Transcriptomic and physiological analysis reveal phytohormone and phenylpropanoid biosynthesis in root of Cynanchum auriculatum

Baishouwu (Cynanchum auriculatum), a medicinal and food dual-use plant, has been cultivated for centuries and is favored by consumers. C. auriculatum tuberous roots contain enormous amounts of flavonoids, lignin, and other nutrients. However, the developmental characteristics and phenylpropanoid metabolic mechanism in C. auriculatum have not been clarified. Here, the phenotypes of C. auriculatum tuberous roots at three stages of developmental were observed, compared with root forming stage (S1), there were significant morphological differences at root expanding stage (S2) and harvest stage (S3). Through next-generation sequencing (NGS) on Illumina HiSeq2500 platform, nine transcriptomic libraries were constructed for transcriptomic analysis. A total of 28,926 DEGs were identified during the development of C. auriculatum tuberous root, and many DEGs were enriched in GO terms/ KEGG pathways of ‘phytohormone signal transduction’ and ‘phenylpropanoid biosynthesis’. The analysis of phytohormone content and gene expression revealed that auxin and cytokinin acted in concert to regulate C. auriculatum tuberous root development, while the biosynthesis and signaling of ethylene were inhibited. Phloroglucinol staining results showed that lignified cells were mainly distributed in the central xylem at S1, followed by ring-like structure formation at S2, and finally formed the connecting rays between the xylem and the phloem at S3. Lignin content increased at S2 and then decreased at S3, and the expression of lignin synthesis genes also presented a similar trend. Total flavonoids content showed a gradually increasing trend, and the expression of flavonoid synthesis genes was also gradually up-regulated. The flavonoid synthesis pathway was enhanced by reducing the activity of key enzymes in lignin synthesis in C. auriculatum. This study provided a basis for the developmental mechanism of C. auriculatum and the further utilization of C. auriculatum tuberous roots.


Introduction
Baishouwu (Cynanchum auriculatum), a plant belonging to the Asclepiadaceae family, is distributed in China, Japan, and Korea . In China, C. auriculatum is widely cultivated in Shandong, Jiangsu, and Anhui provinces, and 95% of C. auriculatum products are produced in Binhai County, Jiangsu Province. As a medicinal and food dual-use plant material, the tuberous root of C. auriculatum has been used by local consumers as a health food for decades (Qi et al. 2009). Recent research has shown that, the biomedical function of C. auriculatum is due to the natural active substances rich in its tuberous root, including 2,4-dihydroxyacetophenone, 4-hydroxyacetophenone, cynandione, bungeiside, and so on (Li et al. 2013;Uchikura et al. 2018;Zhang et al. 2009). At the same time, a variety Communicated by Dawei Xue.
1 3 of active compounds impart pharmacological functions to C. auriculatum products, such as polysaccharide fraction against hydrogen peroxide-induced oxidative stress (Chai et al. 2018), C21 steroidal glycosides against H 2 O 2 -induced damage , and aglycone caudatin against cancer cell (Bailly 2021). However, the developmental mechanism of C. auriculatum tuberous root and the accumulation mechanism of important active substances have not been elucidated, which hinders the basic research and breeding engineering of C. auriculatum.
In plants, lignin is a polymer compound formed by the differentiation of monomers synthesized through the phenylpropanoid pathway, and it plays an important role in the growth and development mechanism (Ralph et al. 2019). As the main structural component of the cell wall in terrestrial plants, lignin maintains the hardness of the xylem, promotes water transportation, and blocks the infection of insects and pathogens (Liu et al. 2018). In certain types of plants, lignin has a significant effect on the germination mechanism of seeds, and the reduction of lignin content in the seed coat seriously affects the germination rate and growth rate (Liang et al. 2006;Mir Derikvand et al. 2008;Vanholme et al. 2012). In anthers, the absence of lignin resulted in severe inhibition of male sterility and loss of apical dominance, which ultimately manifested as stunted plant growth (Schilmiller et al. 2009;Thévenin et al. 2011). In the development of tuberous roots, lignin is primarily dedicated to the formation of xylem parenchyma cells that store starch . However, different from the root development of tree species, high lignin content is not conducive to the storage organ formation of root plant species, as reported in Ipomoea batatas (Firon et al. 2013;Sojikul et al. 2015). In the process of storage root formation, the inhibition of lignin synthesis promoted the proliferation of xylem and cambium cells, which in turn led to the thickening and growth of storage roots (Noh et al. 2013). Therefore, appropriate lignin content is of great significance for the growth and development of tuberous root plants.
Flavonoid is an important active substance of C. auriculatum, which contributes to plant development (Dong and Lin 2021) and has great potential in the fields of food health and pharmaceuticals (Parhi et al. 2020). Like the synthesis mechanism of lignin, the synthesis of flavonoids (tannins, anthocyanins, and flavonols) is also through the phenylpropanoid pathway (Yu et al. 2021). Some researchers have found that the changes in the expression levels of upstream structural genes (PAL, C4H, and 4CL) simultaneously affected the synthesis of tannins, anthocyanins, and lignin (Schilmiller et al. 2009;Thévenin et al. 2011). The contents of phenolic acids and lignin were increased in transgenic N. tabacum and A. thaliana overexpressing CsHCT, while the content of flavonol glycosides was decreased (Chen et al. 2021). Meanwhile, some transcription factors affected the biosynthesis of lignin and flavonoids by regulating the expression of structural genes, and direct redistribution of carbon metabolism in plants (Bhargava et al. 2010;Deluc et al. 2008;Fornalé et al. 2010). In Populus L., the overexpression of MYB6 reduced secondary cell wall deposition and significantly promoted the accumulation of anthocyanins and procyanidins . Hence, the study of the flavonoid biosynthesis pathway in C. auriculatum tuberous root can not only clarify the accumulation mechanism of flavonoids but also be beneficial to explore the transformation rule of lignin content.
In this study, tuberous root phenotypes at three developmental stages (S1: root forming stage, S2: root expanding stage, and S3: harvest stage) were analyzed. We analyzed the enrichment pathways and expression levels of DEGs during the development of C. auriculatum tuberous root by RNA-Seq. Furthermore, through integrated analysis of phytohormones, lignin, flavonoids, and related genes involved in their metabolic pathways, the association between metabolites and genes was explained to provide valuable information for revealing the development characteristics of C. auriculatum tuberous root. This study aimed to elucidate the mechanism of tuberous root development and explore the regulatory mechanism of phytohormones, lignin, and flavonoids in C. auriculatum.

