in metabolite abundance during the first two periods and an extremely rapid increase in the third period. These differential metabolites include dihydrozeatin, 3-indolacetonitrile, ortho-topolin-9-glucoside, GA₃, and dihydrozeatin-7-glucoside (Fig. 6F and Table 4).

Table 3.

KEGG enrichment analysis of DEGs for metabolome data

KEGG pathway	Rich Factor	Group	Num	P-values	Up	Down
Tryptophan metabolism	0.8	S1_vs_S4	4	0.2308	1	3
Metabolic pathways	0.58	S1_vs_S4	7	0.6077	1	6
Biosynthesis of secondary metabolites	0.53	S1_vs_S4	8	1	1	7
Phenylalanine, tyrosine and tryptophan biosynthesis	0.33	S1_vs_S4	1	0.9375	1	0
Biosynthesis of various plant secondary metabolites	0.33	S1_vs_S4	1	0.9375	1	0
Naphthalene degradation	1	S1_vs_S4	1	0.5625	0	1
Biosynthesis of siderophore group nonribosomal peptides	1	S1_vs_S4	1	0.5625	0	1
Microbial metabolism in diverse environments	1	S1_vs_S4	1	0.5625	0	1
Degradation of aromatic compounds	1	S1_vs_S4	1	0.5625	0	1
Tryptophan metabolism	0.8	S4_vs_S6	4	0.2782	2	2
Metabolic pathways	0.58	S4_vs_S6	7	0.7217	2	5
Biosynthesis of secondary metabolites	0.63	S4_vs_S6	10	0.4118	4	6
Phenylalanine, tyrosine and tryptophan biosynthesis	0.67	S4_vs_S6	2	0.6397	0	2
Biosynthesis of various plant secondary metabolites	0.67	S4_vs_S6	2	0.6397	0	2
Glycine, serine and threonine metabolism	1	S4_vs_S6	1	0.5882	0	1
Aminoacyl-tRNA biosynthesis	0.5	S4_vs_S6	1	0.8456	0	1
2-Oxocarboxylic acid metabolism	0.5	S4_vs_S6	1	0.8456	0	1
Biosynthesis of amino acids	0.5	S4_vs_S6	1	0.8456	0	1
Biosynthesis of cofactors	1	S4_vs_S6	1	0.5882	0	1
Tryptophan metabolism	0.6	S4_vs_S6	3	0.7946	1	2

Table 4.

K-means cluster analysis for metabolome data

ID	S1	S4	S6	Gene	cluster	Class
Dihydrozeatin	-0.6305	-0.5225	1.1530	Dihydrozeatin	1	CK
3-Indoleacetonitrile	-0.5774	-0.5774	1.1547	3-Indoleacetonitrile	1	Auxin
ortho-Topolin-9-glucoside	-0.5774	-0.5774	1.1547	ortho-Topolin-9-glucoside	1	CK
Gibberellin A3	-0.5774	-0.5774	1.1547	Gibberellin A3	1	GA
Dihydrozeatin-7-glucoside	-0.4722	-0.6765	1.1487	Dihydrozeatin-7-glucoside	1	CK
N-(3-Indolylacetyl)-L-alanine	-0.5774	1.1547	-0.5774	N-(3-Indolylacetyl)-L-alanine	2	Auxin
3-oxo-2-(2-(Z)-Pentenyl) cyclopentane-1-butyric acid	-0.5774	1.1547	-0.5774	3-oxo-2-(2-(Z)-Pentenyl) cyclopentane-1-butyric acid	2	JA
cis( +)-12-Oxophytodienoic acid	-0.5901	1.1546	-0.5645	cis( +)-12-Oxophytodienoic acid	2	JA
Indole	-0.9970	1.0029	-0.0059	Indole	2	Auxin
ABA-glucosyl ester	-1.0460	0.9466	0.0994	ABA-glucosyl ester	2	ABA
trans-Zeatin	0.4774	0.6718	-1.1492	trans-Zeatin	3	CK
L-tryptophan	-0.0721	1.0341	-0.9620	L-tryptophan	3	Auxin
6-Benzyladenine	0.8346	0.2738	-1.1084	6-Benzyladenine	4	CK
Indole-3-carboxaldehyde	0.8482	0.2545	-1.1027	Indole-3-carboxaldehyde	4	Auxin
9-Ribosyl-trans-zeatin 5'-monophosphate	1.0599	-0.1332	-0.9267	9-Ribosyl-trans-zeatin 5'-monophosphate	4	CK
N-6-iso-pentenyladenosine-5'-monophosphate	1.0655	-0.1474	-0.9181	N-6-iso-pentenyladenosine-5'-monophosphate	4	CK
2-Methylthio-cis-zeatin riboside	1.0512	-0.1118	-0.9394	2-Methylthio-cis-zeatin riboside	4	CK
Dihydrojasmonic acid	1.0128	-0.0261	-0.9867	Dihydrojasmonic acid	4	JA
Jasmonic acid	1.0213	-0.0441	-0.9772	Jasmonic acid	4	JA
Indole-3-carboxylic acid	1.1547	-0.5774	-0.5774	Indole-3-carboxylic acid	5	Auxin

