Background

Anthocyanin is a kind of water-soluble flavonoid that is derived from the branch of flavonoids, and it gives flowers and fruits various and graceful colours [1, 2]. As an antioxidant, anthocyanin can effectively remove free radicals such as reactive oxygen species (ROS) when plants suffer environmental stress, protecting plants from damage [3]. In addition, it has been shown that anthocyanin intake is beneficial to human prevention of cardiovascular diseases and cancer [4]; thus, anthocyanin has been widely studied recently.

The biosynthetic pathway of anthocyanins in higher plants is conserved, and anthocyanins are synthesized from phenylalanine catalysed by a series of enzymes. The enzyme-associated genes involved in anthocyanin synthesis are divided into early biosynthesis genes (EBGs) and late biosynthesis genes (LBGs) [5]. EBGs include chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H) and flavonoid-3′-hydroxylase (F3’H), and they are common to different flavonoid synthesis branches [6, 7]. LBGs mainly include dihydroflavonol 4-reductase (DFR), leucoanthocyanidin dioxygenase/anthocyanin synthetase (LDOX/ANS) and UDP-glucose: flavonoid 3-O-glucosyltransferase (UFGT), and they contribute to the production of various anthocyanin components by catalysing flavanonol and its subsequent derivatives [8]. Anthocyanin biosynthesis-related genes are regulated by many transcription factors, among which MYB-bHLH-WD40 (MBW) has been widely studied. The MBW complex positively regulates the expression of structural genes by binding to cis-acting elements on the promoter regions of genes (such as DFR, LDOX/ANS, UFGT, etc.) and then facilitates the accumulation of anthocyanin in plants [7, 9,10,11,12]. In addition, transcription factors such as COP1 (CONSTITUTIVE PHOTOMORPHOGENIC 1), JAZ (JASMONATE ZIM-DOMAIN), NAC (NAM, ATAF1/2, CUC2), SPL (SQUAMOSA promoter-binding protein-like) and WRKY have been considered to regulate anthocyanin biosynthesis by interacting with the MBW complex [13,14,15,16,17,18].

Anthocyanin accumulation can be affected by light, temperature, hormones and mineral nutrition, and favourable low-temperature conditions are one of the important factors that induce the biosynthesis of anthocyanins [16, 19,20,21,22,3a). To screen the candidate genes related to anthocyanin biosynthesis, our study mainly focused on the DEGs at 10 °C vs. 0 °C and 10 °C vs. 25 °C. There were 2807, 4008, 4450, and 4234 genes that were differentially expressed on Days 7, 14, 21, and 28, respectively, at 10 °C vs. 0 °C and 10 °C vs. 25 °C (Fig. 3b). Among them, 467 genes were consistently differentially expressed (Fig. 3b).

Fig. 3
figure 3

Summary of differentially expressed genes under different storage temperatures in ‘Friar’ plum. a Number of DEGs in different DEG sets. T0–7 vs. T10–7 represents the DEG set in which samples stored at 0 °C for 7 days versus samples stored at 10 °C for 7 days. b Venn diagram shows DEGs in both 10 °C vs. 0 °C and 10 °C vs. 25 °C at different storage time points. DAT represents days after treatment

Full size image

Construction of a WGCNA and coexpression network

To obtain hub genes related to anthocyanin accumulation, the relationships of DEGs, anthocyanin components and storage temperature for each sample were analysed by constructing a WGCNA (Fig. 4). Sample clustering showed that the three biological replicates of each treatment were very good (Fig. 4a). Ten coexpression modules were identified by WGCNA (Fig. 4b), among which the turquoise module was positively correlated with the contents of pelargonidin-3-O-glucoside (r = 0.67, p value = 3e-06), cyanidin-3-O-glucoside (r = 0.80, p value = 8e-11), cyanidin-3-O-rutinoside (r = 0.75, p value = 4e-08), and quercetin-3-O-rutinose (r = 0.83, p value = 6e-11). In addition, the turquoise module was positively correlated with the storage temperature of 10 °C, and the correlation coefficient was 0.90 (p value = 6e-15) (Fig. 4c). According to GO and KEGG enrichment of the candidate genes in the turquoise module (1416 genes in total), 33 genes were mapped to the flavonoid metabolism pathway, and 31 genes were mapped to the starch and sugar metabolism pathway (see Additional file 1: Figs. S1 and S2).

