UHPLC-MS/MS, transcriptomic, and genomic analyses unravel the molecular mechanism on melatonin biosynthesis in Tartary buckwheat sprouts

Abstract

Melatonin, a neurohormone and an antioxidant agent, has been identified in plant-derived foods, such as vegetables, fruits, medicinal herbs, and wine. However, there is no more report about melatonin and its biosynthesis in plant sprouts. This study determined melatonin contents in Tartary buckwheat (TB, Fagopyrum tartaricum L.) sprouts via ultra-high-performance liquid with tandem mass spectrometry (UHPLC-MS/MS) analysis, and the mechanisms underlying melatonin biosynthesis were investigated by integrating genome re-sequencing and transcriptome analyses. Melatonin contents in the sprouts of 16 TB varieties ranged from 0 to 0.187 ng g⁻¹ fresh weight (FW). Specifically, the TB varieties Heifeng No.1 (HF1), Chiku (CK), and Zhaoku No. 1 (ZK1) had relatively high melatonin contents, reaching 0.187, 0.183, and 0.125 ng g⁻¹ FW, respectively, whereas no melatonin was detected in Yunqiao No. 1 (YQ1), YQ2, and Xiqiao No. 3 (XQ3). 330 differentially expressed genes (DEGs) were identified by comparative transcriptome analysis between the high- and non-melatonin varieties. Ten putative melatonin biosynthetic genes were identified, among which FtCYP96A15, FtCYP81D1, and FtCYP82C4 may be the key regulatory genes. Twelve transcription factors such as FtABR1, FtWRKY41, and FtMYB80 may regulate melatonin synthesis. Furthermore, genes involved in auxin, abscisic acid, and cytokinin signaling pathways exhibited more active expression in high-melatonin varieties. Additionally, genome re-sequencing analysis identified 58 genotype-specific DEGs with sequence variations, including RLKs, FtCYP71A26, FtUGT76B1, FtFAAH, and FtSEOC. This work provides valuable insights into the mechanisms underlying melatonin biosynthesis in TB sprouts and proposes candidate genes for breeding high-melatonin TB varieties.

Highlights

• •Melatonin contents in the sprouts of 16 Tartary buckwheat varieties ranged from 0 to 0.187 ng g⁻¹ FW.
• •330 differentially expressed genes were identified between high- and non-melatonin varieties.
• •FtABR1, FtWRKY41, and FtMYB80 transcription factors may regulate melatonin biosynthesis.
• •The activation of auxin, abscisic acid, and cytokinin signaling may induce melatonin accumulation.
• •58 genes with genomic variations and differential expression may involve in melatonin biosynthesis.

1. Introduction

Melatonin (N-acetyl-5-methoxy-tryptamine), a biogenic indolamine, is present in both animals and plants (Fan et al., 2018). In 1958, melatonin was isolated from the bovine pineal gland (Lerner et al., 1958). Melatonin, a neurohormone and an antioxidant agent, exerts diverse physiological activities that enhance cellular and organ health, including sleep and circadian modulation, immunoresponsiveness, regulating vascular functions, cardiac disease treatment, and COVID-19 prevention (Cheng et al., 2021; Muto et al., 2016; Rodriguez-Naranjo et al., 2011; Ziegler et al., 2023). Melatonin performs physiological function by activating high-affinity G-protein-coupled receptors (MT1 and MT2), which are the drug targets for regulating circadian phase (Stauch et al., 2019; Stein et al., 2020). With the growing demand for healthy food, plant-derived melatonin has emerged as an important option for promoting human health through daily diets. Plant-derived melatonin is a natural source of bioactive compound for humans and is widely found in foods, including vegetables, fruits, medicinal herbs, tea, wine, nuts, and sprouts (Cheng et al., 2021; Garcia-Parrilla et al., 2009; Verde et al., 2022). Furthermore, plant-derived melatonin exerts the same regulatory effects on humans as animal-sourced melatonin.

