1 Introduction

Flower color is an important property of higher plants. In addition to providing ornamental value, flower color also attracts many insect and avian pollinators. Anthocyanins are the most important floral pigments and contribute to the wide range of red to purple colors (Iwashina 2015). The anthocyanin biosynthesis pathway has been well described in many horticultural species, such as apples, roses and petunias (Peng and Moriguchi 2013; Liu et al. 2015; Li et al. 2018). This pathway involves a number of reactions that are catalyzed by various enzymes. These include phenylalanine ammonia lyase (PAL), chalcone synthase (CHS), chalcone isomerase (CHI), flavanone-3-hydroxylase (F3H), flavanone 3′-hydroxylase (F3′H), dihydroflavonol 4-reductase (DFR), anthocyanin synthase (ANS) and flavonoid-3-O-glucosyltransferase (UFGT), as well as side-branching enzymes such as flavonol synthase (FLS), anthocyanin reductase (ANR) and leucoanthocyanidin reductase (LAR) (Liu et al. 2018). It is proven by many authors that transcription factors are regulating anthocyanin biosynthesis pathway, such as R2R3-MYB, basic helix–loop–helix (bHLH) and WD40-repeat (WD40) (Zhao et al. 2013; Liu et al. 2015). These transcription factors activate or suppress the transcription and expression of target genes by binding to specific DNA sequences and affecting protein–protein interactions, thereby regulating anthocyanin biosynthesis (Yue et al. 2019). Many studies have shown that the MYB transcription factors primarily regulate flavonoid pathway genes, while bHLH and WD40 transcription factors play roles in the anthocyanin biosynthesis pathway by binding to MYB transcription factors (Zhao et al. 2013; Liu et al. 2015).

Malus hupehensis (Pamp.) Rehder, a member of the Rosaceae family, is an important native species in China (Ran et al. 2014). M. hupehensis has luxuriant and fragrant flowers. Its petals are red at early flowering stages and then gradually fade to white during flower development (Fig. 1). The fading of floral color is a common phenomenon in ornamental plants, including Malus spp. (Zhang et al. 2017), Lycoris longituba Lycoris radiata Herb. (Yue et al. 2019) and Silene littorea Silene littorea Brot. (Casimiro-Soriguer et al. 2016). The loss of floral anthocyanins results in producing white flowers (Zhao and Tao 2015), but the biochemical and molecular basis for this change remains elusive. M. hupehensis is proposed as a model flower to study mechanisms of color fading because of its various merits with regard to phenotype.

Fig. 1
figure 1

Different flowering stages of M. hupehensis: S1: small-bud stage; S2: big-bud stage; S3: initial-flowering stage; S4: full-flowering stage; S5: flower-wilting stage (similarly hereinafter). Different letters between different flower development stages denote significant differences (Duncan test, p < 0.05)

Full size image

In this study, the compositions and concentrations of compounds related to anthocyanin biosynthesis in M. hupehensis petals were determined. The expression levels of corresponding genes in the anthocyanin biosynthesis pathway were also analyzed. The mechanism of flower color fading investigated at the chemical and molecular levels provides a theoretical basis for the flower color breeding in Malus spp.

2 Material and methods

Plant materials and phenotypic characterization

– Malus hupehensis was grown in the crabapple resource nursery of Northwest A&F University, Yangling, China. Flowers were randomly collected at five developmental stages from healthy, approximately uniform eight-year-old trees in April, 2018. Sepals, pistils and stamens were quickly removed, and only the petals were collected and stored at − 80 °C. The petals for extracting pigment were vacuum freeze-dried at − 60 °C for 48 h, then ground to powder and stored in airtight containers. Flower developmental stages of M. hupehensis were described: stage 1 (S1): small-bud stage; S2: big-bud stage; S3: initial-flowering stage; S4: full-flowering stage and S5: flower-wilting stage (Fig. 1).

Color indices and pH value measurements

– The colors of the fresh petals were measured based on the Royal Horticultural Society Color Chart (RHSCC) with a CR-400 chroma meter (Konica Minolta, Tokyo, Japan). Flower color was described using two parameters, L* (lightness) and a* (ratio of red to green) (Li et al. 2018). Ten sets of values were obtained and used to calculate means.

