Mango (Mangifera indica cv. Zill) fruits were obtained from South Subtropical Crops Research Institute (SSCRI) in Zhanjiang, China. Fruitlets were bagged with double layers yellow-black paper bags (Qingdao Kobayashi Co., Ltd., Qingdao, China) at 20 days after full bloom to block out all the light. Green mature fruits (130 days after full bloom) were harvested with bags, transported to the lab, and debagged for postharvest light treatment in plant growth chambers (Conviron, Adaptis A 1000, Winnipeg, Canada). 180 unblemished fruits with uniformed size were divided into two groups, with half fruits subjected to mimic sunlight treatment (mixture of 4.5 μW•cm^-2 UV-B and 16 W•m⁻² white light) and the rest retained in darkness as control. The relative humidity and temperature were 80% and 17°C, respectively. All the conditions for treatment were according to our previous patent (Qian et al., 2022). 30 fruits were regarded as one biological replicate. Fruit peel of 5 fruits per replicate was collected at 0, 6, 24, 72, 144, and 240 hours of light exposure for metabolomic and gene expression analyses. For fruit peel sampling, the exposed side was sampled by a peeler for light-treated fruit, and the up-side fruit peel was sampled for control fruit. Fruit peel was sampled as thin as possible to ensure the minimum collection of flesh.

Firmness was measured by a TA touch texture profile analyzer (Bosin Tech, Shanghai, China). After removing the peel, a 2 mm diameter probe was inserted to the equatorial part of the flesh at a 90 °C angle and depth of 5 mm. The analyzer parameters for operation were set as follows: 2 mm s^-1 for pre-test speed; 4 mm s^-1 for test speed; 3 mm s^-1 post-test speed, and 2 s for intermediate interval. Fruit color indexes in the equatorial part of fruit (L*, a*, and b* values) were measured by a portable colorimeter (LS170, Shenzhen Linshang Technology Co.,Ltd., Shenzhen, China) according to the instruction of the manufacturer. Total soluble solids content, total organic acid, and solidity-acid ratio were measured by a brix-acidity meter (PAL-BX/ACID15, ATAGO, Tokyo, Japan) according to the user manual. Analysis was performed in three biological replicates.

The details for metabolomic profiling of flavonoids were described in the previous study (Shi et al., 2021), which was conducted by Metware Biotechnology Co. Ltd. (Wuhan, China). In brief, fruit peel was successively freeze-dried, ground, and added to the extraction solution (50% methanol containing 0.1% HCl), and the supernatant was used for high-performance liquid chromatography with tandem mass spectrometry (HPLC−MS/MS) analysis. The identification of flavonoids compounds was based on the Metware Database (MWDB). The quantification of flavonoids was according to the area of the chromatographic peak, and the concentration was calculated by using the linear equation of corresponding standard. Analysis was performed in three biological replicates.

Total RNA was extracted using an RNA prep pure plant kit (Tiangen, DP441, Beijing, China). First-strand cDNA was synthesized from 1 µg of total RNA using the HiScript IIQ RT SuperMix (Vazyme, R223-01, Nanjing, China). The Q-PCR primers were designed by Primer 3 (https://bioinfo.ut.ee/primer3-0.4.0/) and synthesized by Sangon Biotech Co. Ltd, Shanghai, China ( Table S1 ). The Q-PCR reactions (15µL) were performed on a real-time PCR machine (qTOWER3G, Jena, Germany) containing 7.5µL SYBR premix ExTaqTMII (Takara, Japan), 5.5µL of cDNA (20 times diluted) and 1µL of both forward and reverse primers (10µM). Relative expression of mRNA was calculated by the cycle threshold (Ct) 2^-ΔΔCt method, with the normalization by the mango actin gene.