In 1995, melatonin was first detected in plants, including tomato (Solanum lycopersicum L.), banana (Musa acuminata Lour.), cucumber (Cucumis sativus L.), and tobacco (Nicotiana tabacum L.) (Dubbels et al., 1995). Since then, melatonin has been identified in hundreds of plant-derived foods, with contents ranging from 0.002 to 34,000 ng g⁻¹ (Cheng et al., 2021). Meanwhile, the melatonin biosynthetic pathway has been elucidated in Arabidopsis thaliana, tomato, and rice (Oryza sativa L.). In plants, melatonin biosynthesis initiates from tryptophan. Tryptophan is converted to tryptamine with the help of tryptophan decarboxylase (TDC). And then, tryptamine is hydroxylated by tryptamine 5-hydroxylase (T5H) to form serotonin. Serotonin is subsequently converted to N-acetylserotonin mediated by serotonin N-acetyltransferase (SNAT). Additionally, serotonin can be converted to 5-methoxytryptamine by affeic acid O-methyltransferase (COMT) or N-acetylserotonin methyltransferase (ASMT). Finally, COMT/ASMT and SNAT catalyze the production of melatonin from 5-methoxytryptamine and N-acetylserotonin, respectively (Liu et al., 2022). However, the melatonin biosynthetic pathway in non-model plants (e.g., vegetables and coarse cereals) remains unclear.

Nowadays, the consumption of plant sprouts, a category of healthy ready-to-eat foods, has become part of human daily diets (Aloo et al., 2021). Sprouts refer to the young buds, leaves, or shoots germinated from plant seeds or other vegetative organs. They are regarded as high-quality functional foods because they exhibit numerous biological activities, including antioxidative, anti-inflammatory, antidiabetes, anticancer, and antiviral activities (Geng et al., 2022). Sprouts germinated from crop seeds have pleasant taste and abundant nutritional components, and thus have been used as raw materials for producing bread, noodles, and yoghurt (Ikram et al., 2021). In addition, sprouts are rich in vitamins, phenolic compounds, flavonoids, γ-Aminobutyric acid, trace minerals, and amino acids (Geng et al., 2022; Ikram et al., 2021). Recent studies have reported the detection of melatonin in various sprouts. In legume sprouts, 1.0, 0.4, 18.8, 26.6, 29.7, and 34.48 ng g⁻¹ dry weight (DW) melatonin were detected in kidney beans (Phaseolus vulgaris L.), lentils (Lens culinaris L.), Vigna angularis, chickpea (Cicer arietinum L.), mung bean (Vigna radiata L.), and soybean (Glycine max L.) sprouts, respectively (Aguilera et al., 2014; Nontasan et al., 2022; Ratha et al., 2022). Therefore, legume sprouts are potential natural sources of melatonin for developing functional food products (Nontasan et al., 2022). In mustard (Brassica juncea L.) sprouts, 22.78 ng g⁻¹ fresh weight (FW) melatonin was detected in the four-day-old germinated sprouts, and methyl jasmonate (MeJA) treatment increased melatonin content by 11.43-fold (Yin et al., 2023). However, there is no report on endogenous melatonin in the sprouts of other plant species. Meanwhile, the physiological and molecular mechanisms underlying melatonin biosynthesis in sprouts remain unclear.

Tartary buckwheat (TB, Fagopyrum tartaricum (L.) Gaertn) is an important edible and medicinal crop widely distributed in the world (Li et al., 2025). TB seeds are commonly used for the production of functional food. The sprouts germinated from TB seeds are also consumed by humans and used as raw material for other food products (Suzuki et al., 2021). TB sprouts are abundant in flavonoids, polyphenols, rutin, γ-aminobutyric acid, amino acids, and organic acids, endowing them with beneficial biological effects on human health (Dong et al., 2023). In addition, our previous study identified 62 disease-resistant metabolites in TB sprouts via UPLC-MS analysis, which are associated with the prevention of various conditions including cancer, thrombosis, hypertension, diabetes, atherosclerosis, and cardiovascular disease (Liu, You, et al., 2023). Several bioactive components such as Qing Hau Sau, 1-deoxynojirimycin, and miglitol were also detected in TB sprouts (Liu, You, et al., 2023). These findings suggest that TB sprouts are rich reservoir of valuable functional substances, and more important bioactive components in TB sprouts remains to be explored. For instance, melatonin has not yet been identified in TB sprouts, and the genes involved in melatonin biosynthesis remain unknown.