The pH value of petal tissue was determined using the method reported by Verweij et al. (2008) with slight modifications. A 0.1 g sample of fresh petal tissue was ground in liquid nitrogen and dissolved in 6 mL of distilled water. The pH value of extract was measured using a ST2011 pH meter (Ohaus Corporation, Shanghai, China). Three biological replicates were performed, and each replicate was measured three times.

Pigment extraction

– 0.5 g of powder was added to 25 mL of methanol and incubated in an ultrasound bath at 25 °C for 30 min; this process was repeated three times. The supernatants were pooled and concentrated by rotary flash evaporation until a volume of 25 mL was achieved. The samples were stored at − 20 °C until further analysis. All experiments were performed in triplicate.

Total flavonoid content determination

– Total flavonoid content was determined using the aluminum chloride method (Wolfe et al. 2003) with slight modifications. Firstly, 1.0 mL of petal extract was mixed with 0.3 mL of 0.5 mol L−1 NaNO2 solution and the mixture was incubated in the dark at room temperature for 5 min. Next, 0.3 mL of 0.3 mol L−1 AlCl3 solution was added and the mixture was again incubated in the dark for 5 min. Then, 2.0 mL of 1.0 mol L−1 NaOH was added, and the mixture was incubated for 2 min. Finally, 6.4 mL of 30% ethanol was added, and the absorbance at 506 nm was measured using a GENSYS-10S UV spectrophotometry (Thermo, USA). Rutin was used to determine the standard curve, and the results were expressed as milligrams of rutin equivalent (RE) per gram of dry weight (mg RE kg−1 DW).

HPLC–DAD analysis

– The petal extracts were filtered through a 0.45-µm filter and analyzed by HPLC coupled with a diode array detector (DAD). HPLC–DAD analysis was carried out using a Shimadzu LC-2030C Liquid Chromatograph (Shimadzu, Kyoto, Japan) equipped with an Inertsil C-18 column (5.0 µm particle size, 4.6 mm × 250 mm). HPLC–DAD separation was performed using a linear gradient of A (0.04% formic acid dissolved in water) and B (acetonitrile) solutions at 40 °C with a flow rate of 0.5 mL min−1. The solvent gradient was as follows: 0 min, 95% A/5% B; 40 min, 60% A/40% B; 45 min, 100% A/0% B, for 15 min. The post-run time was 10 min. Anthocyanin-related compounds were identified by comparing retention times and UV spectral data with authentic standards. The concentration of individual compound was calculated based on peak area and calibration curves derived from corresponding standards. Authentic standards were chosen based on the LC-MC and NMR spectroscopic results from Malus spp. and M. hupehensis (Li et al. 2007; Wang et al. 2013; Hu et al. 2017), including cyanidin-3-galactoside, cyanidin-3-rutinoside, cyanidin-3-arabinoside, cyanidin-3-glucoside, catechin, chlorogenic acid, ferulic acid, caffeic acid, epicatechin, proanthocyanidin B1, proanthocyanidin B2, rutin, hyperoside, dihydroquercetin, phlorizin and quercetin from Shanghai Yuanye Biological Company.

Quantitative real-time PCR

– RNA extraction was performed using an E.Z.N.A Plant RNA Kit (Omega, USA) extraction kit. First-strand cDNA was synthesized from 1 ng of RNA by reverse transcription according to the manual of the PrimeScriptTM RT reagent kit with gDNA Eraser kit (TaKaRa, Japan). Quantitative real-time PCR was performed on a real-time PCR instrument (Applied Biosystems, CA, USA). SYBR® Premix Ex Taq TM (TaKaRa, Japan) was used according to the manufacturer’s instructions with three technical replicates and three biological replicates. Data analysis was performed using the 2−ΔΔCT method with 18S ribosomal RNA that was used as the reference gene (Tian et al. 2017). All primers tested are shown in Table S1.