Plant materials
Annual C. auriculatum (genotype Yanwu 1406, a local variety of Yancheng) plant materials were grown (since 10 May 2021) in the experimental field of Yancheng Xinyang Agricultural Experiment Station (120.16 E,33.52 N), Yancheng, Jiangsu Province of China. C. auriculatum tuberous roots were sampled at three development stages, including the root forming stage (S1, on 15 Sept 2021), root thickening stage (S2, on 15 Oct 2021), and root harvest stage (S3, on 15 Nov 2021), referring to the agronomic details of Cassava (Yao et al. 2017). C. auriculatum materials, with three biological replicates for further analysis, were washed carefully and stored in the ultra-low temperature freezer of − 80 °C (Fig. 1). Part of the samples was then sent to Guangzhou Gene Denovo Biological Technology Co., Ltd for highthroughput transcriptomic sequencing, and the others were kept for physiological analysis.

Phytohormone extraction and content determination
1.0 g of C. auriculatum root samples were added with 2.0 mL of extraction solution (100% methanol) and grounded 1 3 with liquid nitrogen. After standing at 4 °C for 4 h, the supernatant was collected after centrifugation at 4000 rpm for 15 min. Then, after repeating the above step once more, the supernatant was combined for C18 solid phase extraction. Afterward, the supernatant was eluted with 100% methanol and purified using the 0.45 μm microporous filter for subsequent phytohormone content determination.
The endogenous contents of Indole-3-acetic acid (IAA), cytokinin (CK), and ethylene (ETH) were determined with phytohormone ELISA assay kits (Catalog No. QS42270 for IAA, Catalog No. QS441290 for CK, and Catalog No. QS49937 for ETH), respectively. And these ELISA kits were purchased from Beijing Qisong Biotechnology Co., LTD.

Phloroglucinol staining of lignified tissue and lignin content determination in C. auriculatum tuberous root
The chemical staining of lignified tissue of C. auriculatum tuberous root referred to the method described by Sun et al. (2021) with minor modifications. The samples were cut longitudinally into 2.0 mm slices from the midsection Fig. 1 Annual C. auriculatum root tubers at three development stages. A S1 represented root tuber forming stage, S2 represented root tuber thickening stage, and S3 represented root tuber harvest stage.
B The average root length and average root weight of C. auriculatum root tubers at three development stages 1 3 of tuberous roots. The slices were immersed in the 15.0% phloroglucinol solution for 5.0 min and then added with 3.0 mL of 36.0% hydrochloric acid for color reaction. The reddyed tuberous root slices were photographed with the EOS 1300D digital camera, the cell structure was observed and photographed using the MOTIC BA200 optical microscope.
Besides, the lignin content of C. auriculatum tuberous root was assayed with Lignin Content Detection ELISA Kit (Catalog number: BC4200, purchased from Beijing Solarbio Technology Co., Ltd.) and determined with UV spectrophotometry at A 280 .

