Metabolic profiling and transcriptome analysis provide insights into the accumulation of flavonoids in chayote fruit during storage

Chayote (Sechium edulel) fruits are rich in flavonoids, folate, and low-calorie food. However, studies about the flavonoids and the corresponding regulatory mechanism of flavonoid synthesis in chayote fruits was still unclear. In present study, an integrated transcriptome and metabolite analysis of chayote fruits at three different storage stages were conducted to explore the flavonoid compositions and gene expression associated with flavonoid synthesis. Through the UPLC-MS/MS analysis, a total of 57 flavonoid compounds were detected. Of these, 42 flavonoid glycosides were significantly differential accumulation in chayote fruits at three different storage stages. Many genes associated with flavonoid synthesis were differentially expressed in chayote fruits at three different storage stages through RNA-seq analysis, including structural genes and some TFs. There was a high correlation between RNA-seq analysis and metabolite profiling, and the expression level of candidate genes in the flavonoid synthesis pathway were consistent with the dynamic changes of flavonoids. In addition, one R2R3-MYB transcription factor, FSG0057100, was defined as the critical regulatory gene of flavonoid synthesis. Furthermore, exogenous application of phenylalanine increased the total content of flavonoids and promoted some flavonoid biosynthesis-related gene expression in chayote fruits. The above results not only make us better understand the molecular mechanism of flavonoid synthesis in chayote fruits, but also contribute to the promotion and application of chayote products.


Introduction
Sechium edulel (Jacq.) Swartz (chayote), an herbaceous perennial climbing plant, is cultivated in most tropical and subtropical areas (1)(2)(3)(4). The chayote fruits are not only rich in nutrients, such as amino acids, vitamins, α-aminobutyric acid, and dietary fiber, but also contain a lot of flavonoids (5-9). Flavonoids are the low molecular weight polyphenolic secondary metabolic compounds produced by plants, which play important roles in human health benefits (10,11). Thus, research on flavonoid synthesis in chayote fruits has won attracted considerable attention.
Flavonoids have garnered considerable attention because of numerous nutritional and biological benefits, which were often applied in food processing to help prolong the shelf-life of food items due to their antioxidant property (18). Previous studies revealed that the total phenolic contents of rowanberry (Sorbus aucuparia) fruits and banana (Musa nana) significantly reduced during the 20-or 2-day storage at 22°C, respectively (19,20). Aziz et al. (21) investigated the contents changes of flavonoids in Cerasus Humilis fruits during storage revealed that the total content of flavonoids was reduced during storage. Robles-Sánchez et al. (22) researched the contents changes of flavonoids in mangoes during Low-Temperature storage revealed that the total content of flavonoids was reduced during the 15-day storage. Chayote has a short shelf life, which can sprout in 4-6 weeks if the storage conditions allowed (23). However, the dynamic changes of flavonoids contents and corresponding molecular mechanism during shelf life was still unclear.
Combination transcriptomic and metabolomic analyses may be a good approach to to explore the functions of genes associated with metabolism (24). Maoz's teams explored the response of table grapes under storage at anaerobic and cold stresses with an integrated transcriptomic and metabolomic analysis (25), and found that genes associated with phenylpropanoid pathway and pyruvate metabolism were differentially expressed under storage at anaerobic and cold stresses. A transcriptome and metabolome profiling were performed to explore the bluing response of radish roots under storage, and found that low reduction system: ascorbic acid (ASA), glutathione (GSH), ascorbate peroxidase (APX), glutathione peroxidase (GPX), glutathione-S-transferase (GST), or high content of glucosinolates, oxidation system: reactive oxygen species (ROS), catalase (CAT), peroxidase (POD) could promote the blue discoloration of radish roots (26). LTC condition is the prestorage at 8°C for 5 days before storage at 0°C, and Wang's teams conducted the integrated analysis of transcript and metabolite profiling of 'Hujingmilu' peach stored at 0°C conditions and LTC conditions. The results indicated that LTC can relieve chilling injury in peach fruit (27). Although the chayote fruit quality deteriorates with storage, the dynamic changes of the flavonoids in chayote fruits at room temperature for a fairly long time was still unclear. Thus, in present study, the chayote fruits placed in plastic containers at room temperature was used to perform the transcriptome and metabolome analysis.
In this study, we here performed metabolome and transcriptome profiling of chayote fruits during storage at room temperature to examine the dynamic changes of flavonoid compounds and the corresponding regulatory mechanism of flavonoid synthesis. A total of 57 flavonoid derivatives were detected in this study. Among them, 42 flavonoid glycosides were differentially accumulated in different storage stages. Moreover, the DEGs, including structural genes and transcriptional factors (TFs) for flavonoid synthesis were also identified, and the expression pattern of some DEGs were further verified using qRT-PCR. Correlation analysis between RNA-seq analysis and metabolite profiling had revealed that the identified structural genes and TFs displayed better correlation with the quantitative changes of flavonoids. This study not only makes us better understand the molecular mechanism of flavonoid synthesis in chayote fruits, but also contributes to the promotion and application of chayote products.