Statistical analysis

– To further examine the major compounds that influence flower color formation, the total contents of different kinds of compounds were calculated based on the concentrations of single compounds determined by HPLC–DAD. The ratio of Sn/S5 was obtained based on the total contents of compounds at five different developmental stages, where “Sn” refers to S1, S2, S3, S4 or S5. All statistical analyses were carried out using SPSS 20.0 software. All the data are shown as mean values ± SD. A one-way analysis of variance (ANOVA) was used to detect the significant differences for each experimental test condition, and P < 0.05 was considered statistically significant.

3 Results

Color indices and pH value

– The flower color of M. hupehensis changes gradually during development. The flower bud showed no red when it is wrapped by the sepals (S1). The petal color changes quickly to red when the buds are exposed to light (S2). Later, the flower color gradually fades to white. Therefore, petals are the reddest at S2 (Fig. 1). Color differences among petals were determined based on their L* and a* values. In the uniform color space, lightness increases as L* value increases from 0 to 100. Negative values of a* indicate green, whereas positive values of a* indicate red. The results showed that the L* value of M. hupehensis petals gradually increased over time. The a* value initially increased and then decreased; the highest a* values occurred at S2 (Table 1), consistent with the observed changes in flower color.

The pH value of petal extracts was determined to assess the influence of pH on flower color. The pH values varied from 5.48 to 5.75, and the lowest value occurred at S2. However, the pH value did not differ significantly among the five stages (Table 1). These results indicated that pH value of petal has only a slight influence on flower color in M. hupehensis.

Table 1 Color indices and pH values of M. hupehensis flower at five flowering stages
Full size table

Total flavonoid content

– The total flavonoid content of petals was measured using the aluminum chloride method. Total flavonoid content increased initially and then decreased; it varied from 48,753.39 to 72,479.67 mg RE kg−1 DW at different stages (Table 2). The highest content occurred at S2, when total flavonoid content was 1.5 times higher than the lowest content (S5).

Composition of flavonoids

– The components and concentrations of flavonoids in M. hupehensis petal extracts (Table 3), three anthocyanins, including cyanidin-3-galactoside, cyanidin-3-arabinoside and cyanidin-3-rutinoside, were detected in M. hupehensis petals at whole flower development. Of these, cyanidin-3-galactoside was the major anthocyanin. Its concentration increased firstly and it was 1430.75 mg kg−1 DW, a value 10.9 times higher than the lowest content (at S5). The concentration of cyanidin-3-galactoside declined dramatically after S2. Cyanidin-3-arabinoside and cyanidin-3-rutinoside concentrations remained low and there was no significant difference during flower development.

Table 2 Total flavonoid contents of M. hupehensis flower at five flowering stages
Full size table
Table 3 Components and concentrations of polyphenols in M. hupehensis at five flowering stages/mg kg−1 DW
Full size table

Besides, other flavonoids were detected in petals, including flavonols (rutin, hyperoside and quercetin), flavanols (catechin, epicatechin, procyanidin B1 and procyanidin B2), phenolic acids (chlorogenic acid, ferulic acid and caffeic acid), dihydroflavonols (dihydromyricetin and (+)-dihydroquercetin), and dihydrochalcone (phloridzin). Almost all of the flavonols and total dihydroflavonols reached their highest concentration at S2, with the exception of (+)-dihydroquercetin. Catechin and epicatechin, which were the most abundant of flavanols, achieved their highest value at S2 and S3, respectively. However, after decreasing somewhat, the concentrations of these two compounds showed an uptrend, resulting in an increase in total flavanols. Chlorogenic acid and ferulic acid were the major phenolic acids, and their maximums were detected at S3. Phlorizin peaked at S1 and declined thereafter (Table 3).

To further determine the major compounds that influence flower color formation, the Sn/S5 ratio of various compounds was calculated at different developmental stages. The Sn/S5 ratio of total anthocyanins changed dramatically during flowering development, and the S2/S5 was 6.1 times, the highest ratio of any compounds. Compared with total anthocyanins, the Sn/S5 ratio of total flavonols showed a similar change trend but a lower intensity. The Sn/S5 ratios of total flavanols, phenolic acids, dihydroflavonols and dihydrochalcone showed little change across five stages, although contents of these substances were higher compared with total anthocyanins in petals (Fig. 2; Table 3). These results further demonstrate that the decline in total anthocyanin content is the major reason for the flower color fading in M. hupehensis.