Total flavonoids content determination of C. auriculatum tuberous root
The total flavonoids content of C. auriculatum tuberous root was assayed and analyzed according to the colorimetric method of sodium nitrite-aluminum nitrate-sodium hydroxide ). Briefly, a standard curve was prepared with rutin (Catalog number: R8170, purity ≥ 95.0%, purchased from Beijing Solarbio Technology Co., Ltd.) as the standard sample. 1.0 mL of C. auriculatum tuberous root powder (2.0 mg/mL) was added with 1.0 mL of sodium nitrite solution (50.0 mg/mL), 1.0 mL of aluminum nitrate solution (100.0 mg/mL), 4.0 mL of sodium hydroxide solution (200.0 mg/mL), and 3.0 mL of 37.0% ethanol. After static settlement for 15.0 min, the absorbance of the mixture was measured at A 510 .

mRNA extraction and high-throughput transcriptomic library sequencing
The mRNA of C. auriculatum tuberous root was extracted by TIANSeq mRNA Capture Kit (Catalog number: NR105-01, purchased from Beijing Tiangen Biotechnology Co., Ltd.).
To ensure the quality of sequencing result, the construction quality of the library was strictly checked, and the detection standards were as follows: agarose gel electrophoresis for the analysis of RNA integrity and DNA contamination, Nano-Photometer spectrophotometer for the determination of RNA purity, Qubit2.0 Fluorometer for the accurate quantification of RNA concentration, and Agilent 2100 bioanalyzer for the precise detection of RNA integrity.
Using random oligonucleotides as primers, the first strand of cDNA was synthesized in the M-MuLV reverse transcriptase system. Next, RNA degradation and the second strand of cDNA synthesis occurred. The purified double-stranded cDNA was end-repaired and connected to a sequencing adapter. AMPure XP beads were used to screen cDNAs and purify the PCR products. Finally, highthroughput transcriptomic libraries were obtained by Illumina HiSeq2500. In addition, the quality inspection of the transcriptomic library was evaluated according to the manual of DNA 1000 Assay Kit (Catalog number: 5067-1504, purchased from USA Agilent Technology Co., Ltd.).

Bioinformatics analysis of C. auriculatum transcriptome
The original reads were further filtered by fastp (version 0.18.0) to obtain high-quality clean reads (Chen et al. 2018). Through StringTie (version 1.3.1), the mapped reads of each sample were assembled (Pertea et al. 2015). The analysis of differentially expressed genes (DEGs) between different groups was performed with DESeq software (Love et al. 2014). Principal component analysis (PCA) was performed with R package gmodels (http:// www. rproj ect. org/ org/). Gene Ontology (GO) analysis (Ashburner et al. 2000) was performed by mapping DEGs to the Gene Ontology database (http:// www. geneo ntolo gy. org/). Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed with edgeR software (Robinson et al. 2010).

qRT-PCR analysis of DEGs from C. auriculatum transcriptome
To verify the reliability of C. auriculatum transcriptomic profiling data, qRT-PCR was conducted by SYBR® Premix Ex Taq™ II (Catalog number: RR820Q, purchased from Beijing TaKaRa Biotechnology Co., Ltd.) and performed on Thermal Cycler Dice™ Real-Time System Lite (Code No. TP700/TP760, purchased from Beijing TaKaRa Biotechnology Co., Ltd.). Sixteen genes related to starch synthesis, lignin synthesis, flavonoids synthesis, and auxin signaling, respectively, were selected for qRT-PCR assay, and CaAc-tin7 was used as the reference gene.
The relative expression levels of selected genes were calculated through the 2 −∆∆CT method (Livak and Schmittgen 2001). The primers used in this study were designed by NCBI primer design online tool (https:// www. ncbi. nlm. nih. gov/ tools/ primer-blast/) and synthesized by Anhui General Biol Co., Ltd. The nucleotide sequences and amplification characteristics of the primers had been shown in Supplementary Table 1.
In addition, to ascertain whether the correlation between qRT-PCR analysis data and FPKM values from RNA-Seq was significant, Ozone correlation analysis was applied with GraphPad Prism 8.0 software.