In conclusion, maximum levels of multiple Auxin, CKs, JAs, SA, MLT, and ETH classes appeared at stage S1, indicating that most of the endogenous hormones may play a key role in regulating the initiation of tuberous root expansion.

Conjoined transcriptome and targeted-metabolome analysis

To investigate the mechanisms underlying the differences in tuberous root expansion across the three periods, we conducted a conjoined analysis of gene expression and metabolite accumulation. Differentially expressed genes (DEGs) and differentially expressed metabolites (DEMs) from the same comparison groups were mapped to KEGG pathways to identify co-enriched pathways. The results highlighted 7 pathways for S1_vs_S4, 8 pathways for S1_vs_S6, and 10 pathways for S4_vs_S6 (Fig. 7A).

Fig. 7.

Transcriptome-metabolome correlation analysis. A KEGG enrichment correlation analysis among S1_vs_S4 (a), S1_vs_S6 (b), and S4_vs_S6 (c). B Left, transcriptome loading plot (a); right, metabolome loading plot (b)

For further insights, we employed Orthogonal Projections to Latent Structures (O2PLS) modeling. This approach allowed us to identify variables with high correlations and weights across the different datasets. Loadings plots were used to screen for important variables influencing other phenological indicators. The distance of each point from the origin, or the height of the bar, represents the magnitude of the correlation between substances and other omics, with darker colors indicating stronger correlations. The top 10 substances with the most significant influence on other omics are highlighted (Fig. 7B). These include TRP, TRA, Indole, and IAA-Ala (Auxins), as well as DHZ7G, oT9G, DZ, and tZ (Cytokinins), and OPC-4 and OPDA (Jasmonic Acids).

We focused on the ‘Tryptophan metabolism’ pathway (ko00380) due to its high number of entries and significant differences. Indole-3-acetic acid (IAA), an essential phytohormone, promotes tuberous root growth through polar and lateral transport from the roots. Tryptophan, the precursor of IAA, undergoes a series of transformations starting from decarboxylation to form tryptamine, which is then converted through oxidative deamination to IAA. Additionally, tryptophan is the precursor for melatonin synthesis. It is first converted to tyrosine by tryptophanase, then to dopamine by tyrosine oxidase, and further to dopanone by dopamine decarboxylase. Dopamine is eventually converted to melatonin through a series of enzymatic reactions. The expression levels of genes involved in this pathway, including ALDH, ACAT, DLD, katG, and DDC, showed significant changes (Fig. 8).

Fig. 8.

Mechanisms underlying the tryptophan metabolic pathway. This pathway was constructed based on the KEGG pathway and literature sources. Each colored cell represents the values of each compound ion normalized by converting into log10 units according to the color scale. ACAT, acyl-coenzyme A-cholesterol acyltransferase; DLD, dihydrolipoamide dehydrogenase; ALDH, acetaldehyde dehydrogenase; katG, peroxidase gene; DDC, dopa decarboxylase

Typically, tryptophan is a precursor for Auxin hormone synthesis, and our results suggest that tryptophan metabolism may influence Auxin synthesis during tuberous root expansion.