Apple MdMYB22 (AAZ20438.1), MdMYB12 (XP_008337875.1), MdHY5 (NP_001280752.1), and MdHYH (XP_008369576.1) sequences were used to search for the mango homologs by blasting to mango genome database using TBtools (Chen et al., 2020; Wang et al., 2020a). The amino acid sequences of R2R3 domain of flavonoids-related MYBs, and the whole amino acid sequences of HY5 and HYH from mango and other plant species were used for multiple sequence alignment by Jalview (Waterhouse et al., 2009). Full-length flavonoids-related MYBs, HY5, HYH protein sequences were used for phylogenetic tree construction by MEGA X with the neighbor-joining (NJ) method.

Data of fruit quality indexes were subjected to a one-way Analysis of Variance (ANOVA) using SPSS 27.0 (SPSS, Chicago, IL, USA), and mean values were separated by Tukey’s multiple range test. Probability values of <0.05 were considered statistically significant. Metabolic and gene expression data were subjected to a Student’s t-test using SPSS 27.0 to analyze the statistic difference between control and treatment. Probability values of <0.05 and <0.01 were considered statistically significant and highly statistically significant, marked with one asterisk (*) and two asterisks (**), respectively. Correlation analysis was conducted by ChiPlot (https://www.chiplot.online/).

During postharvest light exposure, no red coloration was detected in the control fruits which were kept in darkness, while the light-treated fruits started turning red at 6 days of exposure, and the red coloration was enhanced at 10 days ( Figure 1A ). After 10 days of treatment, the firmness of both light-treated and control fruits was decreased, and control fruits showed significant lower firmness than light-treated fruits ( Figure 1B ). Red coloration decreased the lightness (L* value), and yellowness (b* value) of light-treated fruits, but tremendously increased the fruit redness (a* value) ( Figures 1C–E ). Light treatment increased the total soluble solids content, but had no effect on the total organic acid content and solidity-acid ratio ( Figures 1F–H ).

A total of 108 standards include 9 flavonols, 6 proanthocyanidins, and 93 anthocyanins, while most substance could not be detected or exists at a very low level in our sample ( Table S2 ). According to the results of metabolites profile, naringenin-7-O-glucoside and quercetin-3-O-glucoside, procyanidin B1 and procyanidin B3, and cyanidin-3-O-galactoside and peonidin-3-O-glucoside were the main flavonols, proanthocyanidins, and anthocyanins components in ‘Zill’ mango fruit peel, respectively ( Figure 2 ). The increase of flavonoids concentration was detected by light treatment, while accumulation pattern differed among components ( Figure 2 ). The concentration of flavonols and proanthocyanidins in light-treated sample peaked at day 6, and then decreased at day 10 but still with a relatively higher concentration ( Figure 2 ). For anthocyanins, significant accumulation of Cyanidin-3-O-galactoside, and Peonidin-3-O-glucoside were detected at day 6 and day 3, respectively, continued and peaked at day 10 ( Figure 2 ). Control samples showed very low content of anthocyanins, but accumulated certain amount of flavonols and proanthocyanidins ( Figure 2 ). Concentration of most flavonoids compositions was relatively stable in control samples during the whole treatment ( Figure 2 ).

Light up-regulated the expression of all the structural genes of flavonoids biosynthesis, while the expression pattern differed among genes ( Figures 3A, B ).

For EBGs, the expression of MiPAL, MiCHI, and MiF3’H was increased at day 1 and day 3 by light treatment, and MiCHS expression was induced at day 1 and day10 ( Figure 3A ). The expression of MiF3H was increased at 6 hours of light treatment, and stayed at a higher level in the light-treated fruits during the whole experiment, with the expression peak at day 6 ( Figure 3A ).

