UV-B promotes flavonoid biosynthesis in Ginkgo biloba by inducing the GbHY5-GbMYB1-GbFLS module

Abstract Ginkgo biloba (ginkgo) leaves have medicinal value due to their high levels of secondary metabolites, such as flavonoids. We found that the flavonoid content in ginkgo leaves increases significantly at high altitudes (Qinghai-Tibet Plateau). Considering that high UV-B radiation is among the key environmental characteristics of the Qinghai-Tibet Plateau, we carried out simulated UV-B treatments on ginkgo seedlings and found that the flavonoid content of the leaves increased significantly following the treatments. Combined with results from our previous studies, we determined that the transcription factor GbHY5 may play a key role in responses to UV-B radiation. Overexpression of GbHY5 significantly promoted the accumulation of flavonoids in both ginkgo callus and Arabidopsis thaliana. Furthermore, yeast two-hybrid and real-time quantitative PCR showed that GbHY5 promoted the expression of GbMYB1 by interacting with GbMYB1 protein. Overexpression of GbMYB1 in ginkgo callus and A. thaliana also significantly promoted flavonoid biosynthesis. GbFLS encodes a key enzyme in flavonoid biosynthesis, and its promoter has binding elements of GbHY5 and GbMYB1. A dual-luciferase reporter assay indicated that while GbHY5 and GbMYB1 activated the expression of GbFLS individually, their co-expression achieved greater activation. Our analyses reveal the molecular mechanisms by which the UV-B-induced GbHY5-GbMYB1-GbFLS module promotes flavonoid biosynthesis in ginkgo, and they provide insight into the use of UV-B radiation to enhance the flavonoid content of ginkgo leaves.