In this study, melatonin content was determined in the sprouts of 16 TB varieties using ultra-high-performance liquid with tandem mass spectrometry (UHPLC-MS/MS) analysis. In order to identify the genes associated with melatonin biosynthesis in TB sprouts, high- and non-melatonin varieties were selected for integrated genomic re-sequencing and transcriptome analyses. This study will not only provide theoretical basis for investigating molecular mechanism on melatonin biosynthesis in TB sprouts, but also may facilitate the development of high-melatonin TB varieties.

2. Materials and methods

2.1. Plant materials

16 TB varieties, namely Chuanqiao No. 2 (CQ2), Liuku No. 3 (LK3), Weiku (WK), Yunqiao No. 1 (YQ1), Yunqiao No. 2 (YQ2), Zhaoku No. 1 (ZK1), Hanxuan (HX), Heifeng No. 1 (HF1), Qianku No. 4 (QK4), Tongliao (TL), Xiqiao No. 2 (XQ2), Xiqiao No. 3 (XQ3), Chiku (CK), Shanxikuqiao (SXKQ), Fenghuang (FH), and Dianning No.1 (DN1), were used in this study. The detailed information of these varieties was listed in Supplementary Table S1. Mature, plump seeds with uniform size were selected. Seeds were surface-sterilized with 0.7% sodium hypochlorite for 30 min and subsequently rinsed thoroughly with double-distilled water (ddH₂O) (Liu et al., 2025). Thereafter, the sterilized seeds were placed in floating culture plates containing full-strength Hoagland's nutrient solution for germination. The culture conditions were set as follows: temperature maintained at 25 °C, relative humidity >80%, and a 16/8 h (light/dark) photoperiod. Hoagland's nutrient solution was refreshed every three days. Ten days after germination, the sprouts were harvested by cutting the roots and immediately frozen for subsequent analysis.

2.2. Determination of melatonin content in sprouts

1 g of fresh sprout sample was homogenized in 5 mL of methanol and subjected to ultrasonic extraction at low temperature in the dark for 30 min. The homogenate was centrifuged, and the supernatant was collected. A Termovap Sample Concentrator was used to dry the supernatant, and the resulting residue was redissolved in 0.2 mL of methanol. The extract was filtered through a 0.22 μm filter membrane prior to instrumental analysis. Melatonin content was determined using an Agilent 1290 Infinity III UHPLC system coupled to a Qtrap 6500 tandem mass spectrometer (UHPLC-MS/MS). The parameters for liquid phase conditions were as follows: chromatographic column, Agilent poroshell 120SB-C18 reversed phase chromatographic column (2.1 × 150, 2.7 μm); column temperature, 30 °C; mobile phase, A: B = water/0.1% formic acid: methanol; elution gradient, A = 20% at 0-2 min, A = 20% at 2–5 min, A = 80% at 5–12 min; flow rate, 0.3 mL/min; injection volume, 5 μL (Lin et al., 2019). The parameters for mass spectrometry (MS/MS) conditions were as follows: spray voltage, 3500 v; air curtain gas, 25 psi; atomizing gas pressure, 55 psi; auxiliary gas pressure, 65 psi; atomization temperature, 450 °C. Three biological replicates were conducted for each experiment.

