Overexpression of RuFLS2 Enhances Flavonol-Related Substance Contents and Gene Expression Levels

As an emerging third-generation fruit, blackberry has high nutritional value and is rich in polyphenols, flavonoids and anthocyanins. Flavonoid biosynthesis and metabolism is a popular research topic, but no related details have been reported for blackberry. Based on previous transcriptome data from this research group, two blackberry flavonol synthase genes were identified in this study, and the encoded proteins were subjected to bioinformatics analysis. RuFLS1 and RuFLS2 are both hydrophobic acidic proteins belonging to the 2OG-Fe(II) dioxygenase superfamily. RuFLS2 was expressed at 27.93-fold higher levels than RuFLS1 in red–purple fruit by RNA-seq analysis. Therefore, RuFLS2-overexpressing tobacco was selected for functional exploration. The identification of metabolites from transgenic tobacco showed significantly increased contents of flavonoids, such as apigenin 7-glucoside, kaempferol 3-O-rutinoside, astragalin, and quercitrin. The high expression of RuFLS2 also upregulated the expression levels of NtF3H and NtFLS in transgenic tobacco. The results indicate that RuFLS2 is an important functional gene regulating flavonoid biosynthesis and provides an important reference for revealing the molecular mechanism of flavonoid accumulation in blackberry fruit.


Introduction
Blackberry (Rubus spp.) is a perennial shrub to small tree in the genus Rubus, family Rosaceae, that produces small berries [1]. Blackberry is native to North America and has a long history of cultivation in Europe and the United States. In 1986, blackberry was first introduced into China and popularized by the Institute of Botany, Chinese Academy of Sciences in Jiangsu Province. It has the advantages of drought resistance, vigorous growth, and easy reproduction [2,3]. The fruits are rich in nutrients such as sugars, vitamin C, vitamin E, polyphenols, and flavonoids [4,5] and have extremely high nutritional and medicinal value, including antibacterial, anti-inflammatory, antioxidant and antiaging effects [6,7]. These fruits are widely used in medical and health care fields with good prospects for further applications.
Flavonoids are a class of plant-produced secondary metabolites belonging to the group of polyphenolic compounds with a typical C6-C3-C6 carbon skeleton [8]. Flavonoids can be divided into six subclasses, including flavonoids, isoflavones, dihydroflavonoids, flavonols, flavonones, and anthocyanins [9]. Flavonols can be modified by acylation, hydroxylation, and glycosylation. These modifications can coexist, resulting in diverse flavonol structures [10], so these compounds have a variety of biological activities, such as antioxidation, antiviral, and anticancer activities [11][12][13]. In addition, these compounds can aid in resistance to ultraviolet radiation damage [14], enhance the ability of plant roots 2 of 16 to resist drought stress [15], and play an important biological role in plant growth and development. Thus, flavonols are among the important substances that participate in plant resistance to external biotic and abiotic stresses.
Flavonol synthase (FLS) is one of the key enzymes in the biosynthetic pathway of flavonoids. The FLS genes play an important role in the appearance of white or yellow plant organs, and the overexpression of the Muscari aucheri FLS gene resulted in a white flower color [16]. Plant FLS proteins are often encoded by multiple gene copies, and the number of copies is different in different plants [17]. The expression level of FLS has significant spatial and temporal differences in different tissue parts, metabolic stages, and growth and developmental stages of plants [18], which affects the content of flavonoids in plants. FLS genes have been identified and functionally studied in some plants, with similar functions found in most plants. For example, OcFLS1 and OcFLS2 of Ornithogalum caudatum can both catalyze the production of flavonol from dihydroflavonol, and both have flavanone 3-hydroxylase (F3H) activity [19], while several FLSs in some plants have different functions. In Arabidopsis thaliana, FLS1 is involved in the biosynthesis of flavonoids, while FLS3 catalyzes the production of flavonols [20]. Therefore, although FLS is involved in the synthesis of flavonols, its expression pattern and specific regulatory mechanism in blackberry still need to be further studied and verified.
In this study, based on the transcriptome data of immature and mature blackberry fruits obtained previously by this research group [21], the FLS gene showed significant differential expression in the flavonoid synthesis pathway and was selected for further research. Its effect on flavonoid synthesis in blackberry fruit was investigated along with the effects of related substances. The purpose of this study was to (i) identify, clone and bioinformatically analyze the potential functions of FLSs; (ii) analyze the expression patterns of differentially expressed FLSs in different tissues and fruit developmental stages in blackberry; and (iii) study the function of FLS overexpression and explore its effects on the accumulation of related metabolites and gene expression to provide a basis for the study of flavonoid biosynthesis and FLS function in blackberry fruit.