Fig. 2
figure 2

The ratio (Sn/S5) of contents of flavonoids in petals of M. hupehensis

Full size image

Expression levels of anthocyanin biosynthesis-related genes

– Transcript levels of genes related to anthocyanin biosynthesis were analyzed by qRT-PCR at the five flowering stages of M. hupehensis. (Fig. 3). Most structural genes, including PAL, FHT, CHS, CHI, FHT, DFR, FLS, ANS and UFGT, followed a consistent expression pattern whereby their transcripts levels increased markedly, reached a maximum at S2 and sharply decreased thereafter. Expression of PAL, ANS, FLS and CHS showed striking differences during the flowering development, and their relative expression levels at S2 compared with S5 were 285.33, 100.12, 67.77 and 42.94, respectively. The transcript levels of ANR and LAR peaked at S2 and S3, respectively, and slightly increased after decreasing somewhat. These two genes can catalyze the transformation of anthocyanidins and anthocyanins to flavanols, and their transcription profiles were similar to the pattern of total flavanols content (Fig. 3).

Fig. 3
figure 3

Relative expression of structural genes in anthocyanin synthesis pathway in M. hupehensis during different flower development stages. Different letters between different flower development stages denote significant differences (Duncan test, P < 0.05)

Full size image

The expression levels of multiple transcription factors at different flowering stages were also determined. MYB12 expressions differed significantly among the five stages and reached a maximum at S3 that was 28.1-fold higher than that at S5. There was no significant difference in MYB12 expression between S2 and S3. MYB10 expression reached its highest level at S2 and decreased thereafter. The expression of MYB6 decreased initially, reached a minimum at S2 and then increased. The expression profiles of WD40 and bHLH33 were similar to that of MYB6 (Fig. 4).

Fig. 4
figure 4

Relative expression of transcription factors in anthocyanin synthesis pathway in M. hupehensis during different development stages. Different letters between different flower development stages denote significant differences (Duncan test, P < 0.05)

Full size image

Correlation of flavonoids and genes

– Correlation analyses were performed to investigate the relationship between the anthocyanin biosynthesis-related compounds and gene expression during flower development. Total anthocyanins level was significantly positively correlated with the expression of PAL, CHI, CHS, FLS, DFR, ANS, MYB10 and MYB12, but negatively correlated with the expression of MYB6, bHLH33 and WD40. Moreover, the strongest relationship (0.903**) was observed between total anthocyanins and ANS. Also, total flavonols and flavanols content were positively correlated with the expression of CHS, CHI, DFR, LAR, MYB10 and MYB12.

MYB10 expression was significantly correlated with that of CHS, CHI, DFR and ANS. Large correlations were observed between MYB12 and the structural genes CHI, CHS, F3H, F3′H, DFR and LAR. By contrast, MYB6, bHLH33 and WD40 expression was negatively correlated with the levels of many compounds and expression of most structural genes. The expression levels of these three transcription factors were significantly positively correlated with one another. (Fig. 5).

Fig. 5
figure 5

The Pearson correlation analysis between anthocyanin biosynthesis-related gene expressions and flavonoids contents in petals of M. hupehensis. *Indicates significant correlation at P < 0.05, and **indicates extremely significant correlation at P < 0.01

Full size image

4 Discussion

Malus hupehensis flower is red during early stages and fade to white during flower development. Anthocyanins are the key pigmentation molecules in Malus petals (Jiang et al. 2014; Rehman et al. 2017). In this study, abundant anthocyanin biosynthesis-related compounds were detected in the petals of M. hupehensis. Of these compounds, cyanidin-3-galactoside was the major anthocyanin. Total anthocyanin contents changed dramatically as petal color changed. The expression level of anthocyanin biosynthesis-related genes (PAL, CHS, CHI, DFR, FLS, ANS and UFGT) were initially elevated, peaked at S2 and declined afterward, coincident with a change in a* value that reflected fading trends of flower color. The results suggest that these genes play important roles in regulating anthocyanin biosynthesis. In particular, ANS expression was strongly correlated with total anthocyanins and appeared to play a principal role in the anthocyanin accumulation of M. hupehensis.