Introduction
Ultraviolet (UV)-B radiation (280-315 nm) is not only an important component of sunlight but also an important signaling molecule [1]. UV-B radiation can directly or indirectly affect the growth, development, morphology, and metabolism of plants, and it can significantly affect the biosynthesis of various secondary metabolites [2]. High doses of UV-B radiation typically cause damage to plant leaves, which presents as curling and wilting [3]. In addition, certain doses of UV-B radiation cause dwarfism due to shortened internodes, and higher intensities of UV-B radiation are associated with more pronounced effects on growth [4]. Damage to plants by UV-B radiation is largely attributable to the accumulation of large amounts of reactive oxygen species (ROS), which damage cells and destroy intracellular proteins and DNA [5]. To counteract the damage caused by UV-B, plants increase antioxidant enzyme activity, absorb UV-B, and scavenge ROS via f lavonoid accumulation [6]. For example, under UV-B treatment the f lavonol content of Camellia sinensis leaves increased 1.5-to 2-fold [7,8] and the f lavonoid contents of rice, Tartary buckwheat, and apple increased significantly [9][10][11].
UV-B affects plant growth and development, as well as the biosynthesis of secondary metabolites, by regulating gene transcription. UV RESISTANCE LOCUS8 (UVR8), a specific receptor for UV-B, is present as a dimer in the cytoplasm in the absence of UV-B, but rapidly depolymerizes into monomers in the nucleus upon exposure to UV-B radiation [12][13][14]. UV-B-activated UVR8 inhibits the degradation of ELONGATED HYPOCOTYL 5 (HY5) by binding to CONSTITUTIVELY PHOTOMORPHOGENIC 1 (COP1) [15]. HY5, a transcription factor in the basic leucine zipper (bZIP) family, is a key player downstream of COP1 that controls photomorphogenesis and f lavonoid biosynthesis [16][17][18]. In Arabidopsis thaliana, HY5 levels are directly correlated with inhibited hypocotyl growth and increased anthocyanin content in seedlings [19]. HY5 promotes anthocyanin accumulation by activating the expression of the anthocyanin regulatory gene SlAN1, as well as anthocyanin biosynthesis genes (e.g. chalcone synthase, dihydrof lavonol 4-reductase, and anthocyanidin synthase) [20]. In addition, the transcriptional activity of MYB12 and MYB111, key regulators of f lavonol biosynthesis, is dependent on HY5 in the presence of both UV-B and visible light [21]. In apple, HY5 can activate the f lavonol synthase (FLS) promoter to increase the f lavonoid content, and MYB22 enhances this response [22].
The ancient, 'living fossil' plant Ginkgo biloba (ginkgo) is not only an ornamental tree but also an important medicinal plant due to high contents of secondary metabolites, such as f lavonoids and terpene lactones, in the leaves [23,24]. Ginkgo leaf extract (GBE) is effective in the treatment of cardiovascular diseases, dementia, and asthma [23]. Among the various compounds that constitute GBE, f lavonoids are the most abundant, and the pharmacopoeias of multiple countries stipulate that the f lavonoid content must comprise >24% of GBEs [25]. Previous studies have found that various environmental factors can affect f lavonoid accumulation in plants, and that UV-B radiation can significantly promote f lavonoid biosynthesis [6,[26][27][28]. We previously found that long-term (7-, 14-, and 21-day) UV-B treatment of ginkgo seedlings can significantly promote f lavonoid accumulation [29]. Transcriptome sequencing has revealed that GbHY5 may be a key gene in promoting f lavonoid accumulation in response to UV-B signals [29], but its function and regulatory mechanisms remain unclear.
In this study, we found that the phenotypes of 5-year-old ginkgo trees from low-altitude (LA) Pizhou and high-altitude (HA) Qinghai-Tibet Plateau (Nyingchi) differed significantly, and that ginkgo leaves from HA-grown trees had significantly higher contents of total f lavonoids and f lavonol glycosides than did leaves from LA-grown trees. Using simulated UV-B treatments, we further confirmed that UV-B radiation is the main factor promoting f lavonoid accumulation in ginkgo seedlings. Our results, along with previous transcriptomic data, point to GbHY5 as a key transcription factor that responds to UV-B radiation and promotes f lavonoid biosynthesis in ginkgo. Based on transcriptome sequencing, yeast two-hybrid (Y2H) experiments, and dual-luciferase reporter assays, we identified the molecular mechanisms by which the GbHY5-GbMYB1-GbFLS pathway promotes f lavonoid biosynthesis in ginkgo in response to UV-B radiation.

Phenotype and flavonoid content of ginkgo plants grown at different altitudes
Five-year-old trees from the LA and HA plantations exhibited distinct phenotypes (Fig. 1A-F). Compared with trees growing at LA, the HA trees were smaller and had a smaller diameter at breast height (DBH), leaf fresh weight, leaf dry weight, number of leaf lobes, and leaf area ( Fig. 1G-L). However, the total leaf f lavonoid content of trees grown at HA was 1.59 times than that of trees grown at LA, and the total content of f lavonoid glycosides was 2.4 times higher ( Fig. 1M-N). Specifically, the quercetin content of the HA trees was 2.95 mg/g (dry weight), 2.76 times that of the LA trees; the kaempferol content was 1.22 mg/g, 1.85 times that of LA trees, and the isorhamnetin content was 0.14 mg/g, 2.8 times that of LA trees (Fig. 1O).

