Major classes of color compounds in mulberry fruit
To examine the biochemical information of the lack of color phenotype of white mulberry fruit, we compared the anthocyanin content of the two mulberry varieties, Da 10 and Baisang, at different developmental stages. In Da 10, fruit color changed from green, to purple, and then to black at the S1, S2, and S3 stages, respectively (Fig. 1). For the Baisang variety, mulberries changed from green, to light green, and then to white at the S1, S2, and S3 stages, respectively (Fig. 1). The anthocyanin content in Da 10 comprised Cy and Pg. At S1, none of the measured anthocyanins were detected, whereas Cy and Pg levels were 23.533 ∝g/g 4.197 µg/g, respectively, at S2. Cy and Pg levels increased during the ripening of Da 10, reaching 375.29 and 24.423 µg/g, respectively, at stage S3. In Baisang, Cy was the only anthocyanin detected; its levels ranged from 4.87 ∝g/g at S2 to 10.957 ∝g/g at S3 (Table 1). The other anthocyanins, Dp, Pn, and Mv, were not detected in either variety at any developmental stage. Although the Baisang fruit appear white, they contain Cy at a low level. The two anthocyanins detected in this experiment, Cy and Pg, are cyanidin derivatives, which is consistent with earlier reports (Dugo et al. 2001; Qi., 2104). These results confirm, at the metabolite level, the genotype-dependent difference in the accumulation of anthocyanins in mulberry varieties, and the association between color and anthocyanin levels. In that regard, Cy and Pg seem to be the main anthocyanins determining the color of mulberry fruit. Lou et al. reported that Cy is present in the white flowers of grape hyacinth with a very short time (Lou et al., 2014). Once it is formed, the unstable Cy would be transferred to the colorless epicatechin, which would permanently prevent the formation of stable color pigments by later glycosylation and other reactions.
Transcriptome sequencing, clustering, and functional enrichment of DEGs
After quality filtering, 33.77–74.56 million high-quality pair-end reads of 150 bp were obtained. Next, rRNA reads were removed and sequencing reads were mapped to the Morus notabilis Schneid. genome (https://morus.swu.edu.cn/morusdb/), resulting in the annotation of more than 27,085 genes, including 1,735 new genes (Table S1).
DEGs were identified using the edgeR package (http://www.rproject.org/). Comparing the sequenced transcripts at the S1 and S2 stages, 5,513 and 3,973 DEGs were detected in the fruit of Da 10 and Baisang, respectively. Comparing S2 and S3, more DEGs were identified: 7,204 and 5,359, for Da 10 and Baisang, respectively. The comparison between the varieties detected 4,256, 5,612, and 5,226 DEGs at the S1, S2, and S3 developmental stages, respectively (Fig. 2a).
To analyze DEG expression patterns, the expression data at each stage were normalized to 0, log2(v1/v0), log2(v2/v0), and then clustered using Short Time-series Expression Miner software (STEM). The DEGs in the fruit of Da 10 and Baisang were clustered into eight profiles based on the gene expression patterns. Profiles 0, 1, and 3 were significantly enriched in Da 10, whereas profiles 0, 3, and 7 were significantly enriched in Baisang (Fig. 2b). The DEGs in profiles 0 and 3 were downregulated in both genotypes. To understand the relationship between the enriched genes and metabolite accumulation, the DEGs in profiles 0, 3, and 7 were annotated for both varieties. Profile 7 contained 25 genes in Da 10 and 10 genes in Baisang that were related to biosynthesis of secondary metabolites. In profile 7, more of the genes involved in carbohydrate metabolism, lipid metabolism, and energy metabolism were from Baisang than from Da 10 (Fig. 2c). Pathway enrichment analysis showed that phenylalanine, tyrosine, and tryptophan biosynthesis, phenylpropanoid biosynthesis, flavonoid biosynthesis, and anthocyanin biosynthesis were significantly enriched in profile 7 for Da 10. This indicates that anthocyanins were synthesized from the precursor, phenylalanine, followed by phenylpropanoid, flavonoid, and anthocyanin biosynthesis. In Baisang, profile 7 contained genes involved in fructose and mannose metabolism, fatty acid metabolism, glycolysis/gluconeogenesis, and fatty acid biosynthesis, which may cause the higher sugar content in Baisang than in Da 10 (Fig. 2d).