2.3. RNA extraction and Illumina sequencing

Based on the melatonin contents of different TB varieties, four varieties (CK, HF1, YQ1, and YQ2) were selected for transcriptome analysis. Three biological replicates were conducted for each variety. Total RNA was extracted following the manufacturer's instructions for the RNA extraction kit (Tiangen Biochemical Technology Co., Ltd., Beijing, China), and residual genomic DNA was digested with DNase I. RNA purity, concentration, and integrity were determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, Massachusetts, USA), a Qubit® 2.0 fluorometer (Life Technologies, USA), and an Agilent Bioanalyzer 2100 system (Agilent Technologies Inc., California, USA), respectively. A minimum of 1 μg of total RNA per sample was used for cDNA library construction. The quality and concentration of each library were assessed using an Agilent 2100 Bioanalyzer. These libraries were paired-end sequencing using next-generation sequencing technology on the Illumina platform (Liu et al., 2025). The raw sequencing data were deposited in the NCBI Short Read Archive (SRA) database (accession no.: PRJNA1273265).

2.4. Transcriptome data processing and analysis

Following sequencing on the Illumina HiSeq 2500 platform, the sequencing images were converted into raw reads. Subsequently, low-quality reads were filtered out. Clean reads from each library were aligned to the TB reference genome (Zhang et al., 2017). In order to obtain comprehensive gene function information, all identified genes were against the Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology (GO), eggNOG, Swissprot, and NR databases. DESeq2 software was used to analyze the differentially expressed genes (DEGs) based on read count data. A significance threshold of |log2FoldChange| >1 and P-value <0.05 was applied for DEG screening. Enrichment analysis of DEGs was performed using GO database (http://www.geneontology.org/) and KEGG database (https://www.kegg.jp/).

2.5. Integrating transcriptome and genomic re-sequencing analysis

The genome re-sequencing data of YQ1 and YQ2 have been previously reported (Qiu et al., 2023). In this study, genome re-sequencing was additionally performed for CK and HF1. The methods for DNA extraction, library construction, sequencing, and data analysis were as described in our previous study (Liu, Qiu, et al., 2023). The raw sequencing data was deposited in the NCBI Short Read Archive (SRA) database (accession no.: PRJNA1320530). Genes with sequence variations were identified from the re-sequencing data. Finally, the DEGs with genomic variations were screened by integrating transcriptome and genomic re-sequencing data.

2.6. Real-time quantitative RT-PCR (qRT-PCR) analysis

Total RNA was used as a template to synthesize the first strand cDNA, following the manufacturer's instructions for the PrimerScript™ FAST RT reagent Kit with gDNA Eraser (TAKARA, Japan). qRT-PCR analysis was performed according to the protocol of TB Green® Premix Ex Taq™ II FAST qPCR kit (TaKaRa, Japan) (Liu et al., 2021). The gene expression was calculated using 2^-ΔΔCt method, and FtActin7 gene was used as the internal reference gene. Three biological replicates were set for each experiment. The primers were listed in Supplementary Table S2.

2.7. Statistical analysis

The data were analyzed by using Excel 2021 and SPSS 17.0 software, and the results were showed as mean ± standard deviations (SDs). Venn analysis and gene expression heatmap were performed using TBtools program (https://github.com/CJ-Chen/TBtools). Results were visualized by using TBtools program or GraphPad Prism 10.0.