Fig. 4
figure 4

Weighted gene coexpression network analysis of ‘Friar’ plum fruit under different storage temperatures. a Sample clustering. A01–03 (initial sample); B01–03, C01–03, D01–03 (sample stored at 0 °C, 10 °C, and 25 °C for 7 days, respectively); B04–06, C04–06, D04–06 (sample stored at 0 °C, 10 °C, and 25 °C for 14 days), B07–09, C07–09, D07–09 (sample stored at 0 °C, 10 °C, and 25 °C for 21 days), B10–12, C10–12, D10–12 (sample stored at 0 °C, 10 °C, and 25 °C for 28 days). b Hierarchical clustering showing modules of coexpressed genes. c Module/trait correlations and corresponding p values. The right panel shows a colour scale for module/trait correlations from −1 to 1

Full size image

Identification of candidate genes involved in anthocyanin biosynthesis

A total of 43 structural genes involved in anthocyanin biosynthesis were obtained in the turquoise module, and a heatmap of their expression profiles in the flesh of ‘Friar’ plum fruit was drawn based on their FPKM value (log10(FPKM+1)) (Fig. 5). The 43 structural genes from all major steps of the anthocyanin biosynthesis pathway were distributed as follows: four phenylalanine ammonia-lyase genes (PAL), one 4-coumarate: coenzyme A ligase (4CL), fourteen chalcone synthase genes (CHS), four chalcone isomerase genes (CHI), one flavonoid-3′-hydroxylase gene (F3’H), four flavanone 3-hydroxylase genes (F3H), three dihydroflavonol 4-reductase genes (DFR), five leucoanthocyanidin dioxygenase/anthocyanin synthase genes (LDOX/ANS) and seven UDP-glucose: flavonoid 3-O-glucosyltransferase genes (UFGT).

Fig. 5
figure 5

Analysis of genes related to anthocyanin biosynthesis in the turquoise module. a Anthocyanin biosynthetic pathway. The bold font indicates the genes obtained in the turquoise module and the anthocyanin components detected in the flesh of ‘Friar’ plum fruit. b Heatmap of the expression levels of differentially expressed genes (DEGs) involved in anthocyanin biosynthesis. c qRT–PCR detection of anthocyanin synthesis-related structural genes, PsPAL (Pd.00 g835470), PsCHS (Pd.00 g300780), PsCHI (Pd.00 g402750), PsF3H (Pd.00 g891590), PsF3’H (Pd.00 g637710), PsDFR (Pd.00 g1089860), PsLDOX (Pd.00 g746630), and PsUFGT (Pd.00 g247850). d Correlation analysis of the expression profiles in qRT–PCR (qPCR) and transcriptome data (RNA-seq) (N = 13, |r| > 0.55 represents a significant correlation between the two sets of data)

Full size image

Furthermore, the expression patterns of eight representative structural genes involved in anthocyanin biosynthesis, PsPAL (Pd.00 g835470), PsCHS (Pd.00 g300780), PsCHI (Pd.00 g402750), PsF3H (Pd.00 g891590), PsF3’H (Pd.00 g637710), PsDFR (Pd.00 g1089860), PsLDOX (Pd.00 g746630) and PsUFGT (Pd.00 g247850), were studied via qRT–PCR, and the transcripts of these genes were significantly higher in the flesh of plum fruit stored at 10 °C than in that stored at 0 °C and 25 °C (Fig. 5c), which was consistent with the results of transcriptome analysis based on the correlation analysis (Fig. 5d).

Identification of genes involved in carbohydrate metabolism

Carbohydrates are considered substrates for anthocyanin synthesis, and the change in soluble sugar content was detected by HPLC. The contents of glucose, fructose and sorbitol showed downward trends at three different storage temperatures. The sucrose content decreased in the flesh of plum fruit during 0 °C storage but increased during 10 °C and 25 °C storage; moreover, it was higher at 0 °C than at 25 °C (Fig. 6a). In the turquoise module, which was related to anthocyanin synthesis, five genes were involved in starch and sugar metabolism, including two hexokinases (HXKs) and three sucrose synthases (SSs) (Fig. 6b). Correlation analysis showed that the expression patterns of the SS genes were positively correlated with sucrose content and anthocyanin content (Fig. 6c), suggesting that higher expression levels of these genes were beneficial to carbohydrate metabolism, which contributed to anthocyanin accumulation in the flesh under storage at 10 °C.

Fig. 6
figure 6

Soluble sugar content and expression pattern of carbohydrate metabolism-related genes under different storage temperatures. a Glucose, fructose, sorbitol and sucrose contents in ‘Friar’ plum fruit under different storage temperatures. b Heatmap of the expression levels of carbohydrate metabolism-related genes in the turquoise module. c Correlation analysis of anthocyanin content, soluble sugar content and expression profiles of carbohydrate metabolism-related genes (N = 13, |r| > 0.55 represents a significant correlation between the two sets of data). Different lowercase letters above the bars indicate significant differences (P < 0.05), which were obtained based on one-way ANOVA by LSD and DUNCAN tests

Full size image

Screening potential transcription factors that regulate anthocyanin synthesis

To further explore the molecular regulatory mechanism of anthocyanin biosynthesis in the flesh of ‘Friar’ plum fruit, a coexpression network was constructed based on the genes present in the turquoise module. In the network, three F3H genes (Pd.00 g799840, Pd.00 g891590 and Pd.00 g617550) and two CHS genes (Pd.00 g276460 and Pd.00 g113960) were identified as hub genes (Fig. 7a), and five transcription factor genes, MYB10, APL, WIN1, bHLH111 and bZIP43, were found to be coexpressed with anthocyanin biosynthesis-related genes. Among the five transcription factor genes, MYB10 and WIN1 were closely related to CHS, F3H, DFR and LDOX, APL was closely related to CHS, F3H and DFR, bHLH111 was only closely related to F3H, and bZIP43 was closely related to CHS and F3H. In addition, there was a positive correlation between APL and WIN1 (Fig. 7b).