Plant material, growth conditions, and treatments
The Chayote was purchased from a local supplier (Walmart, Huaxi District, Guiyang City, Guizhou Province, China) at their commercial maturity stage, and then planted in the experimental field of Guizhou university. The mature chayote fruits with the same size were collected from the experimental field of Guizhou University in late October 2021. To avoid excessive dehydration, all chayote fruit samples were collected and placed in plastic containers at optimum storage temperature (10°C) for 1 month (28). Three storage stages (S1, optimal harvest stage; S2, 15 days after storage; S3, 30 days after storage) were collected to perform RNA-seq analysis and metabolome profiling, respectively. Three independent biological replicates were performed.

Phenylalanine treatment
The mature chayote fruits with the same age and size were collected on October 10, 2021 and used for phenylalanine treatment. The phenylalanine concentration was determined based on Watanabe's research (29), and the operation steps are as follows: the cottons containing phenylalanine solution with different concentrations (0.5-and 1 mM) were wrapped on the surface of fruits. The cotton was changed every 3 days. The cotton-wrapped chayote Frontiers in Nutrition 03 frontiersin.org fruits were is placed in plastic containers at optimum storage temperature (10°C) for 10 days. After 10 days of treatments, the chayote fruits were sampled and immediately placed into liquid nitroge.

Total flavonoid measurement and metabolite extraction
Determination of total flavonoid contents in chayote at three different stages (S1 ~ S3) was conducted using the Plant Flavonoids Content Assay Kit (Beijing Boxbio Science & Technology Co., Ltd.). The frozen chayote were crushed to fine powder using grinding bead (6 mm diameter, 6 min, 50 Hz, MM400, Retsch). Fifty milligrams of fine-powdered samples were extracted from the ultrasonic and cryogenic instrumentation at 5°C (40 KHz) using 400 μL of 70% aqueous methanol. The supernatants were collected by centrifugation (13,000 g, 15 min) and then passed through a hydrophilic nylon membrane filter (Pore size: 0.22 μM). Pipette 20 μL of supernatant for each sample and mix them as the quality control samples (QC).
The quadrupole-time-of-flight mass spectrometer (MS) was used to profile the metabolites of fruit samples (31). The analytical MS parameters were as follows: source temperature, 550°C; ion sources GS1, GS2, and curtain gas were set as 50, 50 and 30 psi, respectively; collision energy (CE), declustering potential (DP), ion-spray voltage floating in negative mode, and in positive mode were set as 20 ~ 60 V, 80 V, 4 kV, and 5 kV. MS acquisition was performed in a datadependent acquisition (DDA) mode.

Qualitative and quantificatiive analysis of metabolites
Data preprocessing were performed by Progenesis QI software to obtain the m/z values, retention time (RT), and peak intensity. Metabolites were found by searching and comparing the m/z values, RT, and fragmentation patterns to standards in Human metabolome database (HMDB) 1 and Metlin database. 2 The variable importance of the projection (VIP) score of OPLS-DA model was performed to detect global metabolic changes among comparable groups. Metabolites with t test p-value <0.05 and VIP ≥ 1 were defined as differential metabolites (32). The R package from Bioconductor on Majorbio Cloud Platform 3 was used to perform multivariate statistical analysis. An unsupervised method was used to perform principal component analysis (PCA) (33).

RNA-seq analysis
Extraction of the total RNA from chayote fruits at three different storage stages (S1 ~ S3) was performed using the EAZY spin Plus Plant RNA Kit. The mRNA purification, library construction, and RNA-seq was conducted by Majorbio Technology Co., Ltd (Shanghai, China). The heatmap was drawn with the log2-RPKM of individual DEGs. The RNA-seq data set is available in the Gene Expression Omnibus (accession no. GSE214641).