Physiological changes and GbHY5 expression
Intense UV-B radiation is a key characteristic of HA regions: the average daily dose of UV-B in Nyingchi can be as high as 7.14 kJ/m 2 /day, far exceeding that of lowland areas [29]. Studies have found that moderate UV-B radiation promotes f lavonoid accumulation in plants [7]. To explore whether the high f lavonoid content observed in the leaves of HA trees is related to strong UV-B radiation, we subjected ginkgo seedlings to treatments simulating the average UV-B dose in Nyingchi. The malondialdehyde (MDA) and hydrogen peroxide (H 2 O 2 ) contents exhibited similar trends in response to the treatment; both increased slightly between days 1 and 5, then increased significantly as of day 7 ( Fig. 2A and B). The activity of the ROS scavengers superoxide dismutase (SOD) and catalase (CAT) increased as the treatment period progressed ( Fig. 2C and D). Flavonoids may also scavenge ROS under stress, and the total f lavonoid content of the leaves increased significantly following treatment (Fig. 2E). We previously used RNA sequencing (RNA-seq) to demonstrate that the expression of GbHY5 in ginkgo leaves increases significantly under long-term UV-B treatment (14 days), implying that this gene is important in responses to UV-B [29]. Here, we used realtime quantitative reverse transcription-PCR (qRT-PCR) to demonstrate that the expression of GbHY5 increased significantly as the duration of the treatment increased, and that trends in its expression corresponded to trends in total f lavonoid accumulation (Fig. 2F).

GbHY5 promotes flavonoid accumulation
GbHY5 belongs to the bZIP transcription factor family and has a typical bZIP domain (Fig. 3A). Phylogenetic analyses have demonstrated that GbHY5 protein is in a clade that also includes PsHY5 of Picea sitchensis (Fig. 3B). Following the transformation of 35S::GbHY5-GFP into Nicotiana benthamiana leaves, green f luorescent protein (GFP) f luorescence was observed in the nucleus, cell membrane, and cytoplasm, indicating the localization of GbHY5 to the nucleus, cell membrane, and cytoplasm (Fig. 3C).
To explore the function of GbHY5, we transformed 35S::GbHY5-GFP into ginkgo calli. qRT-PCR indicated that the expression of GbHY5 increased significantly in all three callus lines (Fig. 3D). The total f lavonoid content in the transgenic calli increased by 1.5-1.75 times compared with wild-type (WT) plants (Fig. 3E). In addition, we transformed 35S::GbHY5-GFP into A. thaliana, obtaining two transgenic lines (Fig. 3F). The transgenic plants exhibited no obvious phenotypic differences compared with WT plants ( Fig. 3G and H), but the total f lavonoid content increased significantly (Fig. 3I). These data indicate that GbHY5 promotes f lavonoid accumulation in plants.

Identification of genes downstream of GbHY5
We conducted RNA-seq of GbHY5-overexpressing transgenic ginkgo calli to explore the mechanisms by which GbHY5 regulates f lavonoid accumulation in ginkgo. In total, 39.54 Gb of clean bases were obtained following the removal of adapters, N-containing bases, and reads with low sequence quality (Supplementary Data Table S1). The clean reads were mapped to the ginkgo genome, and the total mapped rates of the six samples were all >94% (Supplementary Data Table S2). A total of 21 508 expressed genes were obtained by RNA-seq, of which 1173 were specifically expressed in the transgenic calli ( Fig. 4A). Differential expression analysis identified a total of 5948 differentially expressed genes (DEGs), of which 2400 were upregulated and 3548 were downregulated (Fig. 4B). Gene Ontology (GO) enrichment analysis of the DEGs indicated that they were enriched in multiple GO terms related to transcriptional regulation, including 'transcription regulator activity', 'DNA binding transcription factor activity', and 'sequence-specific DNA binding' (Fig. 4C). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis indicated that the DEGs were enriched in multiple secondary metabolic biosynthesis pathways, including f lavonoid biosynthesis, phenylpropanoid biosynthesis, and terpenoid backbone biosynthesis (Fig. 4D).