Anthocyanin biosynthesis pathway in mulberry fruit
Anthocyanin synthesis in leaves, fruits, and flowers of Arabidopsis thaliana, grape, and hyacinth is related to three secondary metabolic pathways: the phenylpropanoid, flavonoid, and anthocyanin biosynthesis pathways (Lou et al., 2014; Xie et al., 2016; Savoi et al., 2016). The lack of color development in white mulberries would require a complete blockage of the anthocyanin biosynthesis pathway, which probably occurs before Dp and Cy are formed (Lou et al., 2014). Therefore, we compared the abundance of the anthocyanin biosynthesis pathway candidate genes between the Da 10 and Baisang transcriptomes to identify the key transcripts involved in anthocyanin metabolism. Three genes involved in phenylalanine, tyrosine, and tryptophan biosynthesis, chorismate mutase (CM), arogenate dehydratase (PDT), and aspartate-prephenate aminotransferase (PAT), were upregulated in Da 10 but not upregulated in Baisang (Table S2). Other genes overexpressed in Da 10 relative to Baisang include those related to phenylpropanoid biosynthesis, flavonoid biosynthesis, and anthocyanin biosynthesis (Table S2, Fig. S1). Therefore, the pathways involved in phenylalanine, tyrosine, tryptophan, phenylpropanoid, flavonoid, and anthocyanin biosynthesis are key for the development of the Da 10 phenotype.
To obtain a global picture of the anthocyanin biosynthesis pathway in mulberry fruit, we compared the transcript levels of genes involved in anthocyanin synthesis, and the main metabolic branches associated with it, in Da 10 and Baisang. The schematic representation of anthocyanin metabolism with its core metabolites and enzymes in mulberry fruit (Fig. 3) highlights the key steps that differ between the two mulberry varieties. Chalcone synthase (CHS) expression levels were more than 40 times higher in black mulberries than in white fruit (Figs. 4 and 5). Such overexpression would lead to a large production of the intermediate metabolite, naringenin chalcone, in black mulberry. CHS catalyzes the first reaction for anthocyanin biosynthesis, and subsequently helps to form chalcone (intermediate), which is the primary precursor of flavonoids (Koes et al., 1989). Therefore, when chalcone synthesis is constrained, both anthocyanin production and that of nearly all other flavonoids, is affected (Clark et al., 2011). Another gene upregulated in Da 10 is naringenin, 2-oxoglutarate 3-dioxygenase (F3H, M026681), whose expression levels in black mulberries were about 12 times higher than in white fruit (Figs. 4 and 5). Such a high expression is expected to produce large amounts of the other intermediate metabolite, dihydrokaempferol. The upregulation of these two pathways would boost the supply of naringenin chalcone and dihydrokaempferol, thereby promoting the synthesis of anthocyanins in black mulberries (Fig. 3).
At the early developmental stage, both black and white mulberries shared the same levels of anthocyanins and patterns of gene expression. The presence of both Cy and Pg in black mulberries later in development indicates that the anthocyanin biosynthesis pathway must be induced fairly far downstream, in one of the late reactions catalyzed by enzymes such as dihydroflavonol 4-reductase (DFR), anthocyanidin synthase (ANS), or anthocyanidin 3-O-glucosyltransferase (UFGT). In contrary, the presence of myricetin in the white fruit indicates that the flavonol synthase/flavanone 3-hydroxylase (FLS) was significantly expressed in upregulation (Figs. 3 and 4). These hydroxylation reactions are essential to the formation of different kinds of anthocyanins, and eventually produce various colors. Indeed, anthocyanin accumulation are associated with the expression of the genes encoding the enzymes that participate in these reactions (Castellarin and Gaspero, 2007; Wang et al., 2010; Yuan et al., 2013).
Genes related to anthocyanin biosynthesis
Genes that regulate the anthocyanin biosynthesis pathway are generally classified as early and late genes (Deroles 2009). The early genes include CM, Arogenate dehydratase/prephenate dehydratase 6 (PDA), phenylalanine ammonia-lyase (PAL), 4-coumarate-CoA ligase (4CL), CHS, chalcone-flavonone isomerase (CHI), and F3H, which lead to the formation of dihydroflavonols. The late genes, DFR, ANS, UFGT, and cyanidin-3-O-glucoside 2-O-glucuronosyltransferase (UGAT), lead to the generation of anthocyanins (Table S2, Fig. 3). We identified 15 core genes and 5 transcription factors related to anthocyanin biosynthesis (Table S2). Eighteen of these were upregulated; only FLS and the MYB transcription factor (M011278) were downregulated in black mulberries relative to white fruit (Fig. 4). Notably, both the early gene (CM, PAT, PAL, CYP, 4CL, and CHS) and late genes (DFR, ANS, UFGT, and UGAT) were significantly downregulated in white mulberries relative to black mulberries (Fig. 4). The outcome of the downregulation of UFGT would be synthesis of lower levels of Mv and Pg, whereas the downregulation of UGAT would lead to lower levels of Cy. Chalcone synthase is a plant-specific polyketide synthase that plays a key role in flavonoid biosynthesis. In strawberries, CHS is expressed not only in the petals but also in the fruit, where its transcripts are abundant (Almeida et al., 2007). During strawberry ripening, upregulation of the CHS, F3′H, DFR, and UFGT genes corresponds to an increase in enzymatic activity in the fruit (Halbwirth et al., 2006), which results in anthocyanin accumulation at the ripe red stage. Similar results were observed for ANS expression and anthocyanin accumulation in Allium cepa (Kim et al., 2005), Duchesnea indica (Debes et al., 2011), and Morus alba (Li et al., 2014; Qi et al., 2014).