Gene Structure, Physicochemical Properties and Domains of RuFLSs
From the transcriptome data of blackberry at different developmental stages (SRA: PRJNA701162) [21], we found three genes that may encode FLS proteins: one gene did not contain a verified ORF, while the other two genes are named RuFLS1 and RuFLS2. Structural analysis showed that the ORF sequences of RuFLS1 and RuFLS2 were 1017 and 993 bp in length, respectively. The physicochemical properties of RuFLS1 and RuFLS2 proteins were analyzed by ProtParam software (SIB, Geneva, Switzerland), including protein molecular weight, theoretical isoelectric point, charged residue number, instability index, fat index, and hydrophilicity ( Table 1). The instability index was 36.45 and 34.70, and the instability coefficient was less than 40; thus, both proteins were considered to be stable. In addition, the total hydrophilicity of RuFLS1 and RuFLS2 was predicted to be −0.494 and −0.341 (Table 1), respectively, by ProtScale online software (SIB, Geneva, Switzerland) (Figure 1a,b), suggesting that both proteins are hydrophobic and acidic. The functional domains of RuFLS1 and RuFLS2 were predicted and found to contain a DIOX_N functional domain and a 2OG-FeII_Oxy functional domain (Figure 1c,d), indicating that they belong to the α-oxoglutarate-dependent dioxygenase family. The DIOX_N functional domain of RuFLS1 is located at amino acids 38-152, and the 2OG-FeII_Oxy functional domain is located at amino acids 190-287. The DIOX_N functional domain of RuFLS2 is located at amino acids 40-148, and the 2OG-FeII_Oxy functional domain is located at amino acids 192-292. The positions of the functional domains of different members are not exactly the same. RuFLS1 and RuFLS2 have no transmembrane region and are both extramembrane proteins. acids 192-292. The positions of the functional domains of different members are not exactly the same. RuFLS1 and RuFLS2 have no transmembrane region and are both extramembrane proteins.

Multiple Sequence Alignment, Phylogenetic Analysis, and Motif Composition of RuFLSs
Amino acid alignment of the FLS protein sequences of multiple plants by DNAMAN software (Lynnon Biosoft, California, USA) showed that RuFLS1 and RuFLS2, as well as sequences from Prunus dulcis (XM_034342566.1), Prunus persica (XM_007222519.2), Paeonia lactiflora (KM259902.3), Fragaria vesca subsp. Vesca (XM_011460471.1) and Camellia sinensis (DQ198089.1) were highly conserved. In the figure, the red dots indicate important amino acid residues that bind to the substrate dihydroquercetin, and the blue dots indicate amino acid residues that are important for binding to ferrous ions ( Figure 2a). Interestingly, RuFLS1 was mutated at multiple substrate-binding sites and may have lost part of its protein function during blackberry evolution. Differential gene expression analysis based on blackberry transcriptome data showed that RuFLS1 was not differentially expressed, which may have been related to the changes in amino acid residues at key sites of the RuFLS1 protein rather than in key genes regulating important secondary metabolites such as flavonoids. RuFLS2 is a significantly differentially expressed enzymatic gene, and it is speculated that this gene may play an important role in the accumulation of flavonoids

Multiple Sequence Alignment, Phylogenetic Analysis, and Motif Composition of RuFLSs
Amino acid alignment of the FLS protein sequences of multiple plants by DNAMAN software (Lynnon Biosoft, California, USA) showed that RuFLS1 and RuFLS2, as well as sequences from Prunus dulcis (XM_034342566.1), Prunus persica (XM_007222519.2), Paeonia lactiflora (KM259902.3), Fragaria vesca subsp. Vesca (XM_011460471.1) and Camellia sinensis (DQ198089.1) were highly conserved. In the figure, the red dots indicate important amino acid residues that bind to the substrate dihydroquercetin, and the blue dots indicate amino acid residues that are important for binding to ferrous ions ( Figure 2a). Interestingly, RuFLS1 was mutated at multiple substrate-binding sites and may have lost part of its protein function during blackberry evolution. Differential gene expression analysis based on blackberry transcriptome data showed that RuFLS1 was not differentially expressed, which may have been related to the changes in amino acid residues at key sites of the RuFLS1 protein rather than in key genes regulating important secondary metabolites such as flavonoids. RuFLS2 is a significantly differentially expressed enzymatic gene, and it is speculated that this gene may play an important role in the accumulation of flavonoids during blackberry fruit ripening. Therefore, we conducted an in-depth study on the function of RuFLS2 to reveal the mechanism by which RuFLS2 mediates the synthesis and metabolism of flavonoid-related secondary metabolites in blackberry.
during blackberry fruit ripening. Therefore, we conducted an in-depth study on the function of RuFLS2 to reveal the mechanism by which RuFLS2 mediates the synthesis and metabolism of flavonoid-related secondary metabolites in blackberry. The conserved motifs of the two RuFLS proteins were predicted. The results showed that both RuFLS1 and RuFLS2 had eight identical conserved motifs, and their positions within the sequence were basically the same (Figure 2b). The green part in the figure is the DIOX_N superfamily, a highly conserved N-terminal region of proteins with 2-oxoglutarate/Fe(II)-dependent dioxygenase activity. The blue area is the Fe(II)-dependent oxygenase superfamily, and the red area is the 2OG superfamily. The latter two regions combine to form the 2OG-Fe(II)_oxygenase superfamily motif, which is consistent with the predicted domain and the FLS protein family in plants. Two consistently conserved motifs, "IGTKMN" and "YRNEGLR", are arranged in different positions in RuFLS1 and RuFLS2, which may lead to differences in the functions of these proteins.
A phylogenetic tree was constructed with MEGAX with the NJ method. The results showed that RuFLS1 and RuFLS2 were closely related to flavonol synthase/flavanone 3hydroxylase proteins from Rosaceae species such as Fragaria vesca subsp. vesca, Prunus avium, Prunus dulcis, Prunus persica, and Prunus mume. The flavonol synthases of Brassicaceae and the flavonol synthases of Liliaceae and Vitaceae are clustered into a branch, which is in line with the relationship based on species evolution ( Figure 3). Both are Rosaceae plants, and the relationship between flavonol synthase proteins in Prunus persica, Malus domestica, and Prunus avium with RuFLS1 and RuFLS2 is distant. It is speculated that RuFLS1 and RuFLS2 are not monofunctional flavonol synthase enzymes but may be bifunctional enzymes with both FLS and F3H activities. The conserved motifs of the two RuFLS proteins were predicted. The results showed that both RuFLS1 and RuFLS2 had eight identical conserved motifs, and their positions within the sequence were basically the same (Figure 2b). The green part in the figure is the DIOX_N superfamily, a highly conserved N-terminal region of proteins with 2oxoglutarate/Fe(II)-dependent dioxygenase activity. The blue area is the Fe(II)-dependent oxygenase superfamily, and the red area is the 2OG superfamily. The latter two regions combine to form the 2OG-Fe(II)_oxygenase superfamily motif, which is consistent with the predicted domain and the FLS protein family in plants. Two consistently conserved motifs, "IGTKMN" and "YRNEGLR", are arranged in different positions in RuFLS1 and RuFLS2, which may lead to differences in the functions of these proteins.
A phylogenetic tree was constructed with MEGAX with the NJ method. The results showed that RuFLS1 and RuFLS2 were closely related to flavonol synthase/flavanone 3-hydroxylase proteins from Rosaceae species such as Fragaria vesca subsp. vesca, Prunus avium, Prunus dulcis, Prunus persica, and Prunus mume. The flavonol synthases of Brassicaceae and the flavonol synthases of Liliaceae and Vitaceae are clustered into a branch, which is in line with the relationship based on species evolution ( Figure 3). Both are Rosaceae plants, and the relationship between flavonol synthase proteins in Prunus persica, Malus domestica, and Prunus avium with RuFLS1 and RuFLS2 is distant. It is speculated that RuFLS1 and RuFLS2 are not monofunctional flavonol synthase enzymes but may be bifunctional enzymes with both FLS and F3H activities.