Statistical analysis
All experiments were performed within three or more biological replicates to eliminate errors. The data were statistically analyzed and graphed by GraphPad Prism 8.0 software, and the significant difference analysis was measured with One-Way or Two-Way ANOVA multiple comparisons (P value < 0.05 or 0.01).

Relationship analysis between C. auriculatum tuberous root samples
To investigate the developmental characteristics of C. auriculatum tuberous root, the samples at three development stages were observed and analyzed. It could be observed that the samples (S1 in Sept, S2 in Oct, and S3 in Nov) exhibited obvious morphological changes (Fig. 1A). Compared with S1, the average root length (14.12 cm of S2, 23.63 cm of S3) and root weight (22.59 g of S2, 54.49 g of S3) were significantly higher at later developmental stages (Fig. 1B).
In the meantime, the analysis results of PCA, sample clustering, violin plot, and correlation heat map showed that ( Fig. 2), there was significant variability among the samples collected at different developmental stages, and the three biological replicates of the samples were quite reproducible.

Quality assessment and basic information of C. auriculatum tuberous root transcriptome data
C. auriculatum tuberous root samples at three development stages, with three biological replicates, were used for the construction of nine transcriptomic libraries, including S1-Sept-1, S1-Sept-2, S1-Sept-3, S2-Oct-1, S2-Oct-2, S2-Oct-3, S3-Nov-1, S3-Nov-2, and S3-Nov-3. To reduce the interference caused by invalid data, data filtering was performed on the original data, and the obtained clean reads were as follows: 5,792,103,000 bp of S1-Sept-1, 6,402,806,400 bp of S1-Sept-2, 6,158,593,800 bp of S1-Sept-3, 5,621,652,900 bp of S2-Oct-1, 6,548,790,900 bp of S2-Oct-2, 7,550,866,800 bp of S2-Oct-3, 7,554,501,300 bp of S3-Nov-1, 7,505,005,500 bp of S3-Nov-2, 8,607,777,900 bp of S3-Nov-3. The balance analysis of base composition showed that the quality of the Sample clustering analysis, the abscissa represented the Euclidean distance between samples, and each branch represented a sample. C Violin plot, which showed the data density at any location for all genes in the sample, the white dots represented the median, and the black rectangles were the range from the lower quartile (Q1) to the upper quartile (Q3). D Correlation heat map, the abscissa and ordinate represented each sample, respectively, and the shade of color indicated the magnitude of the correlation coefficient between the two samples transcriptome library was high, which was beneficial to the accuracy of subsequent analysis (Supplementary Table 2).
Using our published full-length C. auriculatum transcriptomic database (NCBI accession number: SRX10889362) as the reference, sequence alignment and quantification were performed by RSEM, and the gene detection in each sample was counted (Supplementary Table 3). The reads on the isoform of our libraries were evenly distributed in each part of C. auriculatum gene, indicating that the randomness of mRNA interruption was satisfactory ( Supplementary Fig. 1). In addition, FPKM distribution and expression violin plot of all samples showed high gene expression abundance and high data density at corresponding sites in these transcriptome libraries ( Supplementary Fig. 2).

