Integrated Transcriptome and Metabolome Analyses Provide Insights into the Coloring Mechanism of Dark-red and Yellow Fruits in Chinese Cherry [Cerasus pseudocerasus (Lindl.) G. Don]

Chinese cherry [Cerasus pseudocerasus (Lindl.) G. Don] is an important fruit tree from China that has excellent ornamental, economic, and nutritional values with various colors. The dark-red or red coloration of fruit, an attractive trait for consumers, is determined by anthocyanin pigmentation. In this study, the coloring patterns during fruit development in dark-red and yellow Chinese cherry fruits were firstly illustrated by integrated transcriptome and widely-targeted metabolome analyses. Anthocyanin accumulation in dark-red fruits was significantly higher compared with yellow fruits from the color conversion period, being positively correlated to the color ratio. Based on transcriptome analysis, eight structural genes (CpCHS, CpCHI, CpF3H, CpF3’H, CpDFR, CpANS, CpUFGT, and CpGST) were significantly upregulated in dark-red fruits from the color conversion period, especially CpANS, CpUFGT, and CpGST. On contrary, the expression level of CpLAR were considerably higher in yellow fruits than in dark-red fruits, especially at the early stage. Eight regulatory genes (CpMYB4, CpMYB10, CpMYB20, CpMYB306, bHLH1, CpNAC10, CpERF106, and CpbZIP4) were also identified as determinants of fruit color in Chinese cherry. Liquid chromatography-tandem mass spectrometry identified 33 and 3 differential expressed metabolites related to anthocyanins and procyanidins between mature dark-red and yellow fruits. Cyanidin-3-O-rutinoside was the predominant anthocyanin compound in both fruits, while it was 6.23-fold higher in dark-red than in yellow fruits. More accumulated flavanol and procyanidin contents resulted in less anthocyanin content in flavonoid pathway in yellow fruits due to the higher expression level of CpLAR. These findings can help understand the coloring mechanism of dark-red and yellow fruits in Chinese cherry, and provide genetic basis for breeding new cultivars.


Introduction
Anthocyanins and proanthocyanidins (PAs), known as flavonoids, belong to the group of the ubiquitous secondary metabolites. Anthocyanins are a group of important natural water-soluble pigments that commonly produce red/purple/blue colors to flowers and fruits of plants [1,2]. They have shown health-promoting properties, including antioxidant activity, cholesterol decomposition, visual acuity, and prevention of cardiovascular disease in humans [3]. In nature, anthocyanins existed as glycosides of polyhydroxy and polymethoxy derivatives mainly including cyanidin, pelargonidin, peonidin, delphinidin, malvidin, and petunidin [1,4]. The composition and proportion of anthocyanins determine the coloration of plant tissues. PAs are essential taste factors affecting astringency and bitterness of fruits, which are also considered as important determinants of fruit quality [5].

Color Phenotypic Characterization of Chinese Cherry during Fruit Development
Phenotypic observation at various developmental stages revealed significant differences in fruit peel color among the four Chinese cherry accessions. 'HP31', 'HF', and 'HP5' turned pink blush at S3 stage and the whole fruit (peel and flesh) color became dark-red, red, or light-red at S5 stage; whereas 'PZB' remained yellow throughout the corresponding developmental stage ( Figure 1A). The a*, b*, and a*/b* ratio were also considered as indicators of fruit color ( Figure 1B-D). The a*/b* ratio was significantly higher in dark-red fruits than in yellow fruits from the S3 stage, reaching the most significant differences at S5 stage compared with yellow fruits ( Figure 1D). Consistent with the color change in the fruit, anthocyanin biosynthesis started at S2 stage, significant differences in anthocyanin content among the four accessions appeared at S3 stage, and anthocyanins accumulated in large quantities by S5 stage ( Figure 1E). The total anthocyanin content was 23-fold higher in 'HP31' (75.96 mg/kg FW) than in 'PZB' (3.31 mg/kg FW) fruits at mature stage (S5), indicating that the anthocyanin accumulation was enhanced in the dark-red Chinese cherry fruits. Significant higher anthocyanin content was also detected in red fruits 'HF' and 'HP5' than that in 'PZB' ( Figure 1E). The green fruits (S1) showed the highest total flavonoids content for the all four accessions, and decreased to at red stage (S4) for 'HP31' and increased to 6.08 mg/g at dark-red stage (S5) ( Figure 1F). At mature stage, the highest and the lowest contents of flavonoids were detected in 'HP31' and 'HP5', respectively ( Figure 1F).

