Anthocyanins, an important kind of secondary metabolites that widely exist in the plant kingdom, are water-soluble plant pigments mainly stored in the vacuole of plant cells that are responsible for giving plants their pink, orange, red, purple, and blue colors (Tsuda, 2018). Anthocyanins can also attract insects for pollination and seed transmission, protect photosynthetic tissues from oxidative damage, protect cells from ultraviolet radiation damage, and confer resistance to pathogens as an antibacterial agent (Harborne and Williams, 2000; Winkel-Shirley, 2001). Another key characteristic of anthocyanins is their antioxidant activity; hence, anthocyanins are indispensable for scavenging free radicals, resisting cancers, inhibiting tumors, and preventing neuronal diseases and cardiovascular diseases (Liu et al., 2018).

Anthocyanins are synthesized via the flavonoid biosynthesis pathway, a branch of the phenylpropanoid metabolic pathway (Grotewold, 2006). Anthocyanin biosynthesis is a process catalyzed by enzymes that are encoded by a series of structural genes. Along phenylpropane branches, the following enzymes are catalyzed by multi-step biochemical reactions (Winkel-Shirley, 2001): phenylalanine ammonia-lyase (PAL), cinnamic acid by cinnamate-4-hydroxylase (C4H), 4-coumaric acid CoA ligase (4CL), chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), flavone 3’-hydroxylase (F3’H), dihydroflavonol 4-reductase (DFR), anthocyanin synthase (ANS), and UDP sugar flavonoid glycosyltransferase (UFGT). The above related enzymes involved in anthocyanin synthesis exist in cell solute, and anthocyanin is transferred to the vacuole after synthesis (Jorgensen et al., 2005; Li et al., 2012). At present, the research on the regulation of anthocyanin synthesis is more thorough (Lin-Wang et al., 2011; Cavallini et al., 2015; He et al., 2020), while the research on the anthocyanin sequestration is still in its infancy (Zhao and Dixon, 2010; Kou et al., 2019). It is generally believed that anthocyanins are synthesized by a series of synthases on the cytoplasmic surface of the endoplasmic reticulum, which makes it very important to study the mechanism of anthocyanins transferred from the endoplasmic reticulum to vacuoles. Presently, there are two hypothetical models of anthocyanin import to vacuoles: (1) vesicle-trafficking model—small vesicles filled with anthocyanins can move toward the central vacuole, as observed in grapevine and Arabidopsis (Kitamura et al., 2010; Gomez et al., 2011) and (2) ligand protein-mediated model—glutathione S-transferases (GSTs) commonly function as ligandins of anthocyanin to bind different anthocyanin monomers, and then, the anthocyanins are escorted from the cytosol to the tonoplast. After that, anthocyanin is transferred to the vicinity of vacuoles, by moving across the tonoplast through transporters, such as MRP-type ATP-binding cassette (ABC) transporter and the multidrug and toxic compound extrusion (MATE) transporter (e.g., TT12) (Debeaujon et al., 2001; Zhao and Dixon, 2010). Many studies have suggested that vesicles, GSTs, and transporters have synergistic effects on anthocyanin sequestration (Zhao, 2015).

Glutathione S-transferases (GSTs) were first found and identified to relieve the toxicity of atrazine, a photosystem II inhibitor (Frear and Swanson, 1970). Subsequent studies found that GST could regulate substance metabolism and participate in biological and abiotic stresses (Edwards et al., 2000; Hayes et al., 2005). Based on the amino acid sequences, GST family was generally divided into different subclasses in Arabidopsis, including phi (GSTF), tau (GSTU), zeta (GSTZ), theta (GSTT), lambda (GSTL), dehydroascorbate reductase (DHAR), and tetrachlorohydroquinone dehalogenase (TCHQD) (Mohsenzadeh et al., 2011). The members of phi subclass are mainly responsible for anthocyanin accumulation (Jiang et al., 2019; Kou et al., 2019; Liu et al., 2019). GST is not an enzyme that directly catalyzes anthocyanin synthesis; rather, it is a carrier responsible for anthocyanins sequestrated in vacuoles (Mueller et al., 2000; Sun et al., 2012). Generally, there is a strong correlation between the subcellular location of GST and flavonoids. In addition, GSTs are relevant for membranes, such as the endoplasmic reticulum (ER) and vacuole, in anthocyanin-rich plant cells (Smith et al., 2003; Sun et al., 2012).