DEGs Enrichment analysis of C. auriculatum tuberous root during development
To investigate the changes at the molecular level during the development of C. auriculatum tuberous root, DEGs enrichment of the nine transcriptomic libraries was analyzed. Genes with FDR adjusted P < 0.05 and |log 2 FC| > 1 were considered significant, and the comparison between groups showed that there was a total of 14,944 DEGs (7,454 of up-regulated, 7,490 of down-regulated) between S1 vs. S2, 14,496 DEGs (9,962 of up-regulated, 4,534 of downregulated) between S2 vs. S3, 20,016 DEGs (12,890 of up-regulated, 7,126 of down-regulated) between S1 vs. S3 (Fig. 3A). In addition, difference comparison volcano plot (Fig. 3B), difference comparison cluster heatmap (Fig. 3 C), and Venn diagram showed 28,926 DEGs among the groups of S1, S2, and S3, indicating that the expression of these genes was affected by the development of C. auriculatum tuberous root.
Gene Ontology (GO) enrichment was analyzed to comprehensively characterize the properties of genes and gene products in C. auriculatum tuberous root during development. The comparison results (S1 vs. S2, S2 vs. S3, S1 vs. S3) had been shown in Supplementary Fig. 3, taking S1 vs. These results demonstrated that with the development of C. auriculatum tuberous root, more DEGs were enriched in 'metabolic process' and 'catalytic activity', which suggested that vigorous metabolism of natural compounds accompanied C. auriculatum root development. In addition, the p-value among three comparison groups showed that in the terms of 'cell wall', 'cell wall macromolecule metabolic process', 'cellular amino acid metabolic process', 'cellular amino acid catabolic process', 'cell wall macromolecule catabolic process', 'cellular carbohydrate biosynthetic process', 'cellular lipid metabolic process', 'cellular carbohydrate metabolic process', and 'intracellular protein transmembrane transport', the expression profiles of DEGs were significantly higher at S3 compared with that of S1 and S2 (Supplementary  Table 4), which indicated that along with C. auriculatum root development, cellular metabolic process and cell wall modification process were active.
Through KEGG pathway significant enrichment, the main biochemical metabolic pathways, and signal transduction pathways in C. auriculatum tuberous root involved in DEGs were determined. The comparison results showed that ( Supplementary Fig. 4), according to the number of DEGs, in the comparison of S1 vs. S2, the top 5 KEGG enrichment classifications included 'Metabolic pathways', 'Biosynthesis of secondary metabolites', 'Biosynthesis of amino acids', 'Carbon metabolism', and 'Ribosome'. In the comparison of S2 vs. S3, the top 5 enrichment classifications changed to 'Metabolic pathways', 'Biosynthesis of secondary metabolites', 'Plant hormone signal transduction', 'Spliceosome', and 'Biosynthesis of amino acid'. Combining the results of GO and KEGG enrichment analysis, we found that many DEGs favored the enrichment of 'Plant hormone signal transduction' and 'Phenylpropanoid biosynthesis', which DEGs enrichment analysis of C. auriculatum tuberous root during development, including A DEGs statistics among the groups of S1, S2, and S3. B difference comparison volcano plot among the groups of S1, S2, and S3. C difference comparison cluster heatmap between three biological replicates at three development stages. D Venn diagram of DEGs among the groups of S1, S2, and S3 1 3 indicated that phytohormones and phenylpropane metabolites played critical roles in the development of C. auriculatum tuberous root.
The crucial role of auxin has been proved in many tuber and storage root crops (Kondhare et al. 2021). Compared with S1, the IAA content of S2 significantly increased by 17.94%, while IAA content increased by 15.21% in S3 (Fig. 5A) CaARF19), whose change trends were generally similar to IAA content (Fig. 5B).
By interfering with auxin transport, CKs promoted PIN protein degradation and regulated root development (Jing and Strader 2019). In the present work, compared with S1, CK content showed a changing trend similar to IAA content, which significantly increased in S2 (increased by 70.26%) and decreased in S3 (Fig. 6A). Meanwhile, CK synthesis, transport, and signaling genes, including E3  CaARR15), whose FPKM values also followed the expected trend with elevated expression in S2 and decreased in S3 (Fig. 6B).
Studies have indicated that ethylene shows significant inhibition of the growth of primary roots (Qin et al. 2019a;Qin and Huang 2018). Here, it has been found that during tuber root development of C. auriculatum, ETH content decreased from 86.64 ng/kg FW (S1) to 42.21 ng/ kg FW (S3) (Fig. 7A) CaERF5), whose expression levels were consistently reduced (Fig. 7B).

Changes in lignin biosynthesis during tuberous root development of C. auriculatum
To explore the lignin biosynthesis mechanism of C. auriculatum tuberous root, C. auriculatum tuberous root slices at three developmental stages were excised for chemical staining and observed by light microscopy. The tuberous root slices stained with phloroglucinol-hydrochloric acid solution showed that, at the tuberous root forming stage (Fig. 8A, D), lignified cells were mainly distributed in the central xylem of the longitudinal section of C. auriculatum tuberous root. As the tuberous root gradually expanded and thickened, the expanded central xylem tore the surrounding endothelial cells, and the lignified cells were observed to form a ring-like structure (Fig. 8B, E). Finally, at the C. auriculatum tuberous root harvest stage, the core part was devoid of lignified cells, and the rays composed of lignified cells extended from the xylem to the phloem, ensuring the connection between the two transport systems (Fig. 8 C, F).