GbMYB1 protein interacts with GbHY5 and promotes flavonoid biosynthesis
Transcription factors, particularly MYB, play important roles in the regulation of plant f lavonoid biosynthesis. A total of 111 differentially expressed transcription factors were identified via transcriptome analysis; these were classified into 11 families, among which AP2 and MYB were the most abundant (Fig. 6A). Twenty-two members of the MYB family were differentially expressed; half were upregulated and the other half downregulated (Fig. 6B). Of these genes, GbMYB1 had the largest fold upregulation (Fig. 6B), and qRT-PCR indicated that its expression increased significantly following the overexpression of GbHY5 (Fig. 6C). Previous studies have demonstrated that ZbHY5 protein interacts with ZbMYB113 in Zanthoxylum bungeanum [30]. To test whether GbHY5 interacts with GbMYB1, we constructed pGBKT7-GbHY5 and pGADT7-GbMYB1 vectors, and, using a Y2H experiment, confirmed the hypothesized interaction (Fig. 6D). Furthermore, split-luciferase (split-LUC) complementation assays further demonstrated the interactions between GbHY5 and GbMYB1 (Fig. 6E).
We isolated GbMYB1, which has a coding sequence (CDS) of 1326 bp and encodes 441 amino acids. Phylogenetic analysis of GbMYB1 with Arabidopsis MYBs revealed that GbMYB1 with  Fig. S2). Phylogenetic tree analysis showed that the MYB1 genes from different plant species can be classified into three clades, and that GbMYB1 is in the same clade as PpMYB1 of P. patens ( Supplementary Data Fig. S2). A 35S::GbMYB1-GFP vector was constructed and transferred into the leaves of N. benthamiana for subcellular localization experiments, and the results indicated that GbMYB1 is localized in the nucleus (Fig. 7A). To explore the function of GbMYB1, the vector was also transferred into ginkgo calli, and the expression of GbMYB1 increased in the three callus lines (Fig. 7B). Subsequently, the total f lavonoid content of GbMYB1-overexpressing callus was measured, and was found to be 1.42-1.51 times that of WT plants (Fig. 7C). In addition, we transformed the 35S::GbMYB1-GFP vector into A. thaliana and obtained two transgenic lines (Fig. 7D). While the phenotype of the transgenic plants did not change significantly, the total f lavonoid content significantly increased (Fig. 7E-G). These results suggest that GbMYB1 promotes f lavonoid accumulation.

GbHY5 and GbMYB1 activate GbFLS promoter activity
GbHY5, GbMYB1, and GbFLS all promote f lavonoid accumulation in ginkgo; however, the regulatory relationship between them remains unclear. To investigate whether GbHY5 and GbMYB1 regulate GbFLS, we cloned the 1937-bp promoter sequence of GbFLS and analyzed it for cis-acting elements. PlantCARE analysis indicated that the GbFLS promoter has a potential G-box binding element for HY5 and a potential MYB binding element/MYB recognition site for MYB (Fig. 8A). To explore whether GbHY5 and GbMYB1 regulate the expression of GbFLS by binding the cis-acting element in the GbFLS promoter, we cloned the GbFLS promoter into the vector pGreen II 0800-LUC and conducted a dual-luciferase reporter assay (Fig. 8B). Compared with the con- trol, GbHY5 promoted a 1.41-fold increase in GbFLS promoter activity, whereas GbMYB1 promoted a 1.84-fold increase (Fig. 8C). Interestingly, GbFLS promoter activity increased 3.44-fold when GbHY5 and GbMYB1 were co-expressed (Fig. 8C). These results indicate that while GbHY5 and GbMYB1 can individually enhance the activity of the GbFLS promoter, their co-expression leads to greater enhancement (Fig. 8C).