Differentially Expressed Genes Analysis between Dark-red and Yellow Fruits
To identify the genes related to the fruit color formation between dark-red and yellow fruits, the fruits at S1-S5 stage for four accessions were subjected to RNA-seq. The raw transcriptome sequences from 60 samples have been submitted to the CNGB database under project number CNP0003682. After filtering the raw data, 2,800,020,582 clean reads were obtained, ranging from 41,582,094 to 48,104,774 per sample. The GC content was more than 46.54%, and the Q20 values ranged from 93.86% to 97.36%. The comparison rate exceeded 85.65% with the genome of Chinese cherry (unpublished data) as the reference genome (Supplementary Table S1).
We compared the transcriptome profiles of dark-red and yellow fruits to identify differentially expressed genes (DEGs) during fruit development. There were more DEGs between dark-red and yellow fruits than that during fruit development of the same accession (Supplementary Figure S1A). For each developmental stage between dark-red and yellow fruits, the down-regulated DEGs were more abundant than the up-regulated DEGs at S1, S2, and S4 stages, but the up-regulated DEGs were more abundant than the down-regulated DEGs at S3 and S5 stages (Supplementary Figure S1A). At total of 477, 591, 1146, 1682, and 1022 common DEGs were identified between dark-red and yellow fruits at S1-S5 stages, respectively (Supplementary Figure S1B). Within each accession, the gene expressions differences were the most significant from stage S2 to S4. A total of 102, 243, 578, and 115 common DEGs were identified in the S1 vs. S2, S2 vs. S3, S3 vs. S4, S4 vs. S5 comparison groups within dark-red fruits (Supplementary Figure S1C). A total of 1047, 2833, 2027, and 1936 DEGs were identified in the S1 vs. S2, S2 vs. S3, S3 vs. S4, S4 vs. S5 comparison groups in yellow fruits (Supplementary Figure S1A).
Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis provided additional information about the enriched biological pathways, including "metabolism", "biosynthesis of other secondary metabolites", "starch and sucrose metabolism", "amino acid metabolism", and "flavonoid biosynthesis", and so on (Supplementary Figure S2). Based on the statistical significance criterion for multiple testing correlation (correlated p-value), "metabolism", "biosynthesis of other secondary metabolites", "flavonoid biosynthesis", and "transporters" were significantly enriched at S1-S5 stages between dark-red and yellow fruits (Supplementary Figure S3). Among them, structural genes such as CHS, CHI, DFR, LAR, and ANR were screened between dark-red and yellow fruits. During developmental stage S2-S4 within dark-red fruits, C4H, CHS, CHI, F3'H, DFR, and UFGT were enriched in the flavonoid biosynthesis pathway (Supplementary Figure S4). In yellow fruits, C4H, FLS, DFR, LAR, ANR, and UFGT were enriched in S2 vs. S3 comparison, and CHI, LAR, and UFGT were identified in S3 vs. S4 comparison (Supplementary Figure S5). These results suggested obvious difference in anthocyanin biosynthesis between dark-red and yellow fruits, especially during color conversion period. (D) Color ratio (a*/b*). The a * coordinate represents the red (positive)-to-green (negative) scale, the b * coordinate represents the yellow (positive)-to-blue (negative) scale. Color ratio (a*/b*) represents the comprehensive color index [32]. (E) Total anthocyanin content. (F) Total flavonoids content. Error bars indicate ± standard deviation (SD) from three independent biological replicates. The lower case letters indicate significant difference at 0.05 level.

Differentially Expressed Genes Analysis between Dark-red and Yellow Fruits
To identify the genes related to the fruit color formation between dark-red and yellow fruits, the fruits at S1-S5 stage for four accessions were subjected to RNA-seq. The raw transcriptome sequences from 60 samples have been submitted to the CNGB database under project number CNP0003682. After filtering the raw data, 2,800,020,582 clean reads were obtained, ranging from 41,582,094 to 48,104,774 per sample. The GC content was more than 46.54%, and the Q20 values ranged from 93.86% to 97.36%. The comparison rate exceeded 85.65% with the genome of Chinese cherry (unpublished data) as the reference genome (Supplementary Table S1).
We compared the transcriptome profiles of dark-red and yellow fruits to identify differentially expressed genes (DEGs) during fruit development. There were more DEGs between dark-red and yellow fruits than that during fruit development of the same accession (Supplementary Figure S1A). For each developmental stage between dark-red and (D) Color ratio (a*/b*). The a* coordinate represents the red (positive)-to-green (negative) scale, the b* coordinate represents the yellow (positive)-to-blue (negative) scale. Color ratio (a*/b*) represents the comprehensive color index [32]. (E) Total anthocyanin content. (F) Total flavonoids content. Error bars indicate ± standard deviation (SD) from three independent biological replicates. The lower case letters indicate significant difference at 0.05 level.