Quantitative real-time polymerase chain reaction
The total RNA was extracted from chayote fruits at S1, S2, and S3 stages and phenylalanine-treated seedlings using with TaKaRa MiniBEST Plant RNA Extraction Kit (TaKaRa) and the corresponding cDNA was obtained with cDNA Reverse Transcription Kit (TaKaRa). qRT-PCR was conducted with SYBR Premix Ex Tag on a DNA Engine Opticon™ 2 system (Bio-Rad Laboratories Inc.). AUG gene was used to normalize gene expression of candidate genes. A list of primers used in this study were presented in Table S1. The expression levels were evaluated by the 2 −ΔΔCt method with three biological replicates.

Statistical analysis
Significant analysis between two sample groups were calculated using T-test analysis (34). All of expression analysis were conducted for three biological replicates. The average data of three biological replicates was displayed with plus or minus the standard deviation (average ± SD).

Measurement of total flavonoid content in chayote fruits at three different storage stages
To explore the dynamic changes of flavonoids in chayote fruits during 1 month storage, the total content of flavonoids in chayote fruits at three different storage periods (S1, S2, and S3) was detected. The total content of flavonoids in S1 stage (11.31 mg. g −1 , DW) was higher than those in S2 stage (6.68 mg. g −1 , DW) and S3 stage (4.88 mg. g −1 , DW), and the total content of flavonoids in S2 stage (6.68 mg. g −1 , DW) was higher than that in S3 stages, indicating that the total content flavonoids was decreased in chayote during 1 month of storage ( Figure 1). To further examine the dynamic changes of flavonoid compounds in chayote fruits during storage, untarget metabolite profiling was conducted through the UPLC-MS/MS system. A total of 57 flavonoid compounds, including 42 flavonoid glycosides, 3 O-methylated flavonoids, 2 flavans, 2 flavones, 1 isoflavonoid, 1 isoflavan, 3 isoflavonoid O-glycosides, 1 O-methylated isoflavonoid, 1 prenylated neoflavonoid, and 1 pyranoisoflavonoid, were identified in chayote fruits at three different storage periods ( Figure 2; Table 1). Moreover, the principal component analyses (PCA) of metabolite accumulation levels in chayote fruits. Two principal components, PC1 (28.84%) and PC2 (54%), explained 82.84% of the variation in the metabolic profile ( Figure 2A). The 18 samples were divided into three groups along PC1, indicating that the model was suitable. A weak correlation among the biological repetitions of the S3 stage and a clear separation among S1, S2, and S3 were observed based on PCA ( Figure 2A). Visualization of the flavonoid profile was performed by hierarchical cluster analysis (HCA; Figure 2B). As shown in Figure 2B, the metabolite profiles were clustered into three clusters: clusters I, II, and III, and the results showed that the flavonoid compounds in chayote fruits were differentially accumulated distributed in the three periods (S1, S2, and S3) of chayote storage ( Figure 2B). In cluster I, the flavonoid compounds in S2 stage of chayote storage was higher than that in S3 stage, and the flavonoids in S1 stage of chayote storage was not detected ( Figure 2B). In cluster II, the flavonoid compounds in S1 stage was higher than that in S2 stage, and the content of flavonoids in S3 stage was relatively low ( Figure 2B). In cluster III, the flavonoid compounds in S1 stage was significantly higher than that in S3 stage, and the content of flavonoids in S2 stage was relatively low ( Figure 2B). In addition, the number of flavonoid compounds in S1 stage was higher than that in S2 stage, and the number of flavonoid compounds in S2 stage was higher than that in S3 stage ( Figure 2B).