Enhancement of Flavonoids biosynthesis during tuberous root development of C. auriculatum
In this study, total flavonoids content had been determined to preliminarily explore the synthesis mechanism of flavonoids in C. auriculatum tuberous root. The data showed that (Fig. 10A), compared with the flavonoids content of S1 (6.23 mg/g), the flavonoids content was significantly higher in S2 (7.41 mg/g) and S3 (7.89 mg/g).
Furthermore, based on the data of the nine transcriptomic libraries, the FPKM values of structural genes involved in flavonoids biosynthesis pathway showed that (Fig. 10B), the relative expression levels of CaCHS, CaCHI,CaF3H,CaDFR,CaANS,CaFLS,CaANR,CaF3'H,CaLAR,and CaUFGT all showed varying degrees of an upward trend, especially CaCHI, CaFLS, CaLAR, and CaUFGT. Among these, compared with that of S1, the expression level of CaCHI1 (PB.26729.1) increased by 1.79-fold at S2 and 8.34fold at S3, and the expression level of CaCHI2 (PB.4236.1) increased by 2.06-fold at S2 and 7.05-fold at S3. Meanwhile, as compared to that of S1, CaFLS1 (PB.34108.1) expression increased by 4.04-fold at S2 and 8.22-fold at S3, CaFLS2 (PB.33009.1) expression increased by 4.89-fold at S2 and 8.65-fold at S3, CaLAR1 (PB.32826.1) expression increased by 3.52 folds at S2 and 8.58 folds at S3, and CaLAR2 (PB.31622.1) expression increased by 4.89 folds at S2 and 7.65 folds at S3. At the same time, compared to gene expression level at S1, CaUFGT1 (PB.29943.1) showed an expression increase of 7.79-fold at S2 and 12.33-fold at S3, and CaUFGT2 (PB.31354.1) showed an expression increase of 7.26-fold at S2 and 14.02-fold at S3, respectively. Interestingly, the variation trend in the expression profiles of these genes was both significantly and positively associated with the changes in the total flavonoids content of C. auriculatum tuberous root at different developmental stages.  CaPIN3), respectively, were selected for qRT-PCR validation, and CaAct7 (PB.15198.1) was used as a reference gene to control the variable.

Expression profile verification of C. auriculatum DEGs by qRT-PCR
According to the heatmap analysis of FPKM values from RNA-Seq data (Fig. 11A) and log 2 (ratio) values from qRT-PCR analysis (Fig. 11B), the expression of selected genes showed similar change trends. For example, in starch metabolism, compared with S1, the expression level of CaSS4 increased by 1.35 times in S2 and 3.86 times in S3, the expression level of CaSS3 decreased by 0.06 times in S2 and increased by 0.26 times in S3, the expression level of CaSBE1 decreased by 0.67 times in S2 and 0.86 times in S3, and the expression of CaSUS2 increased by 1.43 times in S2 and 2.44 times in S3. In the pathway of IAA signal transduction, compared with S1, the expression of CaIAA9 increased by 0.33 times in S2 and decreased by 0.13 times in S3, the expression of CaLAX2 increased by 0.30 times in S2 and decreased by 0.37 times in S3, the expression of CaARF3 increased by In addition, correlation analysis was performed with GraphPad 8.0 software. The results demonstrated that the R square between the FPKM value and qRT-PCR value was 0.7107, and the P value was less than 0.001, indicating that the values of qRT-PCR and FPKM were significantly correlated (Fig. 11 C), and the results of transcriptomic sequencing were credible.

Discussion
The developmental characteristics of C. auriculatum tuberous root revealed by RNA-Seq at the molecular level C. auriculatum tuberous root provides a variety of carbohydrates, proteins, and vitamins for human health. At the same time, the high levels of ascorbic acid and carotenoids Fig. 9 The lignin biosynthesis mechanism of C. auriculatum tuberous root during development. A The lignin content of C. auriculatum tuberous root at the stages of S1, S2, and S3. B Differential expres-sion profiles of structural genes involved in the lignin biosynthesis pathway of C. auriculatum tuberous root at the stages of S1, S2, and S3 make the root barks appear to be bright yellow , which could also be observed on the root bark of the samples we collected (Fig. 1A). During the root forming stage (S1), expanding stage (S2), and harvest stage (S3), the morphology of C. auriculatum tuberous roots changed obviously (Fig. 1A), which was embodied in average root length and average root weight (Fig. 1B).
Along with development progresses, multiple metabolic processes were activated in C. auriculatum tuberous root ( Supplementary Fig. 3). For example, the KEGG pathway of 'Starch and sucrose metabolism' was found to be enriched in a lot of DEGs ( Supplementary Fig. 4). Carbohydrate metabolism plays a vital role in the development of tuberous root in plants (Mehdi et al. 2019). Root crops and tuber crops use starch stored in the roots as an energy reserve to support growth and development (Kondhare et al. 2020). Here, the values of FPKM and qRT-PCR showed that the expression levels of genes encoding enzymes related to starch metabolism, including SS4 (Lu et al. 2018), SS3 (Gámez-Arjona and Mérida 2021) were up-regulated (Fig. 11), and the expression of SBE1 (Utsumi et al. 2021), the gene inhibiting digestion-resistant starch synthesis (Schwall et al. 2000), was down-regulated. Besides, related results had been partly confirmed in Ipomoea batatas (Utsumi et al. 2020).