As one of the largest families of transcription factors in the regulatory network of anthocyanin biosynthesis pathways, MYBs participate in a very complex regulation process. (Liu et al. 2015). In Malus species, MdMYB10 regulates anthocyanin biosynthesis by directly binding to the DFR promoters in apple (Espley et al. 2007, 2013), and McMYB10 positively regulates McF3′H by directly binding to promoter motifs in crabapple leaves. In this study, the expression of MYB10 that was significantly correlated with that of structural genes (CHS, CHI, DFR, ANS) suggested that MYB10 regulates anthocyanin biosynthesis by means of influencing expressions of these genes in flowers. In addition to MYB10, high expression levels of MYB12 were observed in flowers, and a strong correlation was observed between expression of MYB12 and that of CHS, CHI, DFR, F3H, F3′H and LAR (a key gene of flavanol synthesis). These results indicated that MYB12 may have wide ranging influence on the expression of structural genes not only of anthocyanin synthesis, but also of flavanol in M. hupehensis flower. MYB12 isolated from leaves co-regulated anthocyanin and flavanol biosynthesis (Tian et al. 2017), and MYB12 from fruits mainly regulated flavanol by affecting the promoter activity of LAR (Wang et al. 2017). These researches demonstrate that the same transcription factors can have varied functions in different tissues and species, despite the fact that their sequences are highly conserved. In brief, both of MYB12 and MYB10 appear to promote anthocyanin accumulation in flowers by different regulatory mechanisms.

MdMYB6 is a repressor of anthocyanin accumulation in Malus, and MdMYB6 over-expression in Arabidopsis decreased anthocyanin content and down-regulated the expression of corresponding biosynthesis genes (Gao et al. 2011). In this study, the pattern of MYB6 expression was opposite to that of most structural genes as well as that of MYB10 and MYB12, consistent with the reports mentioned above. In most previous studies, bHLH33 interacted with transcriptional activators (like MYB12 and MYB10) to promote anthocyanin biosynthesis (Espley et al. 2007; Wang et al. 2017). In our study, the expression pattern of bHLH33 was opposite to that of MYB10 or MYB12, but bHLH33 expression was significantly and positively correlated with that of MYB6. It has been reported that MYB16 in apple, which is a transcriptional repressor like MYB6, repressed anthocyanin synthesis; however, its inhibition effect was weakened when it interacted with bHLH33 (Xu et al. 2017). In sweet cherry, a bHLH33 closely homologous to bHLH33 in apple strongly inhibited anthocyanin biosynthesis (Starkevič et al. 2015). These researches demonstrate that bHLH33 could interact with multiple MYBs and play variable roles in the anthocyanin biosynthesis pathway in diverse plants. Our results suggested that bHLH33 may interact with MYB6 to regulate anthocyanin biosynthesis; whether it intensifies or suppresses the effect of MYB6 requires further study.

WD40 in apple influences structural gene expressions in the anthocyanin biosynthesis pathway by interacting with bHLHs (such as bHLH33 and bHLH3) (An et al. 2012). In our study, the expression levels of bHLH33 and WD40 were significantly and positively correlated. These results indicate that WD40 may interact with bHLH33 to form a MBW complex with MYB6 to regulate anthocyanin biosynthesis. These results suggest a new researching direction to confirm the interaction among MYB6, bHLH33 and WD40 with additional experiment (such as electromobility shift assays and yeast two hybrid assays).

In conclusion, our study demonstrated that the flower color fading of M. hupehensis resulted from anthocyanin biosynthesis blocked by corresponding structural genes whose expressions were down-regulated during developmental process. The expression of several transcription factor (MYB12, MYB10 and MYB6) appeared to have a significant effect on the expression of structural genes related to anthocyanin biosynthesis. In particular, the bHLH33 and WD40 may have specific function in flower color formation. This study proposes a mechanism for flower color fading in M. hupehensis and provides a basis for further research on floral color formation.