In maize, Bronze-2(Bz2) was the first GST gene to be identified to perform an important role in anthocyanin accumulation (Marrs et al., 1995). Subsequently, Bz2 homologous genes were uncovered in many plants, including Arabidopsis (TT19), petunia (AN9), grape (VvGST1 and VvGST4), Camelina sativa (CmGSTF12), sweet potato (IbGSTF4), kiwifruit (AcGST1), apple (MdGSTF6), carrot (DcGST1), and cotton (GhGSTF12), which implied that the functioning of GST protein in anthocyanin accumulation is highly conserved (Alfenito et al., 1998; Larsen et al., 2003; Kitamura et al., 2010; Wang et al., 2012; Kou et al., 2019; Liu et al., 2019; Meng et al., 2020; Shao et al., 2021). In the Arabidopsis transparent testa19 (TT19) mutant, both anthocyanin and proanthocyanidin accumulation in vacuoles of vegetative tissue is affected (Kitamura et al., 2010). Similar to Bz2 and AN9, TT19 encodes a GST protein, but AN9 can restore the defective phenotype of the Arabidopsis tt19 mutant, which cannot accumulate anthocyanin normally (Kitamura et al., 2004). Both TT19 and AN9/Bz2 may function by stabilizing compounds and sequestrating anthocyanins in the vacuole. The different seed coat color of the Arabidopsis tt19 mutant may be due to the accumulation of procyanidin precursors in a membrane-encapsulated cytoplasmic structure (Kitamura et al., 2010). Studies have shown that Arabidopsis TT19 protein is located not only in the nucleus and cytoplasm but also in the vacuolar membrane. TT19 protein can bind to anthocyanin monomer without catalyzing the formation of GSH-conjugated anthocyanin in vitro. First, TT19 mainly binds to cyanidin on the cytoplasmic surface of the endoplasmic reticulum, yet it also can bind to cyanidin-3-o-glucoside in the cytoplasm. Then, the TT19–anthocyanin complex is glycosylated and acetylated in the cytoplasm or on the surface of the vacuolar membrane (Springob et al., 2003; He et al., 2010). Finally, the modified anthocyanins are released from TT19, received by transporters located on the vacuolar membrane, and then finally stored in vacuoles. In addition to TT19 in Arabidopsis, some anthocyanin-related GSTF12 proteins also have been found in various plant species. In red cotton plants, silencing GhGSTF12 decreased the anthocyanin accumulation but did not affect the expression of anthocyanin biosynthesis-related genes (Shao et al., 2021). CmGSTF12 in Camelina sativa and PcGSTF12 in pear were identified to affect anthocyanin accumulation in vegetative tissues and proanthocyanidin accumulation in seeds (Wang et al., 2012; Zhang et al., 2020).

Radish (Raphanus sativus L., 2n = 18) is an important traditional vegetable in China and cultivated worldwide. The “Yanghua” radish has a characteristic red skin color (i.e., rich in anthocyanins). Some researchers have identified and analyzed the glutathione S-transferase (GST) family in radish but did not study the function of GST gene in detail (Gao et al., 2020). In this study, we identified a glutathione S-transferase gene RsGSTF12 whose function is crucial for anthocyanin sequestration in radish, which provides new insights into the anthocyanin sequestration mechanism in radish.

Previous research has shown that GSTs are essential for the secondary metabolism of plants. Benefiting from the released genomic sequences of radish, here 100 GST candidate genes in this plant were identified, which could be divided into seven distinct subfamilies (Figure 1). Finally, one member, RsGSTF12, was chosen among them for further study. By analyzing the phylogenetic tree, we failed to find any GST members in the GHR subfamily, unlike some AtGHRs and OsGHRs, a result consistent with the findings of a recent study (Gao et al., 2020). Not surprisingly, like the GSTs in other plants, the majority of RsGSTs belong to the tau and phi subfamilies, with those GST members responsible for anthocyanin sequestration generally distributed in the phi family (Figure 1; Shao et al., 2021). In addition, RsGSTF12 in radish has a higher homology with AtGSTF12 (AtTT19) (Figures 2A,B). The GSTs in radish and other species (e.g., sweet potato, strawberry, apple, and cotton) all have multiple, highly conserved regions, which suggests that RsGSTF12 has the potential to play a critical role in anthocyanin sequestration (Luo et al., 2018; Jiang et al., 2019; Li et al., 2019; Shao et al., 2021).

Subcellular localization analysis is a valuable method for understanding protein function. Previous research indicated that AtTT19 was localized in the nucleus, cytoplasm, vacuolar membrane, and endoplasmic reticulum membrane and that it can bind with cyanidin (Cya) in vitro and improve the water solubility of Cya in Arabidopsis (Sun et al., 2012). In our study, subcellular localization analysis in tobacco revealed that RsGSTF12 was located in the nucleus and cytosol, and RsGSTF12 could be overlapped with tonoplast and endoplasmic reticulum markers (Figure 2D), which is consistent with the characteristic of AcGST1 in kiwifruit (Liu et al., 2019). GSTs usually exist in plant tissues where anthocyanin accumulates considerably. Disparities in GST localization may arise from using various tissue types of different materials containing varying kinds and concentrations of flavonoids (Zhao, 2015). In radish, RsGSTF12 was highly expressed in the hypocotyl where abundant anthocyanin accumulated in the peripheral cortex and pericycle (Figures 3E,F; Hara et al., 2003). Further, RsGSTF12’s transcript abundance is positively correlated with anthocyanin accumulation (Figure 3G). However, the transcriptional level of RsGSTF12 was not completely consistent with the anthocyanin content. Whether protein level of RsGSTF12 is highly correlated with anthocyanin content still needs much more attention.