3. Result and discussion

3.1. Detection of melatonin content in different TB varieties

In this study, melatonin content in the sprouts of 16 TB varieties was detected, with values ranging from 0 to 0.187 ng g⁻¹ FW (Fig. 1). Among these varieties, the melatonin contents in HF1, CK, and ZK1 reaching 0.187, 0.183, and 0.125 ng g⁻¹ FW, respectively, whereas no melatonin was detected in YQ1, YQ2, and XQ3. These results indicated that melatonin accumulation varied significantly among the sprouts of different TB varieties. Given that the moisture content of sprouts exceeds 90%, the melatonin content in HF1, CK, and ZK1 was higher than kidney beans and lentils sprouts, but lower than that in Vigna angularis, soybean, chickpea, and mung bean sprouts (Aguilera et al., 2014; Nontasan et al., 2022; Ratha et al., 2022). Compared with other vegetables, the melatonin content in TB sprouts was lower than that in radish (Raphnus sativus, 3.5 ng g⁻¹ FW), tomato (4.11 ng g⁻¹–114.52 ng g⁻¹ FW), garlic (Allium sativum L., 0.59 ng g⁻¹ FW), cauliflower (Brassica oleraceae var. botrytis, 0.82 ng g⁻¹ FW), turnip (Brassica rapa, 0.50 ng g⁻¹ FW), and cabbage (Brassica oleraceae var. capitata, 0.31 ng g⁻¹ FW) (Badria, 2002; Byeon et al., 2015; Stürtz et al., 2011), but higher than that in Indian mustard (Brassica juncea L., 0.005 ng g⁻¹ FW), welsh onion (Allium fistulosum, 0.09 ng g⁻¹), onion (Allium cepa L., 0.03 ng g⁻¹), and cucumber (Cucumis sativus L., 0.02 ng g⁻¹) (Byeon et al., 2015; Hattori et al., 1995).

Fig. 1.

Melatonin contents in the sprouts of different TB varieties. Melatonin contents in the sprouts of 16 TB varieties were determined by UHPLC-MS/MS analysis. Data are means ± SDs (n = 3). Means were compared by Duncan's test. nd indicates the melatonin was not detected.

3.2. Transcriptome sequencing of different TB varieties

In order to explore the potential genes involved in melatonin synthesis in TB, two melatonin-free varieties (YQ1 and YQ2) and two high-melatonin varieties (HF1 and CK) were selected for transcriptome sequencing analysis. A total of 12 transcriptome libraries were constructed, corresponding to three biological replicates for each variety (CK: CK-1/2/3; HF1: HF1–1/2/3; YQ1: YQ1–1/2/3; YQ2: YQ2–1/2/3). Transcriptome sequencing generated 591 Mb of clean reads. For all samples, the Q20 value exceeded 97% and Q30 value exceeded 95% (Supplementary Table S3). Clean reads from each transcriptome library were aligned to the TB reference genome (Zhang et al., 2017), with a sequence alignment rate of over 97% (Supplementary Table S3). The reads per kilobase of transcript per million mapped reads (RPKM) method was used to quantify gene expression level.

3.3. Identification of DEGs between YQ1/YQ2 and HF1/CK

Transcriptome sequencing results showed that 23, 581 genes were expressed across the four varieties. The DEGs were identified by comparing the transcriptomes of high-melatonin varieties and non-melatonin varieties. Pairwise comparisons revealed that HF1 had 2, 606 DEGs relative to YQ1 and 2, 040 DEGs relative to YQ2; similarly, CK had 2, 636 DEGs relative to YQ1 and 1, 514 DEGs relative to YQ2 (Fig. 2A). Venn analysis of these four comparisons identified 330 common DEGs (Fig. 2A; Supplementary Table S4), which serve as candidate genes for investigating the regulatory mechanisms on melatonin synthesis.

Fig. 2.

Identification of the DEGs between high- and non-melatonin varieties. (A) Venn diagram showing the overlapping DEGs among the four comparisons. (B) Heatmap of 330 common DEGs. Gene expression heatmap were performed using TBtools program. Color scale on the heatmap indicates the degree of expression: red, high expression; blue, low expression. (C) Gene ontology enrichment analysis of the 330 common DEGs. (D) KEGG enrichment analysis of the 330 common DEGs. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Expression patterns of the 330 DEGs showed that these genes were clustered into five different expression patterns (Fig. 2B). Specifically, 109 genes exhibited significantly higher expression levels in HF1 and CK than in YQ1 and YQ2, whereas 210 genes showed significantly higher expression in YQ1 and YQ2 than in HF1 and CK. Seven genes were significantly up-regulated in CK relative to the other three varieties, whereas no significant expression differences of these genes were observed among YQ1, YQ2, and HF1. Additionally, FtPinG0005927100.01 and FtPinG0009269800.01 had significantly higher expression in YQ1 and YQ2 than in CK, but lower expression than in HF1. Conversely, FtPinG0000401800.01 and FtPinG0001436900.01 were significantly more highly expressed in HF1 and CK than in YQ2, but less highly expressed than in YQ1.