Quantitative real-time PCR (RT-qPCR) validation of gene expression profiles

Ten representative DEGs in S1, S4, and S6 were selected for RT-qPCR validation, with the screening criterion that the expression of this stage was more than 2-fold higher than the other two stages. Finally, the validation results were largely consistent with the transcriptome data (Fig. 9A, B), indicating a high degree of confidence in the transcriptome results. The ten candidate genes exhibited roughly three expression patterns. The ten candidate genes exhibited roughly three expression patterns. Of these, the expression of COBRA1 (Pmon001G07833), COBRA2 (Pmon002G06148), IAA20 (Pmon002G01012), and YUCCA10 (Pmon001G01339) showed a gradual decrease from S1-S6. In contrast, the expression of IAA16 (Pmon002G00544), IAA27 (Pmon001G00306), and VAN3 (Pmon003G00784) significantly increased from S1 to S4 and then significantly decreased by S6. The expression of IAA13 (Pmon002G02161), ACAA2 (Pmon001G00706), and ARF (Pmon007G00373) showed a gradual increase from S1-S6.

Fig. 9.

RNA-seq analysis and RT-qPCR validation of 10 genes in PMT tuberous roots at three expansion periods. A Correlation heatmap of the DEGs modules and carotenoid at S1, S4, and S6 of PMT tuberous roots, including four DEGs that decrease consistently with tuberous root expansion (a), three DEGs that ‘rise and then fall’ (b), and three DEGs that increase consistently (c). B RT-qPCR verification of 10 differential genes

Discussion

The most striking phenotypic feature of expanding plants is their enlarged edible organs, which are inextricably linked to cell division. For example, the formation of potato tubers is due to rapid cell division and expansion of the cortex, medullary ring, and pith tissue [21]. The fleshy roots of radish are formed as a result of secondary thickening due to formation layer activity [22], and the enlargement of sweet potato storage roots is also caused by the division and proliferation of cells in the formation layer and the accumulation of intracellular starch [23]. In this study, the tuberous root expansion of “Gange No. 5” was similarly caused by the division activity of the vascular formation layer, in which the fusiform primordial cells in the tuberous root underwent periclinal and anticlinal cell divisions since the S1 period. This caused the formation layer to gradually push outward, resulting in the enlargement of the tuberous root circumference, which was similar to previous results for another local PMT variety, “Gange No. 1” in Jiangxi Province, China [4].

As the main economic organ of “Gange No. 5” the tuberous root initiation of expansion (including cell division and initiation of thickening) and subsequent thickening is a complex biological process involving morphogenesis and the accumulation of assimilated products (carbohydrates, proteins, lipids, secondary metabolites, etc.). After optimizing the cultivation technique, we shortened the cultivation cycle of “Gange No. 5” to one year. However, the peak of tuberous root growth of “Gange No. 5” occurred in mid- to late October (S4 period), which differed from that of “Gange No. 1” which peaked in early September [4]. We speculate that this may be related to the different genetic backgrounds of the two varieties. “Gange No. 1” belongs to an early-maturing variety, and the differentiation activity of the formation layer may occur earlier. In contrast, the peak of tuberous root expansion in “Gange No. 5” came later, suggesting that its growth cycle may be longer. At S4, total phenolics, isoflavones, 3'-hydroxyanthocyanins, puerarin, puerarin-apigenin, and soy sapogenins in tuberous roots of “Gange No. 5” reached their highest levels, while the contents of other medicinal and edible components were also elevated. For example, the soluble and reducing sugar content showed an increasing trend at this stage, while the starch content also increased rapidly, suggesting that sugars in the tuberous roots may be rapidly converted into starch and stored. In contrast, the cellulose content in the tuberous root showed a gradient decrease with the onset of expansion. This decrease could be related to the gradual activation of pectinases or changes in the balance between degradation and reconstruction of the cell wall, similar to results obtained in banana and plantain [24] and wolfberry [25]. Notably, changes in the content of these assimilated products, such as sugars, starch, and cellulose, all characterize the yield of tuberous roots [26].