The role of phytohormones in C. auriculatum tuberous root during development
Phytohormones, such as IAA (Kondhare et al. 2021), GA (Zhou et al. 2021), CK (Jang et al. 2015), and ETH , are actively involved in the changes of tuberous root morphology and the accumulation of starch and amino acids .
Different concentrations of auxin controlled multiple cell activities, including secondary wall formation, cell expansion, and division (Fischer et al. 2019). In Ipomoea batatas, it was speculated that auxin triggered the process of vascular cambium formation in the basal root system . At the same time, genes related to auxin transduction, such as ARF (Roosjen et al. 2018), SAUR , and GH3 (Zou et al. 2019), were found to regulate the function of auxin protein. In this study, in contrast to the root forming stage (S1), IAA accumulated at root expanding stage (S2), and there was its decay afterward at the root harvest stage (S3) (Fig. 5A). At the same time, this study also showed the same expression trend of IAA synthesis and signaling genes (Fig. 5B). Gao et al. had proved that IAA was specifically enriched in the formation and enlargement stages of subterranean organs (Gao et al. 2016). However, Fig. 10 The flavonoids biosynthesis mechanism of C. auriculatum tuberous root during development. A Total flavonoids content of C. auriculatum tuberous root at the stages of S1, S2, and S3. B Differen-tial expression profiles of structural genes involved in the flavonoids biosynthesis pathway of C. auriculatum tuberous root at the stages of S1, S2, and S3 IAA synthesis showed a decline at the root harvest stage of C. auriculatum, which suggested that accompanied by cell growth arrest, the last stage of tuber root development was independent of auxin effects (Kolachevskaya et al. 2015). A similar phenomenon also appeared in CK synthesis, transport, and signaling (Fig. 6), which was possibly due to CK concentrating in the meristematic region of the secondary xylem and in the phloem (Melis and van Staden 1985). As C. auriculatum root developed into the harvest stage (S3), the structures of the xylem and phloem tended to stabilize (Fig. 8), and the effect of CK would decrease or even be ignored.
It has been reported that ETH shows significant inhibition of primary root growth (Qin et al. 2019a). For example, overexpression of OsDOF15 positively regulated primary root elongation by limiting endogenous ethylene biosynthesis (Qin et al. 2019b). In potato, exogenous ethylene affected carbohydrate metabolism to inhibit tuber sprouting through its effects on carbohydrates and key enzymes (Dai et al. 2016). Here, we found that compared with S1, some genes related to ETH synthesis and signaling were down-regulated in S2 and S3, as well as the decreasing ETH content (Fig. 7). These results suggested that tuber root formation of C. auriculatum was sensitive to ETH, and C. auriculatum promoted the enlargement and conversion of tuberous roots into storage organs by suppressing ethylene biosynthesis (Carbonnel et al. 2020) and signaling .

The role of lignin in C. auriculatum tuberous root during development
As a type of tuberous root crop (Ma et al. 2015), the roots of C. auriculatum developed from adventitious roots and gradually expanded to form a storage organ. Different from woody plants, the root system of tuberous root crops is mainly composed of lignified cells that store primary metabolites, such as starch and amino acids (Villordon et al. 2014). In this study, as shown in Fig. 8, during the development of C. auriculatum tuberous root, lignified cells first concentrated in the central xylem of root tuber. Then, as the rate of cell division and expansion increased, more lignified cells were augmented to accommodate the expansion of the root system, and these cells formed a ring structure separated from the central core. At tuberous root harvest stage, the vertical section of the root showed many rays connecting the xylem and phloem of C. auriculatum tuberous root. This morphology was similar to the study of Manihot esculenta (Mehdi et al. 2019), and we speculated that these rays were vascular cells formed by secondary cambium, which helped to support the central root region.
Concordantly, we found that during tuberous root development, lignin content increased firstly and then decreased (Fig. 9A). Likewise, FPKM values of lignin biosynthesis genes (CaCCR , CaCAD, CaPOD, etc.) from RNA-Seq showed associated changes. Notably, when the phenylpropane metabolic pathway shifted to the lignin synthesis pathway, rate-limiting enzyme genes, including CaCSE and CaHCT, were down-regulated in S3 (Fig. 9B). As reported in Rehmannia glutinosa (Gu et al. 2021) and Ipomoea batatas , abnormal accumulation of lignin was not conducive to tuberous root enlargement. Therefore, these results demonstrated that at the early developmental stage of C. auriculatum tuberous root, the lignin content gradually increased with the expansion of the tuber, but then the expression of lignin synthesis genes was down-regulated, thereby promoting the transition from tuberous roots to storage roots (Singh et al. 2020).