Discussion
Increased UV-B radiation leads to phenotypic changes and increased accumulation of secondary metabolites in plants. Previous studies have found that UV-B radiation can reduce the leaf area in rice, sunf lower, and lettuce, and inhibit leaf growth in A. thaliana [31,32]. In cucumber, UV-B radiation not only causes a significant reduction in leaf area but also results in decreased plant height, petiole length, and biomass [4]. UV-B radiation can also lead to shortened internodes, reductions in leaf number, leaf curling, and increased axillary branches [33,34]. In addition, the f lavonoid content increases significantly under UV-B radiation to prevent damage. For example, C. sinensis, Caryopteris mongolica, Cymbopogon citratus, and Withania somnifera exhibit significant increases in f lavonoid content when exposed to UV-B radiation [7,[35][36][37]. Strong UV-B radiation is a key characteristic of HA regions, and HA plants exhibit phenotypes and f lavonoid accumulation patterns similar to those of plants subjected to UV-B treatment. Maca, which grows in the Andes, adapts to the environment by changing leaf morphogenesis and reducing the leaf area [38]. Compared with A. thaliana ecotype Columbia, the HA ecotype Tibet-0, which grows on the Qinghai-Tibet Plateau, has fewer leaves, more branches, and a smaller stature [39]. Peach, apple, and Tartary buckwheat grown at HA exhibit strong f lavonoid accumulation [10,11,40,41]. We found that the plant height, DBH, leaf fresh and dry weights, and leaf area of ginkgo grown at HA were significantly lower than those of plants grown at LA. We also found that the contents of f lavonols such as kaempferol, quercetin, and isorhamnetin, as well as the total f lavonol glycoside and f lavonoid content in the leaves of HA trees were significantly higher than those of LA trees. Our UV-B simulation experiments on ginkgo seedlings indicate that ROS and damage to the membrane system increased as treatment time increased. The scavenging activities of enzymes such as SOD and CAT increased significantly to alleviate the damage caused by excessive ROS, as did the f lavonoid content.
HY5 plays an important role in regulating plant root growth, biosynthesis and accumulation of secondary metabolites, cold tolerance, and responses to various hormonal signals [42][43][44][45]. The hypocotyls of A. thaliana hy5 mutants are significantly longer than those of WT plants when grown under both light and dark conditions [46,47], and they have more lateral roots and root hairs [48]. In addition to inhibiting hypocotyl growth, HY5 plays an important role in regulating the accumulation of secondary metabolites in response to light signals. In Artemisia annua AaHY5 indirectly controls artemisinin synthesis in response to light conditions, and in C. sinensis CsHY5 mediates UV-B light signaling to promote f lavonoid accumulation [49]. The expression of GbHY5 in ginkgo seedlings increases significantly under long-term UV-B treatment (1-4 weeks) [29]. Our analyses demonstrate that GbHY5 also responds strongly to UV-B light signals after 1-7 days of UV-B treatment, and that its expression increases over time. HY5 is localized to the nucleus in both A. thaliana and C. sinensis [49,50], whereas GbHY5 is present in both the nucleus and cell membrane in ginkgo. Similar to other studies of HY5 in angiosperms, our experiments using both ginkgo calli and ectopically transformed A. thaliana demonstrate that GbHY5 positively regulates f lavonoid accumulation. However, no clear phenotypic changes were observed in transgenic plants, likely because HY5 differs functionally in some respects between gymnosperms and angiosperms.
Previous studies have found that multiple MYB transcription factors are involved in the regulation of plant f lavonoid biosynthesis, including SG7 MYB (AtMYB11, AtMYB12, and AtMYB111), which regulate f lavonol biosynthesis in A. thaliana, and AtPAP1, which controls anthocyanin biosynthesis [51,52]. UV-B treatment of blueberry during the green fruit stage increases the expression of HY5, thereby upregulating VcMYBPA1, downregulating VcMYBC2, and promoting the accumulation of anthocyanins [53]. Conversely, an HY5-independent pathway in ripe fruit suppresses excessive anthocyanin biosynthesis when exposed to UV-B light by upregulating VcMYBC2 [53]. In pear, PyHY5, alone or in conjunction with PyBBX18, promotes the expression of PyMYB10 and PyWD40, leading to anthocyanin biosynthesis [54,55]. These results suggest that UV-B radiation, HY5, and MYB have complex and interrelated roles in the regulation of f lavonoid biosynthesis. Phylogenetic tree analysis indicated that GbMYB1, AtMYB1, and AtMYB109 belong to SG23 MYB. Previous studies reported that SG23 MYB plays an important role in plant stress response [56]. For example, AtMYB1 responds to salt stress and participates in its tolerance in Arabidopsis by regulating ABA [57]. VdMYB1 positively regulates grape defense response by activating stilbene synthase gene 2 [58]. The SlMYB75 gene promotes anthocyanin accumulation in tomato as well as volatile aroma production and can enhance resistance to Botrytis cinerea [59,60]. SsMYB113 can regulate the accumulation of Schima superba f lavonoids and improve drought stress tolerance [61]. AtMYB109 negatively regulates stomatal closure in Arabidopsis under osmotic stress [62]. However, our genetic transformation experiments confirmed that GbMYB1 is a positive regulator of f lavonoid biosynthesis in the nucleus. The functional difference between GbMYB1 and angiosperm SG23 MYB genes may be due to its relatively distant evolutionary relationship with angiosperm MYB1. Furthermore, our transcriptome sequencing, qRT-PCR, and Y2H and split-LUC assays demonstrate that GbHY5 interacts with GbMYB1 and promotes the expression of GbMYB1. Previous studies showed that UV-B irradiation promoted the accumulation of total f lavonoids but decreased the content of anthocyanins [63], and the expression of all GbDRF genes decreased significantly under UV-B irradiation, but the molecular mechanism is unclear [64]. In this study, we found that GbHY5 promoted most f lavonoid synthesis-related genes but inhibited the expression of anthocyanin-related genes such as DRF and ANS, which initially explained the decrease in anthocyanin content under UV-B irradiation. We found that GbHY5 was significantly upregulated upon UV-B irradiation and interacted with the GbMYB1 protein. Based on previous reports [65,66] and our dual luciferase experimental results, it is speculated that GbMYB1 and GbHY5 may be able to promote GbFLS expression either by enhancing the activity of the GbFLS promoter alone or to co-regulate GbFLS through protein interaction (Fig. 9). In conclusion, GbHY5-GbMYB1-GbFLS is an important module for

