Introduction

Flower color is one of the key traits considered for genetic studies since the rediscovery of Mendel’s laws. Flower coloration is induced by the deposition of flavonoids (including anthocyanins), carotenoids, and betalains (Mol et al. 1998; Grotewold 2006; Tanaka and Brugliera 2013). Although flower pigmentation has been extensively studied in maize (Zea mays), petunia (Petunia hybrida), and Arabidopsis (Lepiniec et al. 2006), the knowledge is relatively limited for soybean [Glycine max (L.) Merr.], because flower color is used only as marker in breeding programs (Koes et al. 2005). However, flower color and patterning have recently become a topic of interest due to their wide use in applied research studies (Mol et al. 1998).

A variety of flower colors has been identified in soybean. Among them, purple and white are the most common colors (Fig. 1). Early genetic studies showed that purple and white-flower colors are controlled by a single pair of genes in which purple is dominant over white (Takahashi and Fukuyama 1919; Woodworth 1923; Nagai 1926). Out of the 19,649 soybean accessions deposited in the United States Department of Agriculture-Germplasm Resources Information Network (USDA-GRIN) germplasm collection, 13,133 accessions (67 %) have purple flowers, 6344 accessions (32 %) have white flowers, and the remaining accessions (1 %) have different flower colors such as dark purple, purple blue, purple throat, magenta, pink, and near white (Table 1). In contrast, almost all the accessions of wild soybean [Glycine soja Seib. and Zucc.] have purple flowers (Chen and Nelson 2004). Flower pigmentation in soybean is mainly controlled by six independent loci, W1, W2, W3, W4, Wm, and Wp (reviewed by Palmer et al. 2004). In addition, factors, such as anthocyanin structure, pH of the vacuole (where the anthocyanins are localized) and co-pigmentation have been found to influence flower coloration (Mol et al. 1998; Yoshida et al. 2009).

Fig. 1
figure 1

Representation of the soybean genotype W1W1w3w3W4W4 (Harosoy) with purple flowers and the soybean genotype w1w1w3w3W4W4 (Williams 82) with white flowers. (A) Whole flower. (B) Standard petal

Full size image
Table 1 Distribution of values for flower color of soybean in USDA-GRIN (2015)
Full size table

Several research studies have been carried out in soybean to identify the molecular mechanism of flower pigmentation. However, flower pigmentation and regulation mechanisms remain unclear. This review summarizes the recent advancements in molecular and regulatory mechanisms that are responsible for the flower color development in soybean.

Anthocyanin and flavonol biosynthesis pathway and flower pigmentation

Anthocyanins are the most prominent pigments in soybean, which are responsible for purple, blue, and pink color in flowers. In addition, they protect plants from different stresses such as UV irradiation and reactive oxygen species (Nagata et al. 2003; Gould 2004). Three derivatives of anthocyanins, namely delphinidin, pelargonidin, and cyanidin, are the most commonly found in plants (Schwinn and Davies 2004). Delphinidin falls into the blue-purple range, pelargonidin into the pink range, and cyanidin into the red range (Harborne 1967). In soybean, most of the accessions have purple flowers, which show that delphinidin branch of anthocyanin biosynthesis is predominantly active in the determination of flower color (Iwashina et al. 2007).

Anthocyanins are water-soluble pigments and derived from a branch of the flavonoid pathway. In anthocyanin biosynthesis, the enzyme chalcone synthase (CHS) catalyzes the initial step by condensing one molecule of 4-coumaroyl-CoA and three molecules of malonyl-CoA for the synthesis of naringenin chalcones (Fig. 2). Naringenin chalcones act as the precursors for the synthesis of all the classes of flavonoids, isoflavonoids, and anthocyanins. The stereospecific isomerization of chalcone isomerase (CHI) converts naringenin chalcones to colorless naringenin, and then flavanone 3-hydroxylase (F3H) converts naringenin to dihydrokaempferol. Subsequently, the hydroxylation of dihydrokaempferol to dihydroquercetin and dihydromyricetin is catalyzed by flavonoid 3′-hydroxylase (F3′H) and flavonoid 3′5′-hydroxylase (F3′5′H), respectively. Dihydroflavonol 4-reductase (DFR) catalyzes the reduction of dihydroflavonols (dihydrokaempferol, dihydroquercetin, and dihydromyricetin) to colorless leucoanthocyanidins (Holton and Cornish 1995).