Expression Levels of RuFLSs in Different Tissues and Fruit Developmental Stages of Blackberry
We analyzed the expression levels of RuFLS1 and RuFLS2 in the roots, stems, leaves, flowers, fruiting branches and fruits of blackberry and found that the expression level of RuFLS2 was higher in fruits, while the expression level of RuFLS1 was higher in stems and leaves. In comparison, the expression level of RuFLS1 in fruits was significantly lower than that of RuFLS2, but it was higher in other tissues. It was speculated that RuFLS2 might play a major role in blackberry fruits ( Figure 4a). Based on this, we also tested the expression level of the RuFLS genes at different fruit developmental stages and found that the expression level of RuFLS1 was significantly lower than that of RuFLS2 at each fruit developmental stage. The RuFLS2 gene showed more significant differences during fruit development. The expression level of RuFLS2 increased sharply as the color of the fruit changed from red to purple. It then decreased rapidly but was still higher than that in the early stage of fruit development (Figure 4b).

Expression Levels of RuFLSs in Different Tissues and Fruit Developmental Stages of Blackberry
We analyzed the expression levels of RuFLS1 and RuFLS2 in the roots, stems, leaves, flowers, fruiting branches and fruits of blackberry and found that the expression level of RuFLS2 was higher in fruits, while the expression level of RuFLS1 was higher in stems and leaves. In comparison, the expression level of RuFLS1 in fruits was significantly lower than that of RuFLS2, but it was higher in other tissues. It was speculated that RuFLS2 might play a major role in blackberry fruits ( Figure 4a). Based on this, we also tested the expression level of the RuFLS genes at different fruit developmental stages and found that the expression level of RuFLS1 was significantly lower than that of RuFLS2 at each fruit developmental stage. The RuFLS2 gene showed more significant differences during fruit development. The expression level of RuFLS2 increased sharply as the color of the fruit changed from red to purple. It then decreased rapidly but was still higher than that in the early stage of fruit development (Figure 4b).

Expression Levels of RuFLSs in Different Tissues and Fruit Developmental Stages of Blackberry
We analyzed the expression levels of RuFLS1 and RuFLS2 in the roots, stems, leaves, flowers, fruiting branches and fruits of blackberry and found that the expression level of RuFLS2 was higher in fruits, while the expression level of RuFLS1 was higher in stems and leaves. In comparison, the expression level of RuFLS1 in fruits was significantly lower than that of RuFLS2, but it was higher in other tissues. It was speculated that RuFLS2 might play a major role in blackberry fruits (Figure 4a). Based on this, we also tested the expression level of the RuFLS genes at different fruit developmental stages and found that the expression level of RuFLS1 was significantly lower than that of RuFLS2 at each fruit developmental stage. The RuFLS2 gene showed more significant differences during fruit development. The expression level of RuFLS2 increased sharply as the color of the fruit changed from red to purple. It then decreased rapidly but was still higher than that in the early stage of fruit development (Figure 4b). . Different lowercase letters indicate significant differences between different harvest periods (p < 0.05). Different uppercase letters represent significant differences between two RuFLS genes (p < 0.05).