Fig. 7
figure 7

Coexpression network analysis of the potential key genes in turquoise modules based on WGCNA. a Coexpression network of the genes (weight > 0.5). b Coexpression network of transcription factors and anthocyanin biosynthesis-related structural genes. c Heatmap of the expression levels of the candidate transcription factors. d qRT–PCR detection of candidate transcription factor genes PsAPL (Pd.00 g301130), PsWIN1 (Pd.00 g307540), PsMYB10 (Pd.00 g623010), PsbHLH111 (Pd.00 g1033540), and PsbZIP43 (Pd.00 g297380). e Correlation analysis of the expression profiles in qRT–PCR (qPCR) and transcriptome data (RNA-seq) (N = 13, |r| > 0.55 represents a significant correlation between the two sets of data)

Full size image

The heatmap based on the transcriptome data showed that the expression levels of APL, WIN1, MYB10, bHLH111 and bZIP43 were obviously upregulated in the fruit stored at 10 °C, while there was little change in the fruit stored at 0 °C and 25 °C (Fig. 7c). Meanwhile, the qRT–PCR results indicated that the transcripts of these five detected genes in the fruit stored at 10 °C were markedly higher than those in the fruit stored at 0 °C and 25 °C (Fig. 7d), which was consistent with the transcriptome analysis results based on the correlation analysis (Fig. 7e). Furthermore, the expression patterns of the five transcription factor genes were positively correlated with the changes in the expression levels of the structural genes and the anthocyanin content (see Additional file 1: Fig. S3). It was proposed that APL, WIN1, MYB10, bHLH111 and bZIP43 might be involved in anthocyanin accumulation by regulating the expression of structural genes associated with anthocyanin biosynthesis.

Discussion

Intermediate temperature is beneficial to promote anthocyanin accumulation in plum fruit

Temperature is an important environmental factor that affects anthocyanin accumulation. Low temperature (4 °C) can significantly induce anthocyanin accumulation in Arabidopsis seedlings in the presence of light [24]. Intermediate temperature (16 °C) effectively leads to reddening in the leaves of apple and begonia [29, 30], and 15 °C treatment promotes anthocyanin accumulation in grape peel [31]. For postharvest fruits, intermediate temperature also promotes the process of anthocyanin accumulation in peach, kiwifruit, sweet orange and plum fruits [26,27,28, 32,66] to obtain mapped data, and evaluated for library quality, such as insert length tests and randomness tests. Then, structural-level analysis, such as alternative splicing analysis, new gene discovery and gene structure optimization, was carried out. The expression levels of differentially expressed genes, functional annotation and functional enrichment of differentially expressed genes were analysed. Finally, the transcriptome data and phenotypic data were analysed by WGCNA using R language.

RNA extraction and qPCR analysis

Total RNA from flesh was extracted by the CTAB method [67]. RNA (0.8 μg, OD260:OD280 between 1.80 and 2.0, OD260:OD230 > 1.5, no obvious degradation by electrophoresis) was used for reverse transcription by PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara Biomedicals, Dalian, China). The products were diluted 15 times with nuclease-free water and then subjected to real-time fluorescence quantitative PCR (qPCR) with a TB Green® Premix Ex Taq™ II (Tli RNaseH Plus) kit (TaKaRa Biomedicals).

Quantitative real-time (qRT)-PCR assays were conducted using an Applied Biosystems 7500 Fast Real-Time PCR System. The reaction system was 20 μL, including 10 μL of 2X Green Premix Ex Taq II (Tli RNaseH Plus), 0.8 μL each of gene specific upstream primer and downstream primer, 0.4 μL of ROX Reference Dye II (50X), 2 μL of diluted cDNA, and 6 μL of nuclease-free water. The running program was set as follows: 30 s at 95 °C for one cycle, 5 s at 95 °C and 34 s at 60 °C for 40 cycles. PsACTIN7 was used as the internal reference. The primers used in this paper are listed in Additional file 1: Table S1. The relative expression levels of genes were calculated according to the 2-△△Ct method.

Statistical analysis

Each experiment was performed in three replicates. Experimental results were analysed using GraphPad Prism 8, Origin 2021, IBM SPSS Statistics 23, RStudio, and Cytoscape 3.7.1 software. Error bars denote standard deviations. Different lowercase letters above the bars indicate significant differences (P < 0.05), which were obtained based on one-way ANOVA by LSD and DUNCAN tests using IBM SPSS Statistics 23 software.