Weight Gene Co-Expression Network Association Analysis
The weighted co-expression network analysis (WGCNA) was conducted based on the normalized expression data for 46,058 genes from all 60 samples. After filtering, 11,515 genes were retained and classified into 12 distinct gene modules (Supplementary Figure S6). Moduletrait relationship analysis revealed that six, two and six modules were significantly (p < 0.01) correlated with color ratio, anthocyanin content and flavonoid content, respectively ( Figure 2A). Among these modules, the MEblack module of 80 genes were highly positively correlated with both color ratio (r 2 = 0.88, p = 1 × 10 −20 ) and anthocyanin content (r 2 = 0.70, p = 6 × 10 −10 ) (Figure 2A). The KEGG pathway enrichment analysis revealed that the MEblack module genes were significantly enriched in "flavonoid biosynthesis" (ko00941) and "biosynthesis of other secondary metabolites" (ko09110) ( Figure 2B). Gene ontology (GO) enrichment analysis showed that they were significantly enriched in molecular function and biological process category, mainly involved in "flavonoid metabolic process" (GO: 0009812) and "flavonoid biosynthetic process" (GO: 0009813) ( Figure 2C).
The hierarchical clustering heatmap was constructed based on FPKM (fragments per kilobase per million fragments) values in each sample, which illustrated the expression patterns of 80 genes in MEblack module among four accessions at five developmental stages ( Figure 2D). This module harbored most of structural genes involved in anthocyanin biosynthesis and transport pathway, such as CHS, CHI, F3'H, DFR, ANS, UFGT, and GST. Moreover, we annotated two reported MYB homologs involved in anthocyanin biosynthesis in other species. MSTRG.6844 and MSTRG.18742 encode TFs homologous to MYB4-like and MYB10 from Prunus avium, and they are related to anthocyanin biosynthesis [29,33]. As an important part of the MBW transcription factor complex regulating the anthocyanin biosynthesis, a bHLH1 gene (MSTRG.42352), homologous to bHLH from Arabidopsis thaliana [34] was also identified ( Figure 2D). Hub genes were screened in the MEblack module ( Figure 2E Table S2).

Expression of Genes and TFs Related to Anthocyanins Biosynthesis
To further investigate the regulatory mechanism underlying anthocyanin accumulation in Chinese cherry fruit, we focused on the 80 MEblack module genes and other structural genes involved in the flavonoid biosynthesis pathway ( Figure 3A, Supplementary Table S3). As for the expression levels, no significant differences were detected in the three genes (PAL, C4H, and 4CL) involved in phenylpropanoid pathway between dark-red and yellow fruits. The EBGs (CHS, CHI, F3H, and F3'H) and LBGs (DFR, ANS, and UFGT) were significantly up-regulated in the dark-red fruits, especially from S3 to S5 stages ( Figure 3A), consistent with the high anthocyanin content in dark-red fruits ( Figure 1). GST genes was proposed to be involved in the anthocyanin transport, which showed 3.33~4.75-fold higher log 2 fold change values at mature stage in dark-red fruits, being the most significant DEG in the anthocyanin accumulation process (Supplementary Table S3). Interestingly, the expression of LAR was up-regulated in yellow fruits especially at the early stage (S1-S2), although their levels were relatively low compared with other genes (Supplementary Table S3). This might imply that more procyanidin content was accumulated in yellow fruits.
It is widely known that anthocyanin biosynthesis was primarily regulated by the MBW protein complex and other TFs. Two TFs, MYB10 and bHLH1, selected from both DEGs and WGCNA, were significantly up-regulated in the dark-red fruits, especially at the later stages ( Figure 3B). In addition to them, we also obtained 6 MYB, 5 bHLH, and 1 WD40 in Chinese cherry fruits ( Figure 3B, Supplementary Table S3). According to transcriptional levels of candidate genes, 5 MYB and 3 bHLH were upregulated, while MYB4, bHLH148, bHLH10, and WD40 were down-regulated in dark-red fruits. Consistent with previous studies [35][36][37][38][39][40][41], other TFs including 6 NAC, 5 MADS, 8 ERF, 9 WRKY, and 3 bZIP were also selected from the DEGs ( Figure 3B). These TFs may also exert an effect on or participate in the regulation of structural and regulatory genes in anthocyanin biosynthesis.  consistent with the high anthocyanin content in dark-red fruits ( Figure 1). GST genes was proposed to be involved in the anthocyanin transport, which showed 3.33~4.75-fold higher log2fold change values at mature stage in dark-red fruits, being the most significant DEG in the anthocyanin accumulation process (Supplementary Table S3). Interestingly, the expression of LAR was up-regulated in yellow fruits especially at the early stage (S1-S2), although their levels were relatively low compared with other genes (Supplementary Table  S3). This might imply that more procyanidin content was accumulated in yellow fruits.