GO enrichment analysis showed that the 330 DEGs were mainly enriched in the following functional categories and pathways: protein binding, ATP binding, protein kinase activity, oxidation-deduction process, DNA binding, transmembrane transport, protein phosphorylation, transcriptional regulation, transcription factor activity, and other pathways (Fig. 2C). KEGG enrichment analysis showed that these genes were mainly enriched in plant hormone signal transduction, plant-pathogen interaction, circadian rhythm, amino sugar and nucleoside sugar metabolism, glycolysis, and MAPK signaling (Fig. 2D).

3.4. Expression analysis of the 330 common DEGs

Among the 330 DEGs, the genes belonging to the TDC, SNAT, ASMT, or COMT families were not identified. However, ten homologous genes of T5H were found, and these genes also belonging to the cytochrome P450 family. These ten genes were designated as FtCYP71A9, FtCYP81D1, FtCYP82C4, FtCYP71A26–1/2/3/4, FtCYP71B9, FtCYP96A15, and FtCYP711A1, respectively. Gene expression analysis showed that the expression of FtCYP96A15 and FtCYP81D1 were significantly higher in CK and HF1 than that in YQ1 and YQ2, whereas the expression level of FtCYP82C4 in CK was significantly higher than that in the other three varieties (Fig. 3A). In contrast, the other seven genes were highly expressed in YQ1 and YQ2, but lowly expressed in CK and HF1. These results indicated that FtCYP96A15, FtCYP81D1, and FtCYP82C4 may act as the key genes involved in the synthesis of melatonin in TB sprouts.

Fig. 3.

Identification of the key genes involved in melatonin biosynthesis in TB sprouts. (A-C) Expression analysis of the genes encoding CYP and transcription factors, as well as hormone signaling genes. Gene expression heatmap were performed using TBtools program. Color scale on the heatmaps indicates the degree of expression: red, high expression; blue, low expression. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Currently, research into the regulatory mechanisms underlying melatonin biosynthesis has attracted considerable attention. Based on the GO enrichment analysis of the DEGs, several transcription factors (TFs) may be involved in the regulating melatonin synthesis (Fig. 3B). A total of 12 genes encoding TFs were screened from the 330 DEGs, namely FtTGA4, FtABR1, FtPIF1/3, FtPCL1, FtbHLH104, FtTCP9, FtNAC29, FtMYB80, FtHHO2, FtCRF3, and FtWRKY41. The expression heatmap revealed that the expression levels of FtABR1, FtNAC29, FtWRKY41, FtTCP9, and FtMYB80 in CK and HF1 were significantly higher than those in YQ1 and YQ2, while the remaining seven genes showed higher expression in YQ1 and YQ2 (Fig. 3B). Previous studies have demonstrated that TFs can regulate melatonin synthesis in plants. For instance, in tea plants, an ERF TF, CsERF21-l, promotes tryptophan and melatonin biosynthesis (Zhou et al., 2025). In herbaceous peony (Paeonia lactiflora Pall.), an APETALA2/ethylene-responsive element-binding factor (AP2/ERF) TF, PlTOE3, enhances melatonin production by activating the PlTDC expression (Zhang et al., 2022). In apple (Malus domestica), MdWRKY17 promotes melatonin biosynthesis by mediating the transcriptional activation of MdASMT7 expression (Song et al., 2023). Collectively, these data identify FtABR1, FtWRKY41, and FtMYB80 as candidate TFs for further investigating the transcriptional regulatory mechanism on melatonin synthesis in TB sprouts.