For LBGs, the increase of MiDFR expression was detected at 6 hours, 1 day, and 3 days of exposure, and up-regulation of MiANS started from 6 hours of treatment and lasted during the whole treatment ( Figure 3B ). As a flavonols specific gene, MiFLS expression of light-treated sample was tremendously induced at day 1, peaked almost at day 6, and kept at a high expression level during the treatment ( Figure 3B ). For proanthocyanidins specific genes, MiLAR transcription in light-treated fruits was up-regulated at day 6 and day 10, with the peak at day 6, which is highly correlated to the proanthocyanidins concentration ( Figure 3B ). The expression of MiANR1 responded at the early stages of light treatment, i.e. 6 hours and 1 day, while MiANR2 expression was significantly up-regulated in light-treated fruits at day1 and day 10 ( Figure 3B ). For anthocyanins specific genes, the increased expression of both MiUFGT1 and MiUFGT3 was detected at day 3, day 6, and day10, and the expression peak of MiUFGT1 and MiUFGT3 occurred at day 3 and day 6, respectively ( Figure 3B ).

The MYB regulating anthocyanins biosynthesis in mango has already been reported by Kanzaki et al. (2020), named MiMYB1, while the flavonols and proanthocyanidins-related MYBs are still unknown. In this study, we used MdMYB12 and MdMYB22, which have been proven to regulate proanthocyanidins and flavonols biosynthesis in apple by Wang et al. (2017), to search for the mango homologs MYBs in the mango genome database (Wang et al., 2020a). Finally, Mi12g07270.1, and Mi09g04990.1 were defined as the homologs of apple MdMYB12 and MdMYB22 in mango, named MiMYB12 and MiMYB22, respectively. Multiple sequence alignment showed that MiMYB1, MiMYB22, and MiMYB12 shared a very conserved R2R3 domain with the relevant MYBs in other plant species including Arabidopsis, strawberry, grape, and apple ( Figure 4A ). Phylogenetic analysis based on the full protein sequence showed that MYBs from Subgroups (SGs) 5, 6, and 7 were clustered together, and mango MiMYB1, MiMYB12 and MiMYB22 showed the closest relationship with the homologs from petunia, apple and grape, respectively ( Figure 4B ).

The expression of MiMYB1, MiMYB12, and MiMYB22 was all induced by light, with the most significant response by MiMYB1 ( Figure 4C ). MiMYB1 transcription in light-treated samples was up-regulated from day 1, and lasted during the whole experiment ( Figure 4C ). The increased expression of MiMYB12 and MiMYB22 by light treatment was detected at day 1, day 3, and day 6, and day 1 and day 3, respectively ( Figure 4C ). MibHLH2 expression was slightly increased at day 1 and day 3, and subsequently slightly decreased at day 6, while the expression of MiWD40 showed no response to light ( Figure 4C ).

Using the apple MdHY5 (NP_001280752.1) and MdHYH (XP_008369576.1) sequences, the respective homologs in mango were identified, named MiHY5 (mango009397) and MiHYH (mango023606). Multiple sequence alignment showed that both MiHY5 and MiHYH contain a conserved bZIP domain ( Figure 5A ). The phylogenetic tree showed that HY5 and HYH were clustered into two groups, and all the HY5 or HYH proteins were clustered together ( Figure 5B ). Q-PCR analysis showed that MiCOP1 expression was only induced by light at 6 hours of treatment ( Figure 5C ). The transcription of MiHY5 was mainly up-regulated by light at the early stages of treatment, i.e. from 6 hours to 3 days, while MiHYH expression was significantly in response to light during the whole treatment ( Figure 5C ).