Changes in Flavonoid, Phenol, and Anthocyanin Contents in Blackberry Fruits at Different Developmental Stages
To gain an in-depth understanding of the changes in flavonoid-related substances during the development of blackberry fruits, we also measured the contents of total phenols, flavonoids and anthocyanins in blackberry fruits at different stages. The results showed that the content of total phenols and flavonoids changed in a similar way, showing a sharp decline in the early stage and little change in the later stage. The content increased slightly in the purple fruit stage, and the content in the green fruit stage was significantly higher than that in the other stages. The total phenolic content reached 40.68 mg·g −1 FW, and the flavonoid content reached 11.50 mg·g −1 FW. The accumulation of anthocyanins in the early stage of fruit development was lower and tended toward zero, and the amount of anthocyanins increased sharply from the red and purple fruit stage to the purple fruit stage, reaching a maximum of 2.60 mg·g −1 FW at the purple fruit stage. This result indicated that anthocyanins are mainly synthesized and accumulate in the later stage of fruit development ( Table 2). Table 2. Variation in flavonoid, phenol, and anthocyanin contents at different developmental stages in blackberry.

Fruit Color
Phenols Note: Different letters indicate significant differences between the different developmental stages of blackberry (p < 0.05).

Heterologous Overexpression of RuFLS2 in Tobacco
To reveal the biological functions of RuFLS2, we overexpressed RuFLS2 in tobacco. Three transgenic lines with relatively vigorous growth were randomly selected for expression level detection, and it was found that RuFLS2 was highly expressed in the transgenic plants ( Figure 5a). To explore the effect of RuFLS2 heterologous expression on key enzyme genes in the tobacco flavonoid biosynthesis pathway, we detected the expression level of tobacco NtFLS and the upstream gene NtF3H (Figure 5d). The results showed that the expression level of the NtFLS gene in transgenic tobacco was significantly increased (Figure 5b), and the change trend of the expression level of the F3H gene in transgenic tobacco across the three lines was consistent with that of FLS (Figure 5c). The upregulation of NtFLS and NtF3H in the RuFLS-1 and RuFLS-2 lines was significantly higher than that in RuFLS-3. The results confirmed that heterologous expression of RuFLS2 upregulated the expression of the tobacco NtFLS gene.

Flavonoid and Phenol Content Analysis in Genetically Modified Tobacco
To analyze the effect of overexpression of the RuFLS2 gene on flavonoid-related metabolites, the contents of flavonoids and total phenolics in the leaves of the transgenic FLS and WT lines were determined. The results showed that the flavonoid content of one transgenic tobacco line was the same as that of the WT, and the content of the other two lines increased by 1.3-and 1.8-fold (Figure 6a). The content of total phenol was similar to that of flavonoids, where one line had the same content as the WT, and the other two lines showed increases of 1.25-and 2.84-fold (Figure 6b). Overall, the overexpression of the RuFLS2 gene increased the contents of flavonoids and phenols in tobacco, indicating that the overexpression of RuFLS2 increased the contents of flavonoid-related substances in plants. As there are many kinds of flavonoids in plants, a metabolome analysis needs to be performed in the future.

Flavonoid and Phenol Content Analysis in Genetically Modified Tobacco
To analyze the effect of overexpression of the RuFLS2 gene on flavonoid-related metabolites, the contents of flavonoids and total phenolics in the leaves of the transgenic FLS and WT lines were determined. The results showed that the flavonoid content of one transgenic tobacco line was the same as that of the WT, and the content of the other two lines increased by 1.3-and 1.8-fold (Figure 6a). The content of total phenol was similar to that of flavonoids, where one line had the same content as the WT, and the other two lines showed increases of 1.25-and 2.84-fold (Figure 6b). Overall, the overexpression of the RuFLS2 gene increased the contents of flavonoids and phenols in tobacco, indicating that the overexpression of RuFLS2 increased the contents of flavonoid-related substances in plants. As there are many kinds of flavonoids in plants, a metabolome analysis needs to be performed in the future.

Multivariate Statistical Analysis of Metabolites in Genetically Modified Tobacco
PLS-DA can maximize the distinction between groups and is beneficial for fin differentially expressed metabolites (DEMs). This study showed that the tobacco me lites in the WT control group and the FLS transgenic group were R 2 X = 0.678, R 2 Y = 0 2 Figure 6. Flavonoid (a) and phenol (b) contents in WT and RuFLS2-overexpressing tobacco. WT represents wild-type tobacco. RuFLS2-1, RuFLS2-2 and RuFLS2-3 represent three different transgenic tobacco lines. Significant differences between means (p < 0.05) are indicated by lowercase letters (a, b and c) above the bars.

Multivariate Statistical Analysis of Metabolites in Genetically Modified Tobacco
PLS-DA can maximize the distinction between groups and is beneficial for finding differentially expressed metabolites (DEMs). This study showed that the tobacco metabolites in the WT control group and the FLS transgenic group were R 2 X = 0.678, R 2 Y = 0.999, and Q 2 Y = 0.763 in positive ion mode ( Figure 7a); in negative ion mode, they were R 2 X = 0.427, R 2 Y = 0.998, and Q 2 Y = 0.688 (Figure 7b), in line with our expectations for the model of the experimental data. These results indicate that the parameters of the PLS-DA model established in this study were reasonable and stable and could be used for metabolomic analysis. OPLS-DA is a derivative algorithm of PLS-DA (Figure 7c,d). Both multivariate statistical analyses of supervised pattern recognition found that the FLS group was on the left side of the confidence zone, while the WT group was on the right side of the confidence zone, indicating that both models can effectively distinguish the metabolites of transgenic and nontransgenic tobacco. In addition, we validated the OPLS-DA model by permutation and found that the OPLS-DA model had good reliability (Figure 7e-h), and the subsequent model checking and differential metabolite screening were analyzed using the OPLS-DA results.