KEGG enrichment analysis showed that many genes enriched in hormone signal transduction, which suggested that hormones may involve in regulating melatonin synthesis in TB sprouts. A previous study in rice suggested that engineering hormonal crosstalk can enhance melatonin content (Cui et al., 2025). In this study, a total of ten genes involved in hormone signaling were found from 330 common DEGs. Among them, six genes were involved in the auxin (IAA) signaling pathway, namely FtGH3.1, FtSAUR50/67, FtORR9, FtTIR1, and FtILL6. Two genes, FtCYP707A2 and FtP2C44, were involved in abscisic acid (ABA) signaling, and another two genes (FtLOG3 and FtCKX6) were involved in cytokinin (CTK) signaling. Gene expression analysis showed that FtGH3.1, FtTIR1, FtSAUR50/67, FtCKX6, and FtP2C44 were highly expressed in CK and HF1, with significantly higher levels than those in YQ1 and YQ2 (Fig. 3C). Notably, although the expression of FtCYP707A2 was lower in CK and HF1 than in YQ1 and YQ2, the ABA 8-hydroxylase encoded by this gene is responsible for ABA degradation and acts as a negative regulator in ABA pathway (Balarynová et al., 2023). These results indicated that the IAA, ABA, and CTK signaling pathways were more active in CK and HF1, which may contribute to the regulation of melatonin synthesis. Previous studies have demonstrated the regulatory role of hormones in melatonin biosynthesis. In watermelon (Citrullus lanatus), ABA induced melatonin accumulation, while inhibition of ABA synthesis blocked this process (Guo et al., 2021). In hickory (Carya cathayensis Sarg.), the ABA-responsive protein CcAZF2 could activate melatonin biosynthesis pathway genes, CcTDC1 and CcASMT1 (Chen et al., 2021). For CTK, exogenous CTK treatment increased the expression of melatonin synthetic gene AtASMT in Arabidopsis thaliana (Bychkov et al., 2023). In wheat, increased CTK content enhanced melatonin accumulation by up-regulating SNAT expression (Shamloo-Dashtpagerdi et al., 2025). Additionally, in peach fruit (Prunus Persica Batsch cv. Hujing), IAA application increased melatonin content in harvested peaches (Zhou et al., 2023). Collectively, these findings suggest that the activation of ABA, CTK, and IAA signaling may induce melatonin accumulation.

To validate the transcriptome results, nine genes were selected from the common DEGs for qRT-PCR validation, including six CYP genes and three TF encoding genes. The transcriptome data of these nine genes were basically consistent with the qRT-PCR results (Fig. 4).

Fig. 4.

Validation of expression profile through qRT-PCR. The expression of nine genes was validated using qRT-PCR against their expression from transcriptomic data. The relative expression levels of gene were calculated using 2^–ΔΔCt.

3.5. Integrating transcriptome and genomic analyses to identify the genes involved in melatonin biosynthesis

In this study, genomic differences among four TB varieties were analyzed by genome re-sequencing. The re-sequenced genome data of YQ1 and YQ2 have been previously reported, with 2, 179 genomic varied genes identified in YQ1 and 2, 545 in YQ2 (Qiu et al., 2023). The genome of CK and HF1 were also re-sequenced. Compared with the TB reference genome, CK had 702, 681 single-nucleotide polymorphisms (SNPs), 244, 411 insertions/deletions (InDels), 2, 155 copy number variations (CNVs), and 1, 512 structural variations (SVs), while HF1 had 827, 259 SNPs, 262, 445 InDels, 1, 996 CNVs, and 1, 454 SVs (Fig. 5). A total of 3, 708 genes with genomic variations were identified in CK, and 4, 504 in HF1 (Fig. 6). Across the four varieties, 5, 640 genomic varied genes were generated, including 1, 651 common varied genes and 3, 989 unique varied genes (Fig. 6A). Apart from the 1, 651 common varied genes, the expression patterns of 3, 989 varied genes across four varieties was analyzed. Venn diagram analysis showed that 58 genomic varied genes showed differential expression levels among the sprouts of four varieties (Fig. 6B; Supplementary Table S5). The expression heatmap showed that the expression of 21 genes in HF1 and CK were higher than those in YQ1 and YQ2, whereas the remaining genes showed the opposite expression pattern (Fig. 6C). Of these genes, four genes encoding receptor-like protein kinase (RLK) were identified, namely CRK10, CRK11, LRR-RLK, and FLR1 (Supplementary Table S5). A previous study in apple demonstrated that the receptor-like cytoplasmic kinase MdPBL34 phosphorylates MdSNAT5, thereby enhancing melatonin content (Yan et al., 2025). Thus, these RLK encoding genes may be involved in regulating melatonin synthesis in TB sprouts.