How RsGSTF12 mediates anthocyanin transferred from the endoplasmic reticulum to tonoplast is an important issue to better understand the anthocyanin sequestration mechanism in radish. A kind of GSH substrate [1-chloro-2,4-dinitrobenzene (CDNB)] could be conjugated by GSTs; however, the formation of GSH–anthocyanin has not been identified in plants so far (Marrs et al., 1995). Many efforts have been put in exploring whether GSTs contribute to anthocyanin sequestration by catalyzing the formation of GSH–anthocyanin. For example, AtTT19 in Arabidopsis, AcGST1 in kiwifruit, and ScGST3 in Senecio cruentus all did not possess the capability to catalyze the formation of GSH–anthocyanin (Sun et al., 2012; Liu et al., 2019; Cui et al., 2021). In this study, there was no GSH conjugation detected at 203 nm, while remarkable peaks were detected at 520 nm under AtTT19 and RsGSTF12 treatment (Figure 4B). These results indicated that RsGSTF12 could improve the water solubility of Cya in vitro to prevent its degradation. In addition, different GSTs are responsible for improving the water solubility of different anthocyanin monomers, which may be because the types of anthocyanin monomers in different species are different (Silva et al., 2016). In Arabidopsis, TT19 increased the water solubility of Cya and cyanidin-3-O-glycoside (C3G), while ScGST3 in Senecio cruentus was able to increase the water solubility of C3G and delphinidin-3-O-glucoside (D3G), rather than Cya and delphinidin (Dp) (Cui et al., 2021). Cya and its derivatives are the main monomers of anthocyanin in radish (Zhang et al., 2019). In this study, we demonstrated that RsGSTF12 may protect the Cya from undergoing degradation so as to improve the water solubility of Cya in solution. Whether RsGSTF12 could improve the water solubility of other anthocyanin monomers still needs further research to explore.

The transient expression system is a direct and important methodological tool used to understand the functions of proteins and is now widely used in various plants, such as apple, pear, and strawberry (Jiang et al., 2019; Castillejo et al., 2020; Zhao et al., 2021). In this study, when RsGSTF12 and GmMYB75 (a positive anthocyanin regulator in soybean) were co-transferred transiently into radish leaves, the leaves appeared much more anthocyanin compared with either of them injected (Figure 5A). Interestingly, the radish leaves injected with RsGSTF12 alone exhibited no significant color change in comparison with the empty vector. Moreover, we observed much more red autofluorescence indicative of anthocyanin content in central vacuoles and vesicle-like structures in those leaves injected with both RsGSTF12 and GmMYB75 vis-à-vis that injected with GmMYB75 or RsGSTF12 only (Figure 5D). One hypothesis to explain this was that when anthocyanin was significantly synthesized, the radish sprouts must also produce more of RsGSTF12 for it to participate in the anthocyanin sequestration process. Recently, new evidence suggested that anthocyanin can be bound and protected by GST, and then, it either was trapped inside the vesicle or formed a complex with the vesicle membrane, to move toward the surroundings of the vacuole (Pourcel et al., 2005; Zhao, 2015). These reports verified our conjecture, to a certain extent, that RsGSTF12 had the potential to be involved in the anthocyanin sequestration. Compared with tt19 mutant, overexpressing RsGSTF12 in tt19 could slightly increase anthocyanin content, while the anthocyanin did not completely recover to the wild-type level (Figure 6A). A similar phenomenon also was observed in maize. ZmGSTIII and Bz2 were both identified as possessing the ability of anthocyanin sequestration, while only knocking out Bz2 could totally result in anthocyanin accumulation deficiency (Marrs et al., 1995). Possible explanations for this phenomenon were that (i) there exist other GSTs responsible for the sequestration of anthocyanin in radish and (ii) RsGSTF12 has the potential to form homodimer with TT19 (or other GSTs in radish). Loss of TT19 in Arabidopsis led to reductions in anthocyanin and proanthocyanidin in that plant’s tissues and seed coat. Our results showed that the mutant phenotypes of the tt19 mutant could be partly restored by the expression of RsGSTF12. Like the functions of MdGSTF6 and GhGSTF12, RsGSTF12 could complement anthocyanin pigmentation in vegetative tissues, despite no color change evident in the seed coat (Figure 6) (Jiang et al., 2019; Zhao et al., 2021). Overall, RsGSTF12 is involved in the process of anthocyanin accumulation.

In this study, a glutathione S-transferase gene, RsGSTF12, was genomically selected that belongs to the phi (F) class. The endoplasmic reticulum- and tonoplast-localized protein RsGSTF12 was highly correlated with anthocyanin accumulation in radish sprouts, which spurred us to explore the functioning of RsGSTF12 in anthocyanin sequestration. Our results confirmed that RsGSTF12 has the ability to improve the water solubility of cyanidin in vitro. Furthermore, expressing RsGSTF12 in radish cotyledons and Arabidopsis seedlings revealed that RsGSTF12 could promote anthocyanin accumulation. We believe that our research findings may pave a fresh path toward enriching the anthocyanin sequestration network in radish cultivars.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

MN, NS, and JCu contributed to the conception of the study. MN conducted the experiments and analyzed the data. MN and CB interpreted the data and wrote the manuscript. JCh, WZ, YZ, and XZ modified the grammar and corrected errors. All authors approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.