Plant materials, growing conditions, and UV-B treatment
Nyingchi (Tibet, China; 34 • 35 N, 117 • 59 E) is located in the Qinghai-Tibet Plateau and has an average elevation of 3000 m above sea level (asl), whereas Pizhou (Jiangsu, China; 29 • 67 N, 94 • 35 E) is located on a plain and has an average elevation of 22 m asl. We introduced individuals of G. biloba 'Fozhi' from LA Pizhou to HA Nyingchi to establish HA ginkgo plantations. The ginkgo seedlings in both places are 5 years old, and the distance between two seedlings is ∼50-60 cm. When sampling, we chose ginkgo trees far away from other tree species around the ginkgo plantation. Seedlings of G. biloba 'Fozhi' were grown in a climate chamber at a temperature of 25 • C/18 • C (day/night), with a 16-hour light/8-hour dark photoperiod. Following previously described methods [29], 4-month-old seedlings were subjected to low doses of UV-B (7.14 kJ/m 2 /day) for 1, 3, 5, or 7 days, with a 16hour light (UV-B)/8-hour dark photoperiod. A control group was grown under white light. Arabidopsis thaliana ecotype Columbia and N. benthamiana were also grown in a climate chamber at a

Gene cloning and vector construction
The CDSs of GbHY5 and GbMYB1 were amplified using an LA Taq Kit (Takara Bio Inc., Shiga, Japan) following the manufacturer's protocols and using ginkgo cDNA as a template. The specific primer sequences are shown in Supplementary Data Table S3. The PCR products were subjected to agarose gel electrophoresis and purified using an AxyPrep DNA Gel Extraction Kit (Axygen Scientific, Union City, CA, USA). The CDSs of GbHY5 and GbMYB1 were cloned into the overexpression vector pRI101-GFP using a ClonExpress II One Step Cloning Kit (Vazyme Biotech, Nanjing, China). The same methods were used to clone GbHY5 into the pGBKT7 vector, and GbMYB1 into the pGADT7 vector. The CDS of GbHY5 was cloned into the pCAMBIA1300-cLuc vector, and the CDS of GbMYB1 was cloned into the pCAMBIA1300-nLuc vector. We used a DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) to extract DNA from the leaves. The DNA was used as a template, and the GbFLS promoter was isolated using specific primers and cloned into the vector pGreen II 0800-LUC (Supplementary Data Table S3).