Transcriptome analysis of chayote fruits during storage
To clarify the molecular mechanisms involved in flavonoid synthesis in chayote fruits during storage, we constructed nine cDNA libraries from the same samples used for metabolic analysis and performed RNA-seq analysis. In this study, we totally achieved 470,430,862 raw reads and 465,666,530 clean reads. The Q30 percentage and average GC content were above over 93 and 46% for all libraries, respectively. Moreover, 84.62 to 86.76% of clean reads in each library were assembled to the reference genome for each library. A total of 57,814 unigenes were identified, with 26,742, 27,925, 27,534, Determined the total content of flavonoids in chayote fruits at three storage stages (S1 S3). Error bars represented ±SD. Different lowercase letters indicate a significant difference between group means (p < 0.05). S1, optimal harvest stage; S2, 15 days after storage; S3, 30 days after storage.  (Table 2). This result indicated that the high quality of RNA-seq data can be suitable for subsequent analyses.
In order to explore the DEGs in chayote fruits during storage, we performed the correlation coefficients and gene expression clustering among the samples from different biological replicates. The k-means method was used to group 57,814 unigenes into eight different clusters, suggesting that the DEGs were differentially expressed in chayote fruit samples at different storage stages ( Figure 4). Gene expression correlations among different biological replicates were higher than 0.85 ( Figure 5B), implying that biological replicates were completely distinguishable, and identification of DEGs has been performed based on the RNA-seq data. Thus, 4,326 (1,761 up-and 2,565 down-regulated), 10,707 (4,647 up-and 6,060 down-regulated), and 10,666 (4,748 up-and 5,918 down-regulated) DEGs were found in S1 vs. S2, S2 vs. S3, and S1 vs. S3, respectively ( Figure 5A).
To investigate the function of the DEGs, Blast-GO was used to perform GO-term enrichment analysis. 8,726 DEGs (S1 vs. S2), 21,554 DEGs (S2 vs. S3), and 22,460 DEGs (S1 vs. S3), were classified into 20 function categories including 8 biological process (BP), 7 cellular component (CC), and 5 molecular function (MF; Figures 5C-E). In biological processes, the most abundant term is the cellular process, followed by the metabolic process. For the cellular component categories, the greatest abundant terms are the cell part and the membrane part. Under the molecular component categories, binding and catalytic activity were the most representative terms.
(C) Differential flavonoid metabolites in S1 vs. S3. (D) Venn diagram of differential flavonoid metabolites in different comparison groups. The red and green color in the plot represents higher and lower metabolite contents, respectively. S1, optimal harvest stage; S2, 15 days after storage; S3, 30 days after storage.

The effect of penylalanine on fruits quality during storage
Phenylalanine is an aromatic amino acid that enters the phenylpropane pathway as the initial substrate, which is the precursor of flavonoids that improve the quality of crops (41). In this study, we treated the chayote fruits with different concentrations of phenylalanine, and the effect of phenylalanine on the flavonoid contents were explored. Our results showed that the total content of flavonoids in chayote fruits treated with 0.5 mM and 1 mM of phenylalanine were 1.27 times and 1.61 times that of the control group, respectively (Supplementary Figure S1). Moreover, the expression patterns of 18 DEGs in chayote fruits treated with phenylalanine were investigated using qRT-PCR. The expression level of all genes was significantly increased after treatment with different concentrations of phenylalanine compared to control group ( Figure 10; Supplementary Table S13). The expression level of most structural genes in chayote fruits treated with 0.5 and 1 mM of phenylalanine were increased by 1.17 ~ 6.98 times and 1.25 ~ 9.92 times in comparison with control group. Moreover, the expression level of MYB TFs in chayote fruits treated with 0.5 and 1 mM of phenylalanine were increased by 1.38 ~ 5.35 times and 5.44 ~ 10.12 times in comparison with control group, respectively. In addition, the expression level of SebHLH1 and SebHLH2 in chayote fruits treated with 0.5 mM of phenylalanine was increased by 1.61 and Clustering of gene expression profiles from chayote fruits during storage. The k-means method was used to perform the cluster analysis of gene expression profiles. The number of DEGs in each cluster. The X-axis and Y-axis represent the different samples and normalized FPKM of DEGs.
Frontiers in Nutrition 10 frontiersin.org 1.54 times in comparsion with control group, respectively, which were also increased by 3.75 and 7.52 times in chayote fruits treated with 1 mM of phenylalanine. These results suggested that exogenous phenylalanine applications could promote the accumulation of flavonoids by increasing key genes involved in flavonoid biosynthesis.