Historically, inputs from hormonal signaling have also been thought to drive organ enlargement, while the internal balance of phytohormones is essential for stage transitions and tuberous root expansion [27]. Although CKs, Auxin, JA, and ABA have been reported to play different roles in tuberous root formation and thickening [28], the mechanism of hormonal regulation of tuberous root expansion still remains unclear. In this study, we focused on the transcriptome and metabolome of “Gange No. 5” tuberous roots from S1 to S4, and finally to S6 to identify key genes involved in phytohormone synthesis and regulation during this process and to elucidate the underlying molecular mechanisms. We screened 1,402 DEGs related to hormone biosynthesis and metabolism from the transcriptome and identified 42 phytohormones in the corresponding metabolome, including eight major classes: Auxin, CKs, JAs, SA, ABA, GAs, MLT, and ETH. Of these, the IAA content reached its maximum at the S1 stage and gradually decreased with tuberous root expansion, consistent with potato tuber development [29, 30]. Similarly, other Auxin-related compounds such as IAA-Phe-Me, ICA, and ICAld also exhibited their highest values at S1. Auxin mediates the direction of plant cell division, and Auxin content in stolon tips increases significantly before tuber expansion [31], which aligns with our results. Furthermore, YUCCA10 and TAR2, a homologous protein of TAA1 [32], which promote indoleacetic acid biosynthesis, were also significantly down-regulated with tuberous root expansion (Table S5), consistent with the observed changes in IAA content. Changes in Auxin-related hormone levels also induced the expression of a series of genes, including the Auxin endotransporters AUX1 and LAX3, which were significantly up-regulated during the gradual expansion of PMT tuberous roots. This up-regulation could promote the polar transport of IAA, thereby inducing tuberous root expansion. The expression of ARF was strongly up-regulated before potato formation, and this gene is associated with the development of vascular tissue, leading to the formation of phloem; however, the expression level of ARF6 decreased significantly with tuber formation [33]. AUX/IAA is a gene involved in the Auxin response and synergizes with TIR1 to regulate ARF expression [34]. In this study, TIR1 and GH3 were significantly down-regulated during PMT tuberous root expansion, but AUX/IAA and ARF were significantly up-regulated (Table S5). This suggests that the Auxin response process may be facilitated by the up-regulation of ARF expression to promote PMT tuberous root expansion.

CKs are another large group of hormones that promote lateral cell division and cell expansion. In this study, the levels of DZ, Ot9G, and tZOG, which are associated with CKs, gradually increased, and BAP, DHZR, IP, IPR, IPRMP, oTR, tZR, and tZRMP gradually decreased; whereas the differential genes associated with their metabolism included both IPT and CKX genes (Table S5). In Arabidopsis, the catalytic reaction of IPT with dimethylallyl diphosphate (DMAPP) as the substrate is the first step in the biosynthesis of CKs, and the same catalytic enzymes are involved in the biosynthesis of zeaxanthin and IP [35]. In contrast, CKX catalyzes the degradation of CKs, and overexpressing CKX7 strains have shown significantly lower cytokinin levels [36]. In our study, the significant down-regulation of both IPT and CKX genes with the gradual expansion of PMT tuberous roots suggests that the late expansion process of PMT tuberous roots may not be dependent on CK content, which is consistent with the results observed in potato [37] and ginger [38].