Promoter element analysis and phylogenetic tree construction
The cis-acting elements in the GbMYB1 promoter were predicted with PlantCARE (http://bioinformatics.psb.ugent.be/webtools/ plantcare/html/) and visualized using TBtools [67]. Protein domains were predicted using Pfam (http://pfam.xfam.org/). A multiple sequence alignment was conducted in Clustal X2.1, and phylogenetic trees were constructed in MEGA 11 using the neighbor-joining method and a bootstrap value of 1000 [68].

Subcellular localization assay
The 35S::GbHY5-GFP and 35S::GbMYB1-GFP vectors were transformed into Agrobacterium tumefaciens strain GV3101 and cultured to OD 600 = 1.0. Following centrifugation, the strain was resuspended in a solution of 10 mM MgCl 2 , 10 mM 2-(Nmorpholino) ethanesulfonic acid, and 150 mM acetosyringone. The resuspended solution was then injected into the leaves of 4-week-old N. benthamiana plants, and GFP f luorescence was observed using a confocal microscope (LSM880; Zeiss, Oberkochen, Germany) after 2 days of culture.

Genetic transformation of ginkgo calli and A. thaliana
We transformed the 35S::GbHY5-GFP and 35S:: GbMYB1-GFP vectors into A. tumefaciens strain EHA105 and infected ginkgo calli following previously described methods [69]. After four days of culture, samples were collected for use in subsequent experiments. We also transformed the two vectors into A. tumefaciens strain GV3101 and transformed A. thaliana using the f loral dip method. Seeds of transformed A. thaliana plants were harvested and sown on medium containing kanamycin, and the seedlings were transplanted into pots. PCR was used to assess the success of gene insertion. Screened, T 3 -generation positive transgenic plants were used for phenotypic observation and total f lavonoid content measurement. The total f lavonoid content was determined using a plant f lavonoid extraction kit (Suzhou Keming Biotechnology Co., Ltd., Suzhou, China).

RNA-seq and analysis
Total RNA was extracted from six ginkgo callus samples using an RNAprep Pure Plant Kit (Tiangen, Beijing, China). cDNA synthesis was conducted using an EasyScript First Strand cDNA Synthesis SuperMix Kit (TransGen Biotech Co. Ltd., Beijing, China). The cDNA library was constructed as described previously [69], and RNA-seq was conducted using an Illumina HiSeq 4000 platform (Illumina Inc., San Diego, CA, USA). Clean reads were mapped to the reference genome using HISAT2 [70]. StringTie was used for novel gene prediction, and featureCounts was used to calculate the read count of each gene [70]. Differential expression was analyzed using DESeq2 [71]. Thresholds for significant differential expression were as follows: adjusted P-value of <.5 and |log2foldchange| of >1. GO and KEGG enrichment analyses of the DEGs were conducted using clusterProfiler 4.0 [72].

qRT-PCR
Total RNA was extracted using an RNAprep Pure Plant Kit (Tiangen), following the manufacturer's protocols. GAPDH was used as an internal reference gene [73], and Primer Premier 5.0 was used to design primers for the target gene (Supplementary Data Table S3). qRT-PCR was performed on a CFX96™ Real-Time System (Bio-Rad, Hercules, CA, USA) using a PerfectStart Green qPCR SuperMix Kit (TransGen Biotech Co. Ltd.). Relative gene expression was calculated using 2 − Ct [74].

Statistical analysis
All data presented in this study were collected from more than three independent replicates and were statistically analyzed by Student's t-test or one-way ANOVA.