Correlation analysis was carried out to understand the relationship among flavonoids contents and gene expression. Results showed that 6 flavonoids components exhibited high positive correlation with each other (r > 0.5, p < 0.01) ( Figure 6 ). The most significant correlations were observed between two proanthocyanidins, i.e. procyanidin B1 and procyanidin B3 (r = 0.98, p < 0.001), and two anthocyanins, i.e. cyanidin-3-O-galactoside, and peonidin-3-O-glucoside (r = 0.98, p < 0.001), respectively ( Figure 6 ). For correlations between flavonoids contents and gene expression, MiF3H, MiLAR, MiUFGT3, and MiMYB1 were highly positively correlated with all the flavonoids components (r > 0.6, p < 0.001) ( Figure 6 ). MiFLS was highly correlated to naringenin-7-O-glucoside concentration (r = 0.72, p < 0.001), but lowly correlated to quercetin-3-O-glucoside concentration (r = 0.44, p < 0.05) ( Figure 6 ). Procyanidin B1 and procyanidin B3 showed high correlation with MiLAR, but low correlation with MiANR1 and MiANR2 ( Figure 6 ). Cyanidin-3-O-galactoside, and peonidin-3-O-glucoside showed high correlation with MiUFGT3, but low correlation with MiUFGT1 ( Figure 6 ). MiMYB12 was highly correlated with quercetin-3-O-glucoside (r = 0.67, p < 0.001), while MiMYB22 showed low correlation with flavonoids concentration ( Figure 6 ).

Light is one of the most crucial environmental factors not only regulates plant growth and development, but also induces secondary metabolites accumulation, for instance, flavonoids. Numerous preharvest and postharvest studies have shown that light promotes the total flavonoids concentration in plant (Huang et al., 2009; Qian et al., 2021), as well as different flavonoids components, including flavonols, proanthocyanidins and anthocyanins (Qian et al., 2013; Scattino et al., 2014; Shi et al., 2021). In this study, light induced the accumulation of flavonols, proanthocyanidins and anthocyanins in mango peel ( Figure 2 ), which was similar with the previous studies. In addition, the accumulation of flavonols and proanthocyanidins peaked at day 6 and subsequently decreased towards day 10 ( Figure 2 ), while the obvious red coloration and anthocyanins accumulation started at day 6, continuously increased, and peaked at day 10 ( Figure 2 ), indicating the competition among different flavonoids compositions. Flavonols, proanthocyanidins and anthocyanins are synthesized from different branches which are derived from the same pathway, and after accumulating certain amount of flavonoids, plants fine-tune the pathway to fulfill the demand of growth and development instead of over-accumulating flavonoids. For example, many fruits accumulate flavonols and proanthocyanidins as the main flavonoids components at the early developmental stage to provide astringent taste against early feeding (Jaakola, 2013), while anthocyanins become the dominant composition in ripe fruits to attract insects and animals for seed dispersal (Harborne and Williams, 2000). The competition between proanthocyanidins and anthocyanins is more obvious because they are derived from the same precursor, anthocyanidins. Therefore, the increased content of proanthocyanidins or anthocyanins usually leads to the decrease of the other substance (Xie et al., 2003; Han et al., 2012).

Among all the EBGs, MiF3H was continuously induced by light ( Figure 3A ), ensuring the sufficient precursor accumulation for flavonoids biosynthesis, i.e. dihydroflavonols. Dihydroflavonols are important intermediate products in flavonoids pathway, which can be directly catalyzed by FLS to form flavonols or further processed for proanthocyanidins and anthocyanins biosynthesis. MiFLS expression was also tremendously induced by light resulting in the increased accumulation of flavonols ( Figures 2 , 3B ). It has been showed that antisense expression of a FLS gene in petunia could inhibit flavonols accumulation and switched the petal color from light pink to red (Holton et al., 1993), while over-expression of a Camellia nitidissima FLS gene in tobacco promoted flavonols accumulation and change floral color from pink to white or light yellow (Zhou et al., 2013). The enhanced expression of MiANS in light-treated fruit peels was detected from 6 hours to 10 days ( Figure 3B ). ANS is crucial for both proanthocyanidins and anthocyanins biosynthesis since it catalyzes the formation of anthocyanidins, which can be further catalyzed by ANR to form proanthocyanidins, or by UFGT to form anthocyanins. The up-regulation of MiUFGT1 and MiUFGT3 mainly occurred form day 3 to day10 ( Figure 3B ), while MiANR1 and MiANR2 responded to light at the early stage ( Figure 3B ), indicating most anthocyanidins were converted to anthocyanins instead of proanthocyanidins at the late stage of light treatment, which was correlated to the increasing accumulation of anthocyanins from day 6 to day 10 ( Figure 2 ). MiLAR expression was highly correlated with proanthocyanidins concentration ( Figures 2 , 3B , 6 ), indicating the conversion from leucoanthocyanins to proanthocyanidins catalyzed by LAR, rather than from anthocyanidins by ANR, was predominant during proanthocyanidins biosynthesis. It has been reported that overexpression of an apple ANR gene in tobacco suppressed the tobacco LAR expression, which also indicated the competition between these two NAPDH-dependent reductases (Han et al., 2012).