GA and ABA are two classical antagonist hormones. In this study, GA₃ content was extremely low at S1 and S4 but reached its highest level at S6 with the gradual expansion of PMT tuberous roots. In its anabolic pathway, GA20ox2, which promotes the synthesis of active GA₃, and the downstream target gene expansin EXP were significantly up-regulated in expression with tuberous root expansion, consistent with the measured changes in endogenous GA₃. A key step in the GA signaling process is the degradation of DELLA proteins through the ubiquitination pathway mediated by the SCF complex [39, 40], while DELLA proteins can interact with and inhibit the function of many transcription factors [41, 42]. In this study, DELLA was significantly down-regulated from S1 (Table S5), suggesting that during the early stages of PMT tuberous root expansion, the gene reduced GA sensitivity and induced PMT tuberous root expansion by feedback regulation of GA synthesis and the expression of GA downstream target genes. Furthermore, ABA showed a “decreasing-increasing” trend in its content. In the synthesis pathway, ABA2 and ABA4, which promote ABA biosynthesis, were significantly up-regulated (Table S5), whereas the expression of CYP707A1, which promotes ABA degradation, showed a significant trend of “upward and then downward” (Table S5). Intriguingly, AAO3, a protein with proven activity on ABA, was significantly down-regulated from the S1 stage (Table S5), which is inconsistent with the finding that ABA levels increase during the S6 period. We suggest that the elevation of ABA may result from inputs from other organs, other aldehyde oxidase isoenzymes, and/or through the shunt pathway [43], which needs to be investigated in the future. Additionally, 9′-cis-epoxycarotenoid dioxygenase (NCED) is an important rate-limiting enzyme in the ABA synthesis pathway [44], but no differential changes in this gene were detected in our study. The biosynthesis of both GA and ABA starts with geranylgeranyl pyrophosphate (GGPP) in the terpene backbone biosynthetic pathway [45]. IspS catalyzes the production of isoprene from DMADP, which is also a synthetic precursor of GGPP. Although GGPP and isoprene share a common precursor, a linear correlation has been found between isoprene and ABA levels [46], and isoprene emission does not affect ABA content [47], suggesting a less competitive relationship between the two biosynthetic pathways. In this study, DXS (1-deoxy-D-xylulose-5-phosphate synthase) and GGPS (Geranylgeranyl pyrophosphate synthase) genes related to the promotion of GGPP synthesis were significantly down-regulated with the gradual expansion of PMT tuberous roots (Table S5), suggesting a gradual decrease in the provision of synthetic precursors for GA and ABA. However, GA₃ content was lower in both S1 and S4, whereas the expression of ABA2 and ABA4 was up-regulated, suggesting the promotion of ABA biosynthesis. Moreover, in the ABA response pathway, PP2CA and ABF4 were significantly up-regulated with PMT tuberous root expansion, while PYR/PYL genes were significantly down-regulated (Table S5). In lotus roots, ABF, PP2C, PYL, and SnRK2 all have positive feedback effects on rhizome development during rhizome formation [48]. In potatoes, most ABF genes are induced during stolon expansion [49], and overexpression of the ABF4 gene significantly promotes tuber formation [50]. PP2CA has been suggested to be a signaling molecule in potato leaves that promotes tuber formation, and its expression is significantly inhibited by GAs [51]. PP2C promotes tuber induction by decreasing GA bioactivity through up-regulating GA2ox1 expression during the early stages of potato tuber formation [52]. Additionally, PP2C can be regulated by ABF gene feedback, and thus induced expression by ABF [53]. The expression of PYL is affected by exogenous ABA and can positively regulate ABA-mediated responses, while ABI5 can negatively feedback regulate PYLs [54]. As a result, these findings suggest that the accumulation of ABA during PMT tuberous root expansion may negatively feedback regulate PYR/PYL and that PP2CA is promoted for expression by ABF4, which promotes tuberous root expansion mainly through feedback regulation of GA (Table S5).

In addition, JA content in PMT tuberous roots gradually decreased with tuberous root expansion, indicating that the role of JA in tuberous root expansion was reduced in the middle and late stages, and it mainly played a role in initiating tuberous root expansion. In potatoes, tuber size and frequency of tuber formation were significantly increased in lines overexpressing the JA methyltransferase gene (JMT) [55]. In the JA synthesis pathway, the expression of most genes related to the promotion of JA synthesis, such as LOX4, AOS3, and AOC4, was down-regulated, which also suggests that JA plays an important role in the pre-expansion phase of PMT tuberous roots. In Callerya speciosa, JAR1 was significantly down-regulated during storage root formation [56], and JAZ and COI1 expression was significantly down-regulated during tuberous mustard expansion [57]. No differential changes in JAR1 were detected in our study, but COI1 was significantly up-regulated, whereas JAZ and MYC2 expression was significantly down-regulated (Table S5), suggesting that the effect of JA on PMT tuberous root expansion may depend on the transduction pathway involving MYC2 expression.

Additionally, ETH is an important phytohormone that promotes germination, maturation, and senescence [58]. In this study, we found that ACC content, an important precursor of ETH synthesis, was highest at S4, which is also a critical point for the rapid expansion of PMT tuberous roots. The reduction of ACC content at S1 and S6 indicated that ETH might hinder the expansion of PMT tuberous roots. In Arabidopsis, EIN3 is considered a key regulator of ETH-sensing factors that mediate ETH-induced gene transcription and promote plant growth by adjusting the circulation of cytoplasmic vesicles and ribosomes [59, 60]. In this study, the expression of EIN3 was significantly up-regulated with the gradual expansion of PMT tuberous roots (Table S5), and the elevated ACC content in S4 (Fig. 6) indicated that ETH had a promotional effect on the expansion of PMT tuberous roots.