RT-qPCR Validation
To verify the reliability of RNA-seq data, real-time PCR was performed to determine the expression levels of seventeen genes ( Figure 4). During fruit development, their expression levels gradually upregulated from S1 to S2 stage and dramatically increased from S3 stage, then reached the highest at S5 stage in dark-red fruits. On the contrary, the expression levels slightly up-regulated from S1 to S2 stage, while significantly down-regulated at S3 stage and recovered slightly at S4 and S5 stages in yellow fruits. CpGST and CpANS showed the highest expression level at the later stages for dark-red fruits (Figure 4). At mature stage, the most significantly different genes were CpGST, CpF3H, and CpUFGT, with about 8.14-, 6.08-, and 5.50-fold log 2 fold change values in the 'PZB' vs. 'HP31' comparison, respectively. However, CpLAR revealed distinct expression profile, which was continuously decreasing during the fruit development in dark-red fruits, while it kept significant higher level in yellow fruits especially at the early stage ( Figure 4). The TFs, CpMYB10, CpMYB20, CpMYB306, CpbHLH1, CpNAC10, and CpERF106 revealed significantly higher expression levels in dark-red than in yellow fruits, especially at S4 and S5 stages ( Figure 4). By contrast, the expression level of CpMYB4 was significant lower in dark-red than in yellow fruits, suggesting its negative role in anthocyanin biosynthesis ( Figure 4). In addition, the expression levels of TFs in different tissues illustrated that these TFs revealed much higher expression levels in fruits and red flower bud than that in root, stem and leaf (Supplementary Figure S7). Finally, the correlation analyses exhibited significant correlation coefficients ranging from 0.7221 to 0.9964 (except for CpF3H, R 2 = 0.5378; CpMYB4, R 2 = 0.1187) between RNA-seq and RT-qPCR (Figure 4), supporting the accuracy of the transcriptome data.

Comparison of Metabolites between Dark-red and Yellow Fruits
To further confirm the key differential expressed metabolites (DEMs) in anthocyanin biosynthesis pathway between dark-red and yellow fruits, we obtained the metabolic pro-
In both dark-red and yellow Chinese cherry, cyanidin-3-O-rutinoside was the highest accumulated anthocyanin, accounting for 81.7369% and 85.3393% of the total anthocyanins, respectively, while it was 6.23-fold higher in dark-red than in yellow fruit ( Figure 6C Figure 6C). In the comparison, the greatest different metabolite was pelargonidin-3-O-rutinoside, being 1.05×10 7 -fold higher in dark-red than in yellow fruits ( Figure 6C, Supplementary Table S5). The heatmap based on relative expression levels of the 36 DEMs showed that 23 anthocyanins were up-regulated, while 10 anthocyanins and 3 procyanidins were down-regulated in dark-red compared with yellow fruits ( Figure 6D). Procyanidin B2 was 2.27 times in yellow fruits than that in dark-red fruits ( Figure 6C). Therefore, it was obvious that the fruit color differences between dark-red and yellow fruits were not only determined by their total anthocyanin contents, but also by their anthocyanin components and percentages, especially cyanidin, pelargonidin, and peonidin derivatives, as well as the total procyanidin contents.

Comparison of Other Flavonoid Compounds in Flavonoid Pathway
To better understand the difference in flavonoid pathway between dark-red and yellow fruits, we further compared the type and number of other flavonoid compounds. The largest three sub-classes of flavonoids were flavonols (132, 34.74%), flavones (96, 25.26%), and flavanones (34,8.95%) (Supplementary Table S5). In addition to anthocyanins, a total of 163 DEMs belonging to flavonoids were detected in the yellow vs. dark-red comparison, with 107 up-regulated and 56 down-regulated metabolites (Supplementary Figure S10). Among them, flavanones, flavanonols, and other flavonoids were all up-regulated, and the majority of DEMs from flavonols, flavones, and chalcones were up-regulated in dark-red fruits. Interestingly, 13 flavanol compounds were down-regulated among 17 differential expressed flavanols in dark-red fruits, suggesting that the accumulation of flavanol content was much higher in yellow fruits than in dark-red fruits (Supplementary Figure S10).