DEM-Related Flavonoid Synthesis in Tobacco
To further explore the effect of overexpression of the RuFLS2 gene on flavonoid-related metabolites, we focused on the analysis of DEMs related to flavonoid synthesis in the transgenic FLS group and WT group. The results showed that six DEMs were detected in the transgenic tobacco flavonoid biosynthesis pathway, and the contents of the other five substances, except naringenin-7-O-glucoside, were decreased. A total of seven DEMs were detected in the flavone and flavonol synthesis pathways, among which the contents of apigenin 7-glucoside, kaempferol 3-O-rutinoside, astragalin and quercitrin were significantly increased, while the contents of luteolin 7-glucoside, quercetin 3-glucoside and ligustroflavone were decreased (Table 3). This result indicated that the overexpression of the RuFLS2 gene consumed upstream flavonoids and significantly increased the flavonol content in plants.

DEM-Related Flavonoid Synthesis in Tobacco
To further explore the effect of overexpression of the RuFLS2 gene on flavonoidrelated metabolites, we focused on the analysis of DEMs related to flavonoid synthesis in the transgenic FLS group and WT group. The results showed that six DEMs were detected in the transgenic tobacco flavonoid biosynthesis pathway, and the contents of the other five substances, except naringenin-7-O-glucoside, were decreased. A total of seven DEMs were detected in the flavone and flavonol synthesis pathways, among which the contents of apigenin 7-glucoside, kaempferol 3-O-rutinoside, astragalin and quercitrin were significantly increased, while the contents of luteolin 7-glucoside, quercetin 3-glucoside and ligustroflavone were decreased (Table 3). This result indicated that the overexpression of the RuFLS2 gene consumed upstream flavonoids and significantly increased the flavonol content in plants.

Discussion
As a third-generation fruit, blackberry is rich in a variety of flavonoids and has high research value [22]. However, research progress on the enzymatic genes related to flavonoid synthesis in blackberry is still lacking. In this study, we identified two members of the FLS gene family from the blackberry transcriptome, named RuFLS1 and RuFLS2, and performed a bioinformatics analysis on them. The results showed that the amino acid sequences of the RuFLS1 and RuFLS2 proteins are very similar; both contain conserved functional domains, but RuFLS1 is mutated at multiple substrate binding sites, which may result in the loss of some protein functions. Relevant studies have shown that in plants, only some members of the multicopy gene family have functions [23]. For example, the FLS gene family in Arabidopsis contains six members, but only FLS1 can affect the biosynthesis of flavonoids [24]. Differential gene expression analysis based on blackberry transcriptome data showed that RuFLS2 was a significantly differentially expressed enzymatic gene. Therefore, we performed preliminary screening and speculated that the FLS protein that mainly functions during blackberry fruit ripening may be RuFLS2.
To understand the evolutionary relationship between RuFLS1, RuFLS2 and FLSs of other species, the phylogenetic tree ( Figure 3) showed that RuFLS2, RuFLS1 and the flavonol synthase/flavanone 3-hydroxylase proteins of other Rosaceae plants, such as Fragaria vesca subsp. Vesca, Prunus avium, and Prunus persica, are closely related. We speculated that blackberry RuFLS2 may have some F3H functions in addition to FLS functions. In the flavonoid biosynthesis pathway, F3H catalyzes the production of flavanones to dihydroflavonols, while FLS is located downstream of F3H and catalyzes the production of flavonols using dihydroflavonols as substrates ( Figure 5). Both FLS and F3H belong to the 2OG-Fe(II)_oxygenase superfamily and have functional similarities [25]. Several experiments have shown that FLS is a bifunctional dioxygenase with F3H activity and FLS activity [26]. FLS extracted from Citrus unshiu can not only convert dihydrokaempferol to kaempferol but also directly use pomelo. Cortexin is a substrate used in the efficient production of kaempferol. Two FLSs isolated from Ornithogalum caudatum also have similar functions. Fe 2+ is not necessary for F3H activity, but OcFLS directly catalyzes flavanones to flavonols and requires a sufficient amount of Fe 2+ to participate in the reaction [27].
The expression of the FLS gene has obvious spatial and temporal differences. FLS is mainly highly expressed in the petals of ornamental plants [28] and regulates changes in flower color. In economic fruit trees, such as grapes, the expression levels and enzyme activities of FLSs are highly positively correlated with the content of flavonoids in the fruit, and the expression level is higher at the veraison and ripening stages of grapes. In this study, it was found that RuFLS1 was highly expressed in roots, stems and leaves. The expression level of RuFLS2 in blackberry fruit was significantly higher than that in other tissues. During the development of blackberry fruit, the expression level of RuFLS1 was always lower than that of RuFLS2. The expression level of RuFLS2 was lower in the early stage of fruit development and was higher when the fruit epidermis developed a red-purple color. The expression level increased sharply during the red-purple period and decreased rapidly when the fruit turned dark purple and became fully ripened, which was consistent with the changes in the expression level of FLS in grapes [29].
To further explore the potential function of the differentially expressed gene RuFLS2 in plants, we constructed an overexpression vector and successfully overexpressed the RuFLS2 gene heterologously in tobacco. We examined the contents of flavonoids and phenols in transgenic tobacco. The results showed that overexpression of RuFLS2 significantly increased the contents of flavonoid-related substances in transgenic tobacco. Based on these results, we performed a metabolomic analysis to further clarify the function of RuFLS2. The content of metabolites in the flavonol synthesis pathway in transgenic tobacco was significantly increased, consistent with the experimental results of Park et al. [30]. The heterologous expression of AcFLS in Allium cepa significantly increased the flavonol content in tobacco. Our study also found that the high heterologous expression of RuFLS2 significantly increased the expression of the NtFLS and NtF3H genes in transgenic tobacco lines. NtF3H, as the upstream gene of NtFLS, was negatively regulated by NtFLS, and its expression level increased. Further investigation suggested that the biosynthesis of flavonoids was regulated not only by structural genes such as F3H and FLS, but also by transcription factors. The regulatory mechanism of flavonol biosynthesis was initially studied in Arabidopsis [31]. The MYB transcription factor regulates the expression of FLS to control the flavonol content in plants. Overexpression of AtMYB12 can significantly increase the flavonoid content of Arabidopsis thaliana, and the flavonoid content of the AtMYB12 mutant is significantly reduced [32]. Related studies have also found that apple MdMYB22 [33] can directly bind to the promoter of FLS and activate its expression, and overexpression of pear PbMYB12b can significantly increase the synthesis and accumulation of quercetin glycosides [34], indicating that the transcription factor MYB plays an important role in flavonoid synthesis. In addition to the MYB transcription factor family, other transcription factor families, such as bZIP and WRKY [35], are also involved in the regulation of flavonoid biosynthesis. In short, the regulatory mechanism of flavonoid biosynthesis is very complex. To explore the molecular mechanism of flavonoid synthesis in blackberry, it is necessary to further study its transcriptional regulation.