Fig. 5.

The distribution of genetic variants of two TB varieties was delineated using the Circos program. Rings (from outer to inner) represent gene, SNP density, indel density, and CNV density between Chiku/Heifeng No. 1 and reference genome.

Fig. 6.

Identification of the genotype-specific DEGs with sequence variations. (A) Venn diagram showing the genes with sequence variations among four genotypes. (B) Venn diagram showing the genotype-specific varied DEGs. (C) Heatmap analysis of the 58 genotype-specific varied DEGs. Color scale in the heatmap indicates the degree of expression: red, high expression; blue, low expression. (D) The correlations among these genotype-specific varied DEGs. The pearson's correlation coefficient was applied for correlation analysis. The correlation index >0.9 or < −0.9 was set as the cut-off. The results were performed by Cytoscape software. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In addition, correlation analysis was performed for the 58 genes with genomic variations. Among them, 57 genes showed high correlation with the expression of more than one gene. Some genes, such as FtCYP71A26, FtUGT76B1, FtFAAH, FtSEOC, FtFPF1, FtBPG2, FtPLAT2, FtRIP30, Thioredoxin, FtXI-E, FtPinG0005435100.01, FtPinG0007096100.01, FtPinG0007096100.01, FtPinG0005829000.01, and FtPinG0006244300.01, may serve as the potential genes contributing to the difference in melatonin accumulation among different TB varieties (Fig. 6D). Although no report has been published on the function of these genes in regulating melatonin biosynthesis, it is worthy of investigating the underlying mechanisms in future studies.

4. Conclusion

In conclusion, this study detected melatonin content in TB sprouts ranging from 0 to 0.187 ng g⁻¹ FW by UHPLC-MS/MS analysis. TB sprouts are potential dietary source of melatonin. Transcriptome comparison between high- and non-melatonin TB varieties identified FtCYP96A15, FtCYP81D1, and FtCYP82C4 genes may act as the key metabolic genes regulating melatonin synthesis in TB sprouts. Meanwhile, TFs such as FtABR1, FtWRKY41, and FtMYB80, along with activated IAA, ABA, and CTK signaling pathways, were implicated in transcriptional and hormonal regulation of melatonin production. Furthermore, integrated genome re-sequencing and transcriptomic analyses identified 58 genes with genomic variations and differential expression, which are the potential genes contributing to the difference in melatonin accumulation among different TB varieties. Collectively, these findings expand the understanding of melatonin biosynthesis and regulatory networks in TB sprouts, providing candidate genes for genetic improvement aimed at enhancing melatonin content and nutritional quality in TB. Future research should focus on investigating the function of melatonin biosynthetic genes and TFs through transgenic approaches, clarifying hormone crosstalk involved in melatonin metabolism, and utilizing the identified genomic variations to breed high-melatonin TB.

CRediT authorship contribution statement

Changying Liu: Writing – review & editing, Writing – original draft, Methodology, Funding acquisition, Data curation. Junjie Yin: Investigation, Data curation. Daiying Xu: Investigation, Data curation. Yaoxuan Zou: Methodology, Formal analysis. Xueling Ye: Methodology, Formal analysis. Liangzhen Jiang: Methodology, Funding acquisition. Yan Wan: Investigation, Formal analysis. Dabing Xiang: Project administration, Funding acquisition. Bangxing Zou: Writing – review & editing, Writing – original draft, Formal analysis, Data curation.

Declaration of competing 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.

Appendix A. Supplementary data

Supplementary material

Data availability

The data that has been used is confidential.