Comparison of Anthocyanin and Procyanidin Compounds in Dark-red and Yellow Chinese Cherry Fruits
The fruit color is largely dependent on anthocyanins classes and their concentrations. Generally, pelargonidin is reportedly as indicating an orange-red color, while cyanidin and peonidin indicate a deep red or purplish-red color [42]. Cyanidin and its glycoside derivatives have been reported as the primary anthocyanins in red-colored cherries [21,26,43]. Cyanidin 3-O-rutinoside and cyanidin 3-O-glucoside are the major anthocyanin components in sweet cherry [24,44,45] and sour cherry fruits [46,47]. Pelargonidin 3-O-glucoside, cyanidin 3-O-rutinoside, and pelargonidin 3-O-rutinoside were the three dominant anthocyanin compounds in red tomentosa cherry [22,26]. Cyanidin-3-O-glucoside was the most abundant anthocyanin in Chinese dwarf cherry, followed by pelargonidin-3-O-glucoside [48,49]. In Chinese cherry, four cyanidin-based anthocyanins were detected, and cyanidin 3-rutinoside and cyanidin 3-glucosyl-rutinoside were the two major compounds for red fruits [22]. In the present study, a total of 52 anthocyanins belonging to 6 types were firstly isolated in both dark-red and yellow fruits, although their contents of major types were much lower in yellow fruits ( Figure 6B,C). Cyanidin-3-O-rutinoside was the largest anthocyanin compound in both of them (>80%), generally similar to previous report [22], but it was 6.23-fold higher in dark-red than in yellow fruits. Meanwhile, two pelargonidins, one peonidin, and one cyanidin were also up-regulated in dark-red fruits, although their proportions were relatively low within total anthocyanin content ( Figure 6C). Two delphinidin derivatives were also the important compounds in yellow fruits, accounting for 4.88% and 3.30% ( Figure 6C). These results indicated that the color difference between dark-red and yellow Chinese cherry was not only dependent on the total anthocyanin content, but also on the anthocyanin components and proportion. In addition, our results further supported that the predominant anthocyanin compounds were similar among Chinese cherry, sweet cherry and sour cherry [22][23][24][25]44,46], while it was obvious different from tomentosa cherry [26] and Chinese dwarf cherry [48,49]. This was generally consistent with their traditional taxonomy classifications, being assigned into subgenus Cerasus and Microcerasus of genus Cerasus, respectively [15].
Procyanidin B type was predominant in cherry fruits as described by previous reports [25,26,50,51]. Procyanidin B2 and B4 accumulated from large green stage, decreasing their accumulation as the fruits ripened in sweet cherry cultivar 'Lapins' [52]. Procyanidin B2 was the major procyanidin (about 95%) in both red and white tomentosa cherry, while no significant difference was detected in total procyanidins between them [26]. The most abundant compounds in Chinese dwarf cherry genotypes were procyanidin B1 (24.54~48.79%) and B2 (4.90~20.35%) [50]. In this study, procyanidin B2 was the largest compound among 11 detected procyanidins in Chinese cherry, accounting for about half, followed by procyanidin B3 and C1 ( Figure 6C). All the relative contents of procyanidin compounds were higher in yellow than dark-red fruits, with three up-regulated compounds ( Figure 6B,C). Therefore, the greater procyanidins accumulation was also responsible for the light color of yellow fruits, strongly supported by the higher content of flavanol (the precursor of procyanidin) in yellow fruits (Supplementary Figure S10).