The formation of flavonols from dihydroflavonols is catalyzed by FLS. It can act on dihydrokaempferol to produce kaempferol, on dihydroquercetin to produce quercetin, and on dihydromyricetin to produce myricetin. Compared with the black mulberry, FLS expression was upregulated in the white mulberry (Fig. 4). In the white variety, the low levels of anthocyanin intermediates might have contributed to the downregulation of core the anthocyanin biosynthesis pathway genes, inhibiting the synthesis of cyanidin, whereas the upregulation of FLS gene expression would block the synthesis of anthocyanins.
The analysis of the transcriptome data resulted in the annotation of 1,217 transcription factors. Among these, 899 transcription factors were differentially expressed, including 78 MYB transcription factors, 83 bHLH (basic helix-loop-helix) transcription factors, and 31 MYB-related transcription factors. Most transcription factors involved in the regulation of genes related to the anthocyanin biosynthesis pathway belong to the MYB and bHLH transcription factor families. The functions of MYB transcription factors in the regulation of genes associated with anthocyanin accumulation have been investigated in numerous plant species, such as rice (Oryza sativa), Arabidopsis thaliana, maize (Zea mays), petunia (Petunia hybrida), grapevine (Vitis vinifera L.), snapdragon (Antirrhinum majus), apple (Malus domestica), and poplar (Populus tremuloides), using the genetic and molecular analyses (Dubos et al., 2010). The bHLH proteins form the second largest family of transcription factors in plants, where they play an important role of metabolic, physiological, and developmental processes (Pires and Dolan, 2010). Here, we found that the expression of such transcription factors differed greatly between Da 10 and Baisang. Transcription factors of the MYB (M022975 and M019093) and bHLH (M026077 and M022662) families involved in anthocyanin biosynthesis were upregulated in black mulberries relative to white mulberries, whereas MYB (M011278) showed an inconsistent expression pattern.
Expression validation of core genes and transcription factors using qPCR
In order to verify the expression levels of core genes and transcription factors related to mulberry anthocyanin synthesis, we selected the Morus010170 gene as a reference gene, because it had stable expression in all stages in both genotypes in the transcriptome data. We then validated the expression of 19 genes using qPCR (Fig. 5). According to the qPCR results, the expression levels of these genes at different developmental stages were consistent with those obtained using RNA-SEq. The expression levels of all 19 genes (and particularly of CHS, F3′H, F3H, DFR, and ANS) were much higher in Da 10 than in Baisang. Gene expression was significantly higher at stages S2 and S3 than at stage S1, in all samples. These results suggest that these 19 genes are involved in the biosynthesis of Cy and Pg through the mulberry anthocyanin biosynthesis pathway.
In conclusion, these biochemical and molecular findings provide deeper insight into the mechanisms of synthesis and accumulation of anthocyanins in different mulberry genotypes. The levels of two anthocyanins, Cy and Pg, increased sharply at late ripening in Da 10, whereas in Baisang, only Cy was detected, and only at the late developmental stage, at low levels. The high content of anthocyanin in black mulberries was associated with the upregulation of genes associated with the initial steps of anthocyanin synthesis (CM, CHS, and CHI), which produce the intermediates, chalcone and dihydroflavonol. The genes that ultimately produce Cy, Pg, and Mv in the later steps of anthocyanin synthesis (DFR, ANS, UFGT, and UGAT) were also upregulated in black mulberries. In contrast, the low anthocyanin content in white mulberries was mainly the result of the downregulation of the early genes (MYB, bHLH, CHS, and CHI), and the upregulation of FLS expression, leading to the generation of quercetin and/or myricetin.