As for other hormones, including SA and MLT, we indirectly investigated their involvement in PMT tuberous root expansion by analyzing the expression of genes related to their synthesis and signaling. For example, ICS1 (encoding isochorismate synthase 1) and PAL (encoding phenylalanine ammonia-lyase), related to SA biosynthesis, and NPR genes involved in SA-mediated signaling, were all significantly up-regulated (Table S5). For MLT, the key genes for its synthesis, ASMT (N-acetylserotonin methyltransferase) and COMT (caffeic acid O-methyltransferase), were consistently reduced (Table S5), which suggests that MLT promotes cell expansion during the pre-tuberous root expansion period. This is consistent with results from many previous studies [61–63]. MLT can induce root primordia formation from pericycle cells in lupine (Lupinus albus), leading to adventitious and lateral root formation [64], which aligns with our study showing that MLT content reaches a maximum in the S1 period and decreases considerably thereafter.

Conclusion

This is the first detailed and comprehensive transcriptomic and metabolomic study of endogenous hormones in PMT tuberous root expansion. In this study, eight plant hormones—Auxin, CKs, JAs, SA, ABA, GAs, MLT, and ETH—were found to be closely associated with PMT tuberous root expansion. First, most of the Auxin, CKs, JAs, SA, MLT, and ETH hormones reached their maximum levels at the S1 stage and then gradually decreased with tuberous root expansion. This pattern was generally consistent with the expression patterns of the genes involved in the synthesis and signaling of these hormones (Fig. 10), suggesting that the main function of these phytohormones might be to promote the initiation of tuber expansion.

Fig. 10.

A summary of the results regarding molecular regulatory mechanisms related to endogenous hormones in the expansion of PMT tuberous roots. “ ” promotion; “ ” inhibition

Secondly, ABA-like hormones changed in the opposite direction, with their levels reaching a maximum at the S6 stage, indicating that ABA was not conducive to PMT root expansion. Moreover, from the morphological characteristics of tuber expansion, S4 was identified as the key expansion period for the PMT local variety “Gange No. 5” suggesting that the increase in cell number and cell volume in the tuber at the early stage might also be related to hormone regulation.

Materials and methods

Plant materials

In this study, we shortened the growth period of “Gange No. 5” (Pueraria montana var. thomsonii) to one year by employing improved cultivation techniques. We transplanted well-grown PMT cuttings, approximately 6–10 cm in height, into the field on March 20th, 2023. Sampling began once the tuberous roots were formed. Samples were taken on July 20th (stage S1), August 20th (stage S2), September 20th (stage S3), October 20th (stage S4), November 20th (stage S5), and December 20th (stage S6), corresponding to six stages of tuberous root expansion.

For cultivation, we used a wooden frame and bamboo frame model, retaining a single vine and tuberous root per plant. The cultivation setup included a border height of 0.3 m, ridge spacing of 1.6 m, furrow width of 0.4 m, and plant spacing of 0.8 m. Three independent biological replicates of tubers were collected at each sampling period. The tuberous root samples were immediately frozen in liquid nitrogen and stored at -80 °C until further processing.

Determination of quality indicators and metabolite components

Tuberous roots were cut into small pieces and snap-frozen in liquid nitrogen. The samples were then ground and placed in 50 mL centrifuge tubes for measurement. Protein content was determined using Coomassie Brilliant Blue G-250 staining [65]. Reducing sugars were measured using the 3,5-dinitrosalicylic acid colorimetric method [66]. Soluble sugar, starch, and cellulose contents were assessed using the anthrone-sulfuric acid colorimetric method [67]. Vitamin C (VC) content was determined using the molybdenum blue colorimetric method [68]. Total flavonoids were measured according to the method described by George et al. [69], and total phenolic content was quantified using a modified Folin-Ciocalteu method [70].

Chemical components of tuberous roots

The flavonoid components (3'-hydroxy geranylgeranyl, geranylgeranyl, geranylgeranyl apigenin, soy glycosides, and genistein) in PMT tuberous roots were determined by HPLC as described by Wang et al. [71] with slight modifications. Briefly, a C18 column (250 mm × 4.6 mm, 5 μm) was used with a flow rate of 1.0 mL/min, a detection wavelength of 254 nm, a column temperature of 30 °C, and an injection volume of 10 μL.

For the control solution, an appropriate amount of control product was placed in 2 mL centrifuge tubes, dissolved in methanol, and then diluted with a 10% aqueous methanol solution.