Key Candidate Genes Involved in Anthocyanin Biosynthesis of Chinese Cherry
It has been widely reported that a series of structural genes, including PAL, C4H, 4CL, CHS, CHI, F3H, F3'H, DFR, ANS/LDOX, and UFGT, co-regulated anthocyanin biosynthesis in many fruit crops. In the present study, a regulatory network of gene expression regulating anthocyanin biosynthesis and transport and the key differential expressed metabolites in Chinese cherry fruit was summarized, as shown in Figure 7. Combining the transcriptome data and RT-qPCR results, the up-regulation of structural genes (CpCHS, CpCHI, CpF3H, CpF3'H, CpDFR, CpANS, and CpUFGT) in (dark-red fruits enhanced flux in anthocyanin pathways. However, the low expression levels of these genes in yellow fruits generated a lack of stable anthocyanin synthesis. The DFR enzyme can selectively catalyze three kinds of substrates to synthesize three specific products: leucodelphinidin, leucopelargonidin, and leucocyanidin. The expression of F3'H promoted the synthesis of dihydroquercetin through DFR and UFGT, further forming cyanidin-3-O-rutinoside. Both the cyanidin and pelargonidin contents of dark-red fruits are considerably higher than that of yellow fruits ( Figure 6C). Thus, the high expression of CpF3'H and CpDFR determined the synthesis of specific anthocyanin component, which is consistent with the anthocyanin biosynthesis in grape [53]. Among these genes, CpANS showed the highest expression level, and CpF3H and CpUFGT exhibited the biggest difference (Figure 4). This suggested that the above seven genes are potential key genes regulating anthocyanin biosynthesis in Chinese cherry, especially CpF3H, CpANS, and CpUFGT. This was largely consistent with the results in many other Rosaceae fruit crops such as sweet cherry [27,28], tomentosa cherry [26], and peach [54]. WD40 was significantly higher in yellow than dark-red fruits ( Figure 3B), indicating its negative regulating role. Therefore, the anthocyanin biosynthesis in Chinese cherry fruit is regulated by the relevant MBW protein complex (Figure 7). We also obtained two NAC, two WRKY, one ERF, and two bZIP that were significantly upregulated in dark-red fruits. These TFs have been proven to directly or indirectly regulate anthocyanin biosynthesis by binding to an MYB promoter or through protein-protein interactions in other fruits [35,[37][38][39]. For example, MdNAC52 can bind to the promoters of MdMYB9 and MdMYB11 to increase the anthocyanin content by regulating MdLAR in apple [70]. The interaction of PyWRKY26 and PybHLH3 could co-target the PyMYB114 promoter, which resulted in anthocyanin accumulation in red-skinned pear [41]. MdbZIP44 enhances MdMYB1 binding to downstream target gene promoters to promote anthocyanin biosynthesis in apple [39]. Overall, eight regulatory genes were confirmed as determinants of fruit color in Chinese cherry. Our findings can enrich the key candidate genes and metabolites involved in the anthocyanins biosynthesis in Chinese cherry fruits, which are of great importance for molecular marker-assisted breeding.  GSTs are known to participate in the anthocyanin transport and accumulation, while the absence of GSTs often results in an anthocyanin-less phenotype with reduced pigmen-tation [7,8]. Here, four GST genes significantly increased in dark-red fruits compared to those in yellow fruits ( Figure 3). Therefore, the upregulation of CpGST played a vital role in anthocyanin transport and resulted in the accumulation in vacuoles. Interestingly, the expression level of CpGST was about 8.91-fold greater at S5 than S4 stage in yellow fruits (Figure 4), which generated yellow with blush at mature stage ( Figure 1A). This suggested that the anthocyanin biosynthesis pathway was fluent in yellow fruits, but the reduction in precursors finally resulted in the less accumulation of anthocyanins due to the lower expression levels of EBGs (CpCHS, CpCHI, CpF3H, and CpF3'H).
LAR and ANR are regarded as two key enzymes in procyanidin biosynthesis. As reported in Medicago truncatula, The loss of function of LAR in seed coats decreased the levels of procyanidins [55]. The higher expression level of CpLAR in yellow fruits (Figure 4) probably contribute to the more accumulated flavanols from the early stage, finally forming more procyanidins at mature stage ( Figure 6B,C). It has been reported that competition existed between anthocyanins and flavonols biosynthesis in red apple fruits during fruit ripening [56], and apple flowers [57]. Higher amount flavonols were detected for all stages of white flower development than red flower in peach [58]. However, this study exhibited a different branch to flavanols and procyanidins rather than flavonols in bicolored (yellow) fruits in Rosaceae family.

Transcription Factors Involved in Anthocyanin Biosynthesis of Chinese Cherry
Anthocyanin metabolism is also regulated by a series of transcription factors, such as MBW (MYB-bHLH-WD40) protein complex, NAC, WRKY, ERF, and bZIP families. MYB TFs are reportedly associated with the regulation of anthocyanin biosynthesis and accumulation in sweet cherry [29], apple [7], and blackberry [59]. bHLH commonly interacts with MYB and WD40 to regulate anthocyanin biosynthesis jointly [60]. Based on to the expression levels, we identified seven MYB and six bHLH TFs ( Figure 3B). Anthocyanin biosynthesis has been proven to be positively regulated by MYB TFs, such as MYB10 [29,61,62], MYB20 [63], and MYB306 [64], by bounding to the promoter of structural genes or interacting with bHLH genes. MYB4, a repressor activated by bHLH3, prevents the formation of the MBW through competitive binding with bHLH3 to inhibit the accumulation of anthocyanins by down-regulation of CHS, ANS, and DFR in mulberry [65], and bananas [66]. In addition, bHLH148 and bHLH10 were specifically expressed in yellow Chinese cherry ( Figure 3B), suggesting their possible negative role in regulating anthocyanin biosynthesis. A WD40 protein, homologous to Arabidopsis TTG1 [67], was identified, which has also been characterized from apple [68] and strawberry [69]. The expression level of WD40 was significantly higher in yellow than dark-red fruits ( Figure 3B), indicating its negative regulating role. Therefore, the anthocyanin biosynthesis in Chinese cherry fruit is regulated by the relevant MBW protein complex (Figure 7).
We also obtained two NAC, two WRKY, one ERF, and two bZIP that were significantly upregulated in dark-red fruits. These TFs have been proven to directly or indirectly regulate anthocyanin biosynthesis by binding to an MYB promoter or through proteinprotein interactions in other fruits [35,[37][38][39]. For example, MdNAC52 can bind to the promoters of MdMYB9 and MdMYB11 to increase the anthocyanin content by regulating MdLAR in apple [70]. The interaction of PyWRKY26 and PybHLH3 could co-target the PyMYB114 promoter, which resulted in anthocyanin accumulation in red-skinned pear [41]. MdbZIP44 enhances MdMYB1 binding to downstream target gene promoters to promote anthocyanin biosynthesis in apple [39]. Overall, eight regulatory genes were confirmed as determinants of fruit color in Chinese cherry. Our findings can enrich the key candidate genes and metabolites involved in the anthocyanins biosynthesis in Chinese cherry fruits, which are of great importance for molecular marker-assisted breeding.