Conclusions
In this study, we successfully identified and isolated the RuFLS2 gene containing a 993-bp ORF. The RuFLS2 gene encodes a hydrophobic acidic protein composed of 331 amino acids belonging to the 2OG-Fe(II)_oxygenase superfamily. A temporal and spatial expression analysis showed that RuFLS2 was highly expressed in blackberry fruit, and the expression peaked when the fruit developed a red-purple color. Heterologous overexpression in tobacco revealed that the synthetic content of flavonols, such as apigenin 7-glucoside, kaempferol 3-O-rutinoside, astragalin, and quercitrin, increased; additionally, the high expression of RuFLS2 increased the gene expression levels of NtF3H and NtFLS in transgenic tobacco. The results indicated that RuFLS2 is a key enzymatic gene regulating the synthesis of flavonoids in blackberry and can affect the expression of other key structural genes in the flavonoid synthesis pathway, providing a new direction for studying flavonoid biosynthesis and its regulatory mechanisms.

Plant Materials and Growth Conditions
The experimental material was the blackberry cultivar 'Chester' from the Baima Experimental Base in Lishui District, Nanjing City (119 • 11 E, 31 • 36 N). The study area belongs to the subtropical monsoon climate zone. Specific growth conditions are listed in Supplementary Table S1. The roots, stems, leaves, flowers, fruiting branches and fruits of blackberry were picked as plant materials for the analysis of gene expression in different tissue parts. In addition, the fruit developmental stages were distinguished according to the color of the fruit. Green fruits, green-red fruits, red fruits, red-purple fruits and purple fruits were picked (Figure 8). All experiments were biologically repeated at least three times. After collection, samples were stored in liquid nitrogen and brought back to the laboratory for storage in a −80 • C refrigerator. According to the color of the fruit, the development of the fruit can be divided into five stages: green fruits, green-red fruits, red fruits, red-purple fruits, and purple fruits.

Total RNA Extraction and cDNA Reverse Transcription
Total RNA was extracted using a general plant total RNA rapid extraction kit (Bio-TeKe, Beijing, China). Total RNA from different fruit parts and developmental stages was reverse transcribed to synthesize cDNA. A subsequent expression analysis was performed using cDNA diluted to a concentration of approximately 300 ng/µL as a template.

Cloning of the RuFLS Gene
Based on the differentially expressed RuFLS unigene sequences identified in the blackberry transcriptome data, the open reading frame (ORF) and third-generation fulllength transcript sets were obtained by BioXM2.6 software (Ji Feng, Nanjing, China) analysis for correction, and primers were designed with Oligo 6.0 software (Reachsoft, Beijing, China). The ORF sequence was cloned using the high-fidelity PCR enzyme PrimeSTAR Figure 8. Appearance of blackberry cv. 'Chester' fruits at various developmental stages. According to the color of the fruit, the development of the fruit can be divided into five stages: green fruits, green-red fruits, red fruits, red-purple fruits, and purple fruits.

Total RNA Extraction and cDNA Reverse Transcription
Total RNA was extracted using a general plant total RNA rapid extraction kit (BioTeKe, Beijing, China). Total RNA from different fruit parts and developmental stages was reverse transcribed to synthesize cDNA. A subsequent expression analysis was performed using cDNA diluted to a concentration of approximately 300 ng/µL as a template.