For the test solution, 1.0 g of powdered tuberous root sample was weighed into a 50 mL centrifuge tube. To this, 50 mL of 30% aqueous ethanol solution was added, and ultrasonic extraction was performed for 45 min. The mixture was then centrifuged at 7500 rpm for 5 min. The supernatant was separated and passed through a 0.22 μm filter membrane for analysis.

The control stock solution was serially diluted with 30% ethanol across six mass concentration gradients. A standard curve was plotted with the mass concentration of the control as the horizontal coordinate (X) and the peak area as the vertical coordinate (Y) (Table 5).

Table 5.

Regression equations and correlation coefficients for isoflavones

Standard	Regression equation	R2
3'-hydroxygeranin	y = 5.7831x - 1.1873	0.9997
Puerarin	y = 16.128x - 8.5827	0.9995
Puerarin apoigenin	y = 10.755x + 3.1616	0.9996
Soybean glycosides	y = 6.1558x - 0.4369	0.9999
Adenosine	y = 6.4948x - 2.5604	0.9997
Soybean sapogenin	y = 8.343x + 3.3658	0.9996
Genistein	y = 15.817x - 3.0654	0.9996

Ultrastructural observations of tuberous roots

Since S4 is a critical expansion period for tuberous root formation in the annual cultivation of “Gange No. 5” we collected tuberous root samples from S1, S4, and S6 for cytological characterization.

The tuberous roots were fixed and preserved in FAA fixative, then processed into conventional paraffin sections through dehydration, wax immersion, sectioning, dewaxing, staining, and sealing. These sections were observed using an orthogonal fluorescence microscope (Nikon ECLIPSE 80i, Tokyo, Japan) and photographed to capture characteristic features.

Finally, the samples from S1, S4, and S6 were stored in the herbarium of the College of Agriculture, Jiangxi Agricultural University.

RNA-seq and data analysis

Tuberous root samples were sent to Metware Biotechnology Co., Ltd. (Wuhan, China) for total RNA extraction and transcriptome sequencing, with the samples transported on dry ice. The extracted total RNA was assessed for integrity, concentration, and purity to ensure it met the quality requirements for library construction. After constructing the cDNA library and performing quality checks, sequencing was conducted on the Illumina NovaSeq platform. The raw data was then filtered to remove low-quality sequences (adapter reads, reads with more than 10% unknown base N content) to obtain high-quality clean data.

Using version 2.1.0 software [72], clean reads for each sample were aligned with the PMT reference genome to determine the positional information of the reference genome or gene and the specific sequence features of the sequenced samples. Based on this positional information, reads were assembled into transcripts using StringTie 1.3.4d [73].

A comparative analysis of transcriptome data from the tuberous roots of “Gange No. 5” at three developmental periods was performed using high-throughput sequencing technology. Differentially expressed genes (DEGs) were identified using DESeq2 1.22.1 software [74, 75], with gene expression expressed as fragments per kilobase of transcript per million fragments mapped (FPKM). DEGs were screened with criteria of |log2 fold change|≥ 1.0 and FDR < 0.05. Genes meeting these criteria were considered DEGs. These DEGs were then compared with KEGG, GO, KOG, NR, Swiss-Prot, and Pfam databases to obtain functional annotations. Subsequently, the DEGs were subjected to KOG classification, GO functional enrichment analysis, and KEGG pathway enrichment analysis.

Targeted metabolomic profiling

Metabolomic data were also obtained from Metware Biotechnology Co., Ltd. (Wuhan, China). Briefly, 0.5 g of tuberous root samples, ground with liquid nitrogen, were placed in a centrifuge tube. The samples were mixed with 0.5 mL of 80% methanol solution in water, shaken, and extracted for 5 min in an ice bath. The mixture was then centrifuged for 20 min at 4 °C and 12,000 rpm. The supernatant was collected, diluted with mass spectrometry grade water to achieve a methanol content of 30%, and centrifuged for an additional 30 min at 8,000 rpm. The final supernatant was collected and injected into the LC–MS for analysis.

The metabolite content data were normalized using the polar deviation method. Accumulation patterns of metabolites among different groups were clustered using R software. Additionally, P-values and fold changes (FC) were calculated. Differential metabolites were selected based on a P-value < 0.05 and further screened under conditions of FC ≥ 2 or FC ≤ 0.5.