Plant Materials
A total of four Chinese cherry accessions including 'HP31' (dark-red), 'HF' (red), 'HP5' (light-red) and 'PZB' (yellow) were grown under field conditions at the cherry germplasm resources of Sichuan Province (Chengdu City), China. Full flowering was set at 0 days after full bloom (DAFB) when 50% of flowers were open in the trees. Fruit samples were collected based on fruit phenology (Table 1) Table S6). Fruits were collected for their uniform size, same appearance, and no defects. Ten cherries were analyzed to measure their color parameter, and the other thirty fruits were immediately frozen in liquid nitrogen and stored at −80°C for subsequent analysis.

Fruit Color Assessment
The fruit peel color was measured using a HunterLab chromameter (Konica Minolta, Inc., Tokyo, Japan) according to the CIE system. Positive a* values indicated red and purple, and negative values indicated green and blue. Positive b* values represented yellow, and negative values represented blue. Three sets of a* and b* values were measured for the equatorial part of each fruit and used to calculate the color ratio (a*/b*) [32]. Three biological replicates per sample point were analyzed, with ten cherries for each replicate.

Total Anthocyanin and Flavonoid Content Measurement
The extraction and measurement of total anthocyanin content was conducted using a pH differential method [71]. About 1.5 g fruit was extracted with 15 mL of extraction solution (acetone:methanol:water:acetic acid = 2:2:1:0.5), after a water bath at 40 • C, the mixture was centrifuged at 8,000 g for 25 min, and the supernatant was used for determination. Two buffer systems were employed, with 0.4 M potassium chloride (pH 1.0) and 0.4 M dibasic sodium (pH 4.5). Total anthocyanin content was calculated according to the equation: A = [(A 510 − A 700 ) pH 1.0 − (A 510 − A 700 ) pH 4.5 ], which was converted into mg cyanidin 3-glucoside per 1,000 g fresh weight (FW). Three independent biological replicates per sample point were analyzed.
Aluminum chloride method was utilized for measurement of the total flavonoid content [72]. Stock solution was prepared by dissolving 50 mg quercetin in 50 mL of methanol. About 5 mg fruit extracts were mixed in 5 mL of distilled water and 0.3 mL of 5% NaNO 2 . 0.6 mL of 10% AlCl 3 and 2 mL of 1.0 M NaOH were combined in the above solution after 5 min. The absorbance of the reaction mixture was measured at 510 nm using a spectrophotometer. Total flavonoid content was calculated as quercetin equivalents (mg/g FW), and performed in triplicates.