Cloning of the RuFLS Gene
Based on the differentially expressed RuFLS unigene sequences identified in the blackberry transcriptome data, the open reading frame (ORF) and third-generation fulllength transcript sets were obtained by BioXM2.6 software (Ji Feng, Nanjing, China) analysis for correction, and primers were designed with Oligo 6.0 software (Reachsoft, Beijing, China). The ORF sequence was cloned using the high-fidelity PCR enzyme PrimeSTAR Max DNA Polymerase (TaKaRa, Dalian, China). A 50 µL PCR system was utilized as follows: 25 µL of Primer Star Max, 1 µL of forward and reverse primers, 1 µL of cDNA template, and 22 µL of ddH 2 O. The PCR program was as follows: 98 • C for 3 min; 35 cycles of 98 • C for 10 s, 55 • C for 5 s, and 72 • C for 15 s; 72 • C for 3 min; and a 4 • C hold. The amplified product was ligated to the vector according to the requirements of the pClone007 Blunt Vector Kit (TSINGKE, Nanjing, China) and transferred to competent Escherichia coli cells; the potential positive bacteria were cultured at 37 • C and 250 rpm for 6-8 h on a shaker with 2 × T5 Super PCR Mix (Colony) (TSINGKE, Nanjing, China) and used for bacterial liquid PCR verification. The positive bacterial liquid was sent to Tsingke Biotechnology Co., Ltd. for first-generation Sanger sequencing verification. Primer sequences are listed in Supplementary Table S2.

Bioinformatics Analysis
The protein sequences encoded by the blackberry FLS genes were analyzed using DNA-MAN software, and their physical and chemical properties and protein hydrophilicity were predicted using ExPASy software (SIB, Geneva, Switzerland) (https://web.expasy.org/, accessed on 19 February 2022). The MEME online tool (https://meme-suite.org/meme/ tools/meme/, accessed on 24 March 2022) was used to predict the motifs of RuFLS proteins. The SMART online tool (http://smart.embl-heidelberg.de/, accessed on 19 February 2022) and NCBI online tool CDART (Conserved Domain Architecture Retrieval Tool) (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi#opennewwindow/, accessed on 19 February 2022) were used to analyze the domains of RuFLSs, including transmembrane domains, functional domains and conserved structure areas. In addition, RuFLS protein sequence alignment (similarity) was performed using MEGAX (MEGA, AZ, USA), and a phylogenetic tree was constructed by the neighbor-joining with 1000 bootstrap replicates.

qRT-PCR Analysis
qRT-PCRs were performed with TB Green Premix Taq II (Tli RNASEH Plus) (Takara, Dalian, China). A 15 µL reaction system was used as follows: 7.5 µL of TB Green Premix Taq II fluorescent dye, 1 µL of cDNA template, 0.6 µL of upstream and downstream primers, and 5.3 µL of ddH2O. The qRT-PCR program was as follows: 95 • C for 2 min; 40 cycles of 95 • C for 10 s, 60 • C for 10 s, and 72 • C for 15 s; and melting curve generation for 6 s. For blackberry, 18S was the internal reference gene (Supplementary Table S2), and for transgenic tobacco seedlings, NtActin was the internal reference gene (Supplementary  Table S2) for real-time quantitative PCR. The 2 −∆∆CT method was used for quantitative data analysis.

Determination of Polyphenol, Flavonoid and Anthocyanin Contents
The total phenol content was determined according to the method of Cheok et al. [36] using a commercial plant total phenol test kit (Nanjing Jiancheng Institute of Bioengineering, Nanjing, China). Fresh samples were ground into powder with liquid nitrogen. Then, 0.1 g of powder was taken, added to 2 mL of extract (60% ethanol aqueous solution), vortexed and mixed for 3 min, extracted at 60 • C for 30 min, and centrifuged at 4000 rpm·min −1 for 10 min. The supernatant was transferred to a centrifuge tube, with reagents added as needed, mixed, and allowed to stand at room temperature for 10 min. The absorbance of each tube was measured at 760 nm. The hydromethanolic gallic acid solution was freshly prepared in a series of concentrations (0.1-1 mM) and tested in parallel to establish a calibration curve. The total phenolic content of the samples was calculated as mg of gallic acid equivalent per g fresh sample (mg/g FW).
The flavonoid content was detected by the aluminum chloride calorimetry method [37] using a commercial plant flavonoid test kit (Nanjing Jiancheng Institute of Bioengineering, Nanjing, China). Samples were washed with normal saline, dried, ground with liquid nitrogen to powder, weighed to 0.05 g and combined with 2 mL of 60% ethanol. The solution was shaken, extracted at 60 • C for 2 h, and centrifuged at 10,000 rpm for 10 min at room temperature. The supernatant was collected for testing. According to the standard curve drawing procedure, each sample was tested three times, and the mean value was taken and substituted into the standard curve to calculate the flavonoid content, which was expressed as mg per g of FW (mg/g FW).
The anthocyanin content was determined by the pH differential method [38] with a slight modification: 50 g of blackberry fruit homogenate was taken, 3 g of homogenate solution was added, 30 mL of 50% ethanol extract was added, and the mixture was shaken well and ultrasonically treated at 60 Hz for 20 min at 35 • C. The mixture was then centrifuged at 5000 rpm for 5 min; then, 0.5 mL of the supernatant was taken and placed in a 10 mL centrifuge tube, 4.5 mL of pH 1.0 buffer was added, and the mixture was reacted at room temperature for 20 min and adjusted to zero with double distilled water. The absorbance was measured at 510 nm. A standard curve was utilized to calculate the anthocyanin content. The equation obtained for the calibration curve of the cyanidin-3-glucoside standard (C832095-5mg) (Nanjing Shoude Biotechnology Co., Ltd., Nanjing, China) solution in the range of 0-0.6 mg/mL was y = 0.0455x − 0.0881 (R 2 = 0.9996). The results are expressed as mg per g of FW (mg/g FW).