By regulating the expression of flavonoids biosynthetic genes, MYBs are regarded as the most important TFs mediating flavonoids synthesis, and this process is often in response to light signal. Over-expression of a light-induced Tartary buckwheat FtMYB6 gene in Tartary buckwheat hairy roots and tobacco could significantly increase the accumulation of flavonols (Yao et al., 2020). Over-expression of Arabidopsis AtMYB111 in tobacco promoted flavonols accumulation, which requires light (Pandey et al., 2014). UV-B light responsive MYB134 promotes proanthocyanidins synthesis in poplar by binding to the promoter regions of PA pathway genes including PAL and ANR (Mellway et al., 2009). Compared with flavonols and proanthocyanidins, anthocyanins are more widely studied due to their great contribution to the plants coloration, and MYBs regulating light-induced anthocyanins biosynthesis has been reported in diverse fruit species including apple (Li et al., 2012; Bai et al., 2016), pear (Feng et al., 2010; Ni et al., 2019), blood orange (Huang et al., 2019), Chinese bayberry (Niu et al., 2010), litchi (Lai et al., 2014), and mango (Kanzaki et al., 2020). In this study, the putative MYBs regulating flavonols and proanthocyanidins biosynthesis, i.e. MiMYB22 and MiMYB12, have been identified, respectively ( Figures 4A, B ), and the expression of MiMYB1, MiMYB22 and MiMYB12 was significantly increased by light treatment ( Figure 4C ), suggesting these MYBs probably regulate light-induced flavonoids biosynthesis.

After being sensed by different photoreceptors, light signal pathway is transduced by COP1, HY5, and HYH. Under darkness, COP1 is located in the nucleus to degrade flavonoids-related regulators such as HY5 and MYBs via ubiquitination (Osterlund et al., 2000; Saijo et al., 2003; Li et al., 2012; Maier et al., 2013). Under light conditions, COP1 is translocated to cytoplasm leading to the accumulation of TFs such as HY5 and MYBs and subsequent flavonoids biosynthesis (Jiao et al., 2007; Li et al., 2012). In this study, the expression of MiCOP1 was induced by light at day 6, but showed no difference between light-treated sample and control sample at the other time points ( Figure 5C ), indicating that COP1 functions mainly through the protein subcellular localization instead of gene expression.

HY5 and HYH have also been widely reported to regulate light-induced flavonoids accumulation. In apple, MdHY5 regulates light-induced anthocyanins biosynthesis via binding to E-box and G-box motifs in the promoter region of MdMYB10 (An et al., 2017). PyHY5 is involved in the light-induced anthocyanins biosynthesis in pear by promoting the expression of PyWD40 and PyMYB10 (Wang et al., 2020b). As the homolog of HY5, HYH has also been shown to contribute to the light-induced flavonoids biosynthesis in peach (Zhao et al., 2022) and Arabidopsis (Holm et al., 2002). In this study, both MiHY5 and MiHYH responded very quickly to light and the increased expression lasted for quite a long time ( Figure 5C ), indicating the essential roles of HY5 and HYH during the light signal transduction.