Transcriptome Analysis
Three biological replicates with mixed three fruits each, including fruit peel and flesh, were used for RNA-seq. A total of 1 µg RNA per sample was used for RNA preparations. The mRNA molecules were purified using oligo(dT)-attached magnetic beads and then were fragmented into small pieces using fragmentation reagent. The first-strand and secondstrand cDNAs were synthesized using random hexamer-primed reverse transcription. End repair, polyadenylation, adapter ligation, PCR amplification, and library quality control were carried out following the DNABSEQ RNA-Seq library preparation protocol. A total of 60 libraries were sequences on an MGI2000 platform to generate raw 150 bp paired-end reads. For downstream analyses, high-quality clean reads were obtained by filtering low-quality reads and those containing adapters or poly-N in SOAPnuke software (−n 0.01-1.20, −q 0.4-A 0.25-cutAdaptor).
The reference genome database and gene annotation files were extracted from Chinese cherry (unpublished data). Clean reads were mapped onto the reference genome using Hisat2 v.2.1.0. The read counts and FPKM values for each gene were calculated in StringTie v.1.3.5. The DEGs were identified using EdegR package with the screening conditions |log 2 fold change| ≥ 2 and DFR ≤ 0.01. Structural and regulatory genes involved in anthocyanin biosynthesis pathway were screened from DEGs.
WGCNA was conducted using the WGCNA R package with default settings (v1.4.1717). All genes were imported into WGCNA to construct co-expression modules using the automatic network construction function block-wise Modules. Correlations between modules and color ratio, and anthocyanin and flavonoid content at each developmental stage were analyzed with respect to all genes in each module. Significant trait-related modules were identified based on high correlation values. Using default settings, genes from the MEblack module were exported for Cytoscape software (v.3.9.1) [73].

Real-Time PCR Analysis
The expression levels of nine structural genes and eight transcriptional factors in the anthocyanin biosynthesis pathway were determined by RT-qPCR. The gene-specific primers were designed by Primer 5.0 and shown in Supplementary Table S7. Total RNA was extracted from the fruits at different developmental stages using the Plant Total RNA Isolation Kit (SK8631; Sangon Biotech, Shanghai, China). The cDNA was synthesized from RNA using the PrimeScriptTM RT-PCR Kit (RR047A; TaKaRa Bio, Kusatsu, Japan). The RT-qPCR was performed in a 20 µL reaction volume using the TransStart ® Green qPCR SuperMix (TransGen Biotech Co., Ltd., Beijing, China) on a CFX96 TouchTM Real-Time PCR detection system (Bio-Rad, Hercules, CA, USA). The reaction procedure is as follows: 95 • C for 30 s, followed by 40 cycles of 95 • C for 5 s, 60 • C for 30 s, and 72 • C for 30 s. The 2 −∆∆CT method was used to calculate the gene expression levels with the geometric mean of the two housekeeping genes (cherry actin and ubiquitin). Three independent biological replicates were analyzed for each sample point.

Metabolome Analysis
The freeze-dried matured cherry samples with mixed peel and flesh were crushed using a mill (MM400, Retsch, Germany) with a zirconia bread at 30 Hz for 1.5 min. Approximately 50 mg of samples was extracted overnight at 4 • C with 1.2 mL of 70% methanol before performing centrifugation at 12,000 rpm for 3 min. The supernatants were pooled and filtered with a microporous membrane (0.22 µm). The relative quantification of widely targeted metabolites in Chinese cherry fruit using an UPLC-ESI-MS/MS system (UPLC, ExionLC TM AD; MS, Applied Biosystems 6500 Q TRAP, https://sciex.com.cn/, accessed on 21 October 2022). Quantification of metabolites was carried out using a scheduled multiple reaction monitoring method. Metabolite profiling and metabolomics data analyses were conducted by Metware Biotechnology Co., Ltd. (Wuhan, China).
PCA and OPLS-DA were conducted to verify the differences and reliability of metabolites. DEMs were determined by a VIP ≥ 1 and absolute log 2 fold change (≥1). Then the DEMs were mapped to the KEGG database and their significance was determined by hypergeometric test's p-values.

Statistical Analysis
The data was analyzed using IBM SPSS Statistics software (v25.0). The results were expressed as mean ± standard deviation (SD). A p ≤ 0.05 was considered a statistically significant difference (Tukey's test).

Conclusions
In summary, this is the first study to investigate the coloring patterns and the corresponding accumulation of anthocyanin between dark-red and yellow Chinese cherry fruits. Based on LC-MS/MS, we identified 33 and 3 differential expressed metabolites related to anthocyanins and proanthocyanidins between mature dark-red and yellow fruits. The anthocyanins were mainly up-regulated, while the proanthocyanidins were all downregulated in dark-red fruits. By transcriptome analysis, eight biosynthesis genes (CpCHS, CpCHI, CpF3H, CpF3'H, CpDFR, CpANS, CpUFGT, and CpGST) were significantly more highly expressed in dark-red fruits, especially CpANS, CpUFGT, and CpGST. CpLAR was higher in yellow fruits than dark-red fruits, especially at the early stage. Eight regulatory genes (CpMYB4, CpMYB10, CpMYB20, CpMYB306, bHLH1, CpNAC10, CpERF106, and Cp-bZIP4) were also identified as determinants of fruit color in Chinese cherry. These findings can enrich the key genes and metabolites involved in the anthocyanins biosynthesis in Chinese cherry, which are of great importance for molecular marker-assisted breeding.