qRT-PCR Analysis
The Agrobacterium culture containing the overexpression vector with the target fragment was expanded to 50 mL in liquid culture containing kanamycin (Solarbio, Beijing, China) and rifampicin (Solarbio, Beijing, China) antibiotics at 28 • C. The OD value of the bacterial culture was measured and maintained at 0.5-0.8. The bacterial solution was transferred to a 50 mL centrifuge tube and centrifuged at 5000 rpm for 10 min, and the supernatant was then discarded. The bacterial pellet was resuspended in 40 mL of MS0 (4.43 g/L MS (Beijing XMJ Scientific Co., Ltd, Beijing, China) + 100 µmol/L acetosyringone (Solarbio, Beijing, China), pH 5.8). The leaves of "Nicotiana tabacum cv. Wisconsin 38" that grew vigorously were collected and cut into small (1 cm 2 ) pieces. The small leaf pieces were immersed in the prepared bacterial solution for 6-12 min and shaken several times during this period. The leaf pieces were spread on MS1 medium (4.43 g/L MS + 25 g/L sucrose + 5.6 g/L agarose (Solarbio, Beijing, China) + 2.0 mg/L 6-BA (Solarbio, Beijing, China) + 0.1 mg/L NAA (Solarbio, Beijing, China) + 100 µmol/L acetosyringone, pH 5.8) and cultured at 25 • C for 2-3 d in the dark. They were then transferred to MS2 medium (4.43 g/L MS + 25 g/L sucrose + 5.6 g/L agarose + 2.0 mg/L 6-BA + 0.1 mg/L NAA + 100 mg/L Timentin (Solarbio, Beijing, China) + 100 mg/L kanamycin, pH 5.8), placed under a 16 h light/8 h dark cycle, and cultured at 25 • C until shoots formed on the callus. The shoots were excised and transferred to MS rooting medium (4.43 g/L MS + 25 g/L sucrose + 5.6 g/L agarose + 100 mg/L Timentin, pH 5.8). The leaves were collected to extract RNA, which was reverse transcribed into cDNA when the transgenic plants showed strong growth. PCR cloning was performed with ORF-specific primers to verify the positive cloned plants.

LC-MS/MS Analysis of Tobacco
After thawing six tobacco samples at 4 • C, precooled methanol/acetonitrile (Merck, Darmstadt, Germany) /water solution (2:2:1, v/v) was added, followed by low-temperature ultrasonic extraction; the supernatant was then vacuum-dried. For mass spectrometry analysis, 100 µL of acetonitrile water solution was added (acetonitrile:water = 1:1, v/v) for reconstitution, and the supernatant was injected for analysis.
The samples were separated using an Agilent 1290 Infinity LC ultrahigh-performance liquid chromatography (UHPLC) (Agilent, CA, USA) HILIC column (Waters, MA, USA), and the samples were placed in an autosampler at 4 • C throughout the analysis. Continuous analysis of samples was performed in random order to avoid the effects of fluctuations in the instrument detection signal. An AB Triple TOF 6600 mass spectrometer (AB SCIEX, MA, USA) was used to collect the primary and secondary spectra of the samples.

Statistical Analyses
Data for blackberry are presented as the mean ± standard deviation (SD); data were obtained using the SPSS 24.0 statistical software program (SPSS 24.0, Chicago, IL, USA). One-way ANOVA was used to compare the mean values, and Duncan's post hoc multiple comparisons were used to calculate the significance of differences in flavonoid, anthocyanin, and total phenolic contents and RuFLS2 expression levels in different tissue parts (flowers, roots, stems, leaves, and fruiting branches) and fruits at different ripening periods (p < 0.05).
For the metabolome analysis of transgenic tobacco, unsupervised dimension reduction principal component analysis (PCA) was applied to all samples using the R package model (http://www.r-project.org/, accessed on 1 January 2022) to initially show the difference between groups of samples. Partial least squares discriminant analysis (PLS-DA) was applied to the control group using the R package ropls (http://www.r-project.org/, accessed on 1 January 2022) to better differentiate between the two groups by screening for variables related to the classification membership metabolomic signature [39]. Latent Structure Orthogonal Projection Discriminant Analysis (OPLS-DA) [40] was further used for cross-validation and permutation tests using the R package model (http://www.rproject.org/, accessed on 1 January 2022) in the control group. The validation of OPLS-DA models in metabolomics was performed with predictive power (Q 2 ) values. Tests were carried out randomly 200 times to generate a distribution of R 2 ' values and Q 2 ' values.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/ijms232214230/s1, Supplementary Table S1. Growth conditions for blackberry in a subtropical monsoon climatic zone; Supplementary Table S2. PCR primer information for the blackberry RuFLS2 gene and flavonoid biosynthesis pathway genes in WT and RuFLS2overexpressing tobacco.