Systematically detect the composition and content of flavonoid compounds In chayote fruits during storage
All organs of chayote plant are edible such as roots, shoots, leaves, and fruits, which was commonly used for medicinal and culinary purposes due to the rich minerals, dietary fibers, proteins, vitamins, polysaccharides, and flavonoids (42). Flavonoids are one of the major plant metabolites found throughout the plant kingdom, especially in fruits and vegetables. Fruits of chayote is rich in a variety of nutrients and bioactive compounds, such as polyphenols and flavonoids (43). The total phenolic contents of rowanberry fruits and banana fruits stored at room temperature for 20 day and 2 days were significantly decreased (19,20). Aung et al. (44) reported that the quality of chayote (fresh weight) was affected by postharvest storage temperature and film wrap. Wang et al. (20) found that the total content of flavonoids in fruits of Cerasus Humilis kept decreasing with storage time. However, the effect of storage time on the total content and the changes in the components of flavonoids in chayote fruits was still unclear. Thus, the dynamic change of total Frontiers in Nutrition 11 frontiersin.org flavonoid content in chayote fruits during storage was investigated, and found that the total flavonoid content was decreased in chayote fruits during 1 month storage (Figure 1), which was consistent with the previous studies (19)(20)(21). Phenylalanine is an aromatic amino acid that enters the phenylpropane pathway as the initial substrate, which is the precursor of flavonoids that improve the quality of crops (41). Previous studies have revealed that exogenous application of phenylalanine can increase the expression level of flavonoid biosynthesis-related genes. Md-Mustafa et al. (45) showed that exogenous phenylalanine increased the expression of 2 PAL, 3 4CL, and 1 CHS, and the addition of phenylalanine (1 mM) increased the folate content up to 2.0-fold. Our results showed that the total content of flavonoids in chayote fruits treated with phenylalanine solutions (0.5 and 1 mM) was 1.27-1.61 times that of the control, and some flavonoid biosynthesis-related genes were significantly increased in chayote fruits treated with phenylalanine. These results suggested that phenylalanine increased the total flavonoid content in chayote fruits during storage by upregulated the expression of flavonoid biosynthesis-related genes. Phenylalanine can be used in fruit or vegetable storage, promote flavonoid biosynthesis, improve plant resistance to biotic and abiotic stresses, and increase plant quality.

Metabolic profiling and transcriptome analysis of chayote fruits during storage
Combined transcriptome analysis and metabolic profiling were widely used to explore the flavonoid metabolites and the corresponding molecular mechanism of flavonoid synthesis in many plants, such as tea, peach, Physalis angulata, and Mulberry (46)(47)(48)(49). In this study, the flavonoid metabolites and corresponding molecular mechanism of flavonoid biosynthesis in chayote fruits during storage were explored using the combined transcriptome and metabolome profiling. A total of 57 flavonoid compounds were identified in chayote fruits during storage using UPLC-MS/ MS. The number of flavonoid compounds in chayote fruits was higher than that in fruit peel of melon (40 differential flavonoids) (16), and less than that in cucumber yellow peel (165 differential flavonoids) (50) and in citrus fruit (117 differential flavonoids) (51). Moreover, 42 (72%) of the 57 flavonoid compounds belong to flavonoid glycosides, which was the main reason for the decrease of the total content of flavonoids (21,52). Previous study showed that flavonoid glycosides are the main form of flavonoids in plants (53). In addition, flavonoid glycosides played important roles in response to abiotic stress (drought, cold, UV) (50,54). Our results showed that the content of flavonoid

Structure genes of phenylpropanoid pathway in plants
The phenylpropanoid pathway is responsible for the biosynthesis of numerous secondary metabolites such as phenylpropenes, anthocyanins, stilbenoids, and flavonoids, which regulated the plant growth and environmental response (60)(61)(62). The complex molecular mechanism of flavonoid synthesis was extensively studied in many plants (63)(64)(65). However, the molecular regulatory mechanism of A B C FIGURE 9 Connection network between DEGs and TFs associated with flavonoid synthesis and DEMs in chayote fruits during storage. (A) Correlation analysis between structural genes of flavonoid synthesis and flavonoids; (B) Correlation analysis between MYB TFs and flavonoids; (C) Correlation analysis between bHLH TFs and flavonoids. The red ellipsoid represents the DEMs. The light blue, yellow, blue, pink and green ellipsoid represent structure genes, MYB, and bHLH TFs, respectively.

FIGURE 10
Expression profiles of 18 DEGs involved in flavonoid biosynthesis after phenylalanine treatment. The expression levels of DEGs relative to actin7, measured by qRT-PCR. Three biological replicates and three technical replicates were obtained for each data point. Lower case letters (a, b, c) above the error bars represent significant difference between the treated and untreated (0 g/l) samples (p < 0.05).