The biosynthesis mechanism of flavonoids in C. auriculatum tuberous root during development
Flavonoids have been proven to be natural antioxidants with high content and strong activity in Cynanchum plants (Wu et al. 2019). The biosynthesis mechanism of flavonoids has been well-described in many plants (Yu et al., 2021), including critical intermediate metabolites and structural genes. Here, we found that during the development of C. auriculatum tuberous root, the content of total flavonoids increased from S1 to S3 stage (Fig. 10A). At the same time, the expression profiles of flavonoids biosynthesis genes in RNA-Seq and qRT-PCR showed a similar increasing trend (Figs. 10B and 11). The results suggested that during the development of C. auriculatum, flavonoids biosynthesis genes (CaPAL,CaC4H,Ca4CL,CaCHS,CaCHI,etc.) were activated, and their increased expression level induced flavonoid accumulation in C. auriculatum tuberous root. Besides, flavonoids have been proven to impart color to petals and promote root growth of Nicotiana tabacum (Park et al. 2020). It could be observed that compared with the morphology of S1, C. auriculatum petioles in S3 were purple-red, and the epidermis of tuberous root was bright yellow, which had an obvious correlation with the accumulation of flavonoids (Fig. 1A). This also suggested that the accumulation of flavonoids promoted the coloration of C. auriculatum organs.
Since the biosynthesis of lignin and flavonoids are both through the phenylpropane metabolic pathway, the changes in the contents of lignin and flavonoids in plants are closely related (Vogt 2010). In transgenic Arabidopsis, mutations in CCR1, CAD1, CAD2 resulted in reduced lignin in stems and anthers, and higher levels of flavonol glycosides, sinapoyl malate, and feruloyl malate (Thévenin et al. 2011). In Nicotiana tabacum, ectopic expression of AtCPC not only reduced the content of flavonoids but also up-regulated the expression of lignin biosynthesis genes, which in turn significantly enhanced lignin accumulation (Shi et al. 2022). As shown in Figs. 9B and 10B, the expression of CaPAL, CaC4H, and Ca4CL were up-regulated during the development of C. auriculatum tuberous root, but at S3 stage, it was found that the expression of lignin biosynthesis genes, particularly rate-limiting enzyme coding genes (CaCSE, CaHCT, and CaCCR ), was suppressed, and the expression of flavonoids biosynthesis genes continued to be enhanced. These results suggested that the decreased activities of the key enzymes in lignin synthesis diverted the precursors to flavonoids synthesis pathway, resulting in reduced lignin content and continuous flavonoids accumulation of C. auriculatum tuberous root.

Conclusion
Tuberous root development is a critical agronomic trait affecting in C. auriculatum. In this study, C. auriculatum tuberous root (S1 in Sept, S2 in Oct, and S3 in Nov) exhibited significant phenotypic differences in root length and root weight. Based on Illumina HiSeq2500, nine high-quality RNA-Seq libraries were constructed for transcriptomic analysis. Many DEGs were identified among the groups of S1, S2, and S3, which were mainly enriched for phytohormone signaling and phenylpropanoid metabolism. Phytohormone analysis indicated that IAA and CK jointly regulated root development, and C. auriculatum inhibited ETH synthesis and signaling to ensure the germination and expansion of tuberous roots. Chemical staining of lignin showed that in S1, lignified cells concentrated in the central xylem of root tuber. In S2, more lignified cells formed a ring structure separated from the central core. In S3, the vertical section of the root showed many rays connecting the xylem and phloem of C. auriculatum tuberous root. Besides, lignin content increased in S2 and decreased in S3, and the expression of lignin biosynthesis genes also showed a similar trend. With the development of C. auriculatum tuberous roots, the content of flavonoids gradually increased, and the expression of flavonoids biosynthesis genes also showed a similar trend. The expressions of rate-limiting genes suggested that C. auriculatum diverted the precursors to flavonoids by inhibiting lignin synthesis. Our results will aid in the investigations of phytohormones and phenylpropanoid biosynthesis in C. auriculatum tuber root development and may be helpful to improve the application value of C. auriculatum.

Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.