RNA isolation and RT-qPCR

Tuberous roots from three expansion periods of PMT were used for RNA isolation with the RNAprepPure Polysaccharide and Polyphenol Plant Total RNA Extraction Kit. The cDNA obtained from reverse transcription served as the template for real-time PCR. Primers were designed using Primer 5 software (Table 6) and PiActin was used as a reference gene.

Table 6.

Primers of RT-qPCR

Gene	Forward sequence
Pmon001G07833	F: GGCTCTGCTGTCTTGTACTTGTTTC R: CCAGCCAGGTGAAGCGATATGAC
Pmon002G06148	F: GCTGGACCCCTGATGGATATGTG R: GAACAATCCCCTTGTTCAGTGGTCT
Pmon002G01012	F: ATTAGCAGCAGCAGCAACCGT R: CATGAGCAGCTTGTGAAGTGCTCAC
Pmon001G01339	F: CCACCTTCATCTCAGGAAGCAGT R: CCCCTTATCGTACTCCGCAAGC
Pmon002G00544	F: GGCAGAGCGAGACAAGTACAAGATG R: CGTCAGATGAGGCATTAGTTGAGGT
Pmon001G00306	F: TGGCCAGAGGTTGTGAACTTCTCA R: AGTGGCGACCAGGCCAAG
Pmon003G00784	F: GAGAGGACTAGTGGTGGCTTGAACC R: CAATGGTGTCTGAGAAACCCCTCT
Pmon002G02161	F: CTCTCTCCCTATGCCACCGCT R: CCGTGGCTGGTGATTTTGCATGG
Pmon001G00706	F: CTCAACCCTTCATCATCTACCCACG R: CATCAGCAAGAGTGTCTTTGAAACCACC
Pmon007G00373	F: GGTGTTCTGGGATGAGGTTCAAGATG R: CTGGCTCATCCCAAGTAACCTGAAG
PiActin	F: TTGTTTACACTTCCAGTGGCTCC R: TCCAACCCATTCAGTTCTCCA

The qPCR reaction procedure was as follows: pre-denaturation at 95 °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 10 s, and annealing/extension at 60 °C for 30 s. After amplification, a melting curve analysis was performed with steps at 95 °C for 15 s, 60 °C for 60 s, and 95 °C for 15 s. Relative expression was calculated using the 2^−ΔΔCT method. Data significance was assessed using the t-test, and graphs were generated with Prism software.

Statistical analysis

All experimental data were presented in triplicate. Data analysis was used SPSS 17.0 software. All data were showed as mean ± standard deviation (SD). Comparisons between and within groups were made using Duncan’s multiple range test. Student’s t-test was used to derive P-values, with P < 0.05 indicating statistical differences.

Abbreviations

IAA

Indole-3-acetic acid

CK

Cytokinin

ETH

Ethylene

JA

Jasmonic acid

GA

Gibberellin

SA

Salicylic acid

ABA

Abscisic acid

MLT

Melatonin

ZR

Zeatin riboside

DHZR

Dihydrozeatin riboside

tZR

Trans-zeatin riboside

TRP

Tryptophan

DMAPP

Dimethylallyl diphosphate

NCED

9′-Cis-epoxycarotenoid dioxygenase

GGPP

Geranylgeranyl pyrophosphate

DXS

1-Deoxy-D-xylulose-5-phosphate synthase

GGPS

Geranylgeranyl pyrophosphate synthase

ICS1

Encoding isochorismate synthase 1

PAL

Encoding phenylalanine ammonia-lyase

ASMT

N-acetylserotonin methyltransferases

COMT

Caffeic acid O-methyltransferases

Authors’ contributions

X.X, Z.W., and G.F designed the experiments. W.L., Y.F., Y.Y., and W.W. conducted the experiments. Z.M. analyzed the data. X.X and W.L. wrote and Y.F. edited the manuscript. All authors read and approved the final manuscript.

Funding

This research was funded by Jiangxi Modern Agricultural Industrial Technology System Construction Project (JXARS-16) and the National Natural Science Foundation of China (32360746).

Data availability

Sequence data that support the findings of this study have been deposited in the National Center for Biotechnology Information with the primary accession code PRJNA1132671.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.