Transcriptional factors involved in regulating the flavonoid synthesis
Previous studies had revealed that MYB, WRKY, bHLH, bZIP, and DOF TFs were involved in regulating the flavonoid synthesis (35)(36)(37)(38)(39). In present study, 45 MYB, 33 bHLH, 25 WRKY, 20 bZIP, and 22 Dof TFs with the differential expression patterns in chayote fruits at three differential storage stages (Supplementary Table S9). Among them, the expression patterns of 24 MYB, 14 bHLH, 10 WRKY, 11 bZIP, and 8 Dof TFs displayed a good linear relationship with structural genes in flavonoid synthesis pathway, which were increased or decreased in S1 vs. S2 and S2 vs. S3, suggesting that these TFs played important roles in regulating the flavonoid synthesis of chayote fruits. Of these 24 MYB, seven MYB TFs, including 4 positive regulatory TFs (FSG0019590, FSG0221650, FSG0057100, and FSG0036280) and 3 negative regulatory TFs (FSG0159340, FSG0079980, and FSG0179850), belong to R2R3-MYB subfamily, which have been reported to be involved in flavonoid biosynthesis. Moreover, the homologous genes of seven MYB TFs in Arabidopsis were MYB86, MYB73, MYB12, MYB78, MYB36, MYB78, MYB62, and MYB116, respectively. Previous studies revealed that seven R2R3-MYB TFs (MYB86, MYB73, MYB12, MYB78, MYB36, MYB78, MYB62, and MYB116) were involved in the positive/negative regulation of flavonoid biosynthesis (71)(72)(73)(74)(75). For example, Li et al. (76) reported that overexpression of SmMYB86 in eggplant reduced the accumulation of anthocyanins by decreasing the expression of SmCHS, SmF3H, and SmANS. Mehrtens et al. (77) revealed that MYB12, as a transcriptional regulator of CHS and FLS, acted as the flavonolspecific activator of flavonoid biosynthesis. With the development and improvement of gene knockout and transgenic technologies, overexpression and knockout of the selected candidate genes can change the flavonoid content by regulating the gene expression, which can improve the quality, enhance the tolerance to biotic and abiotic stress and extend shelf-life.
Correlation analysis between RNA-seq analysis and metabolite profiling indicated that the expression patterns of 18 DEGs (3 PAL, 1 C4H, 3 4CL, 1 CHS, 1 DFR, 1 F3H, 1 FNSI, 1 FLS, 2 IFR and 4 UGT) are closely correlated with the dynamic changes of 13 flavonoids, suggesting that these DEGs may be involved in regulating the flavonoid synthesis in chayote fruits during storage (Supplementary Table S11). Correlation analysis also revealed that 21 MYB and 12 bHLH TFs are closely correlated with eight flavonoid derivatives (r > 0.8; Figures 9B, C; Supplementary Table S12). In addition, correlation analysis between transcriptome and metabolite profiling also revealed that one R2R3-MYB TF (SeMYB1, FSG0057100) displayed a better correlation with five flavonoid glycosides, and the homologous gene of FSG0057100 in Arabidopsis was AtMYB12. Previous studies had revealed that MYB12 have the potential to activate UGT genes known to play a role in the formation of specific flavonol glycosides (39). This result indicated that FSG0057100 is the key regulators of the flavonoid synthesis in chayote fruits based on the expression pattern and correlation analysis.
In this study, the characterization and kinetic pattern of flavonoid constituents and differential flavonoid metabolites were identified in chayote fruits during storage. The DEGs involved in the flavonoid biosynthetic pathway were also detected using RNA-deq data. The expression pattern of some DEGs were further investigated using qRT-PCR. Moreover, correlation analysis between RNA-seq and metabolite revealed the structure genes and some TFs influencing flavonoid accumulation in chayote fruits during storage. Our results enable a comprehensive and systematic understanding of flavonoid compositions and dynamic changes as well as the corresponding molecular regulation mechanisms of flavonoid synthesis in chayote fruits.

Data availability statement
The original contributions presented in the study are publicly available. This data can be found at: NCBI GEO, GSE214641.

Author contributions
XW, YA, and WZ conceived and designed the experiments. YP and CW performed the experiment. XW, YP, WZ, and YJ analyzed the data. XW and YP wrote the manuscript. All authors read and approved the manuscript.