Fig. 2
figure 2

Schematic representation of anthocyanin and flavonol biosynthesis pathway involved in flower development. AMT anthocyanin methyltransferase; ANS anthocyanidin synthase; CHI chalcone isomerase; CHS chalcone synthase; DFR dihydroflavonol-4-reductase; F3H flavanone 3-hydroxylase; F3′H flavonoid 3′-hydroxylase; F3′5′H flavonoid 3′5′-hydroxylase; FLS flavonol synthase; UFGT UDP-flavonoid glucosyltransferase

Full size image

Furthermore, anthocyanidin synthase (ANS) oxidizes leucoanthocyanidins to anthocyanidins with the help of ferrous ion (Sparvoli et al. 1994; Gollop et al. 2001). In nature, anthocyanidins are unstable, but glycosylation increases their stability and hydrophilicity (He et al. 2010). Uridine diphosphate-flavonoid glucosyltransferase (UFGT) catalyzes the O-glycosylation of anthocyanidins and synthesizes more stable anthocyanins (delphinidin 3,5-di-O-glucoside, pelargonidin 3,5-di-O-glucoside, and cyanidin 3,5-di-O-glucoside) (Hostel 1981; Springob et al. 2003). Subsequent methylation of delphinidin 3,5-di-O-glucoside to petunidin 3,5-di-O-glucoside and malvidin 3,5-di-O-glucoside is catalyzed by anthocyanin methyltransferase (AMT) (Tanaka et al. 2005).

Similar to anthocyanins, flavonols and their derivatives are also synthesized from dihydroflavonols by flavonol synthase (FLS) followed by glycosylation, methylation, and acylation (To and Wang 2006). In purple flowers, a high amount of malvidin 3,5-di-O-glucoside and small traces of delphinidin 3,5-di-O-glucoside, petunidin 3,5-di-O-glucoside, and delphinidin 3-O-glucoside have been detected among anthocyanins. In addition, eight flavonol derivatives were also found, and kaempferol 3-O-gentiobioside was detected as the primary flavonol (Iwashina et al. 2007, 2008). In soybean, F3′H and F3′5′H are the two essential enzymes, which are mainly involved in the pigmentation of seed coat and flowers, respectively (Zabala and Vodkin 2003; Han et al. 2010; Moreau et al. 2012).

Among the six flower-color-controlling loci (W1, W2, W3, W4, Wm, and Wp) of soybean, all loci encode enzymes involved in flavonoid biosynthesis, except for W2. The loci W1, W3, W4, Wm, and Wp encode F3′5′H, DFR1, DFR2, FLS, and F3H1, respectively, whereas W2 encodes an MYB transcription factor, which is involved in the vacuolar acidification of flower petals (Takahashi et al. 2008, 2011).

The role of structural genes of anthocyanin biosynthesis pathway in flower pigmentation

F3′5′H gene at the W1 locus

In anthocyanin biosynthesis, F3′5′H (a cytochrome P450 enzyme) hydroxylates dihydrokaempferol to produce dihydromyricetin from which a delphinidin class of anthocyanins (purple color) is synthesized. Buzzell et al. (1987) found that W1 is responsible for the synthesis of delphinidin-3-glucosides and the production of purple flowers in soybean. W1 was mapped between Satt348 and Satt160 on MLG F (chromosome 13) in soybean linkage map (Song et al. 2004).

Zabala and Vodkin (2007a) identified the F3′5′H gene as a candidate gene for W1 and provided molecular evidence for the association with W1 locus. In Williams 82 (w1 allele), a small 65-bp insertion with tandem repeats and a 12-bp deletion in the third exon cause premature termination of translation and make recessive w1 allele non-functional and produced white flowers. Further analysis of F3′5′H showed very low expression levels in both purple- and white-flower lines, implying that the low amount of the F3′5′H enzyme is sufficient to synthesize delphinidins and determines the flower color in soybean (Zabala and Vodkin 2007a). However, anthocyanins were not detected in white flowers of the w1 mutant which strongly suggests that the W1 locus encodes the F3′5′H enzyme (Iwashina et al. 2007). In addition, Yang et al. (2010) developed an indel-based marker, SL019, from the third exon of the F3′5′H gene (where a 65-bp indel was present), which perfectly detects the polymorphism between the W1 and w1 alleles of the W1 locus from all the purple- and white-flower lines, respectively.

Recently, Park et al. (2014) analyzed the W1 locus from 99 landraces with white flowers and showed that all the white-flower landraces have the w1 allele, which is identical in sequence with the w1 allele of Williams 82 (studied by Zabala and Vodkin 2007a). In addition, a phylogenetic analysis showed the possible origin of the w1 recessive allele from a group of purple-flower G. max accessions, which were found to cluster together with white-flower G. max accessions in the constructed phylogenetic tree.

In contrast, almost all the accessions of wild soybean (G. soja) have purple flowers. The absence of flower color variations in wild soybean remains unclear. However, a few G. soja accessions have been identified with color variations. In 1998, a white-flower line, PI 424008C, was found among the progenies of the G. soja accession PI 424008A that has purple flowers. Genetic analysis showed that the white color in PI 424008C was caused by a mutation in the W1 locus similar to that in the white-flower G. max accessions (Chen and Nelson 2004). Takahashi et al. (2010) identified an accession of G. soja (B09121) with light purple flowers and assigned it as the w1-lp mutant. They also showed that the w1-lp allele is dominant over the w1 allele (W1 > w1-lp > w1). In the w1-lp mutants, cDNA sequence analysis of the F3′5′H gene indicated a unique single-base substitution in the nucleotide position 653, which results in a notable amino acid change from valine to methionine (position 210). However, not much difference was detected in the transcription level between the w1-lp and W1 alleles. Flavonoid analysis of the flower petals in the w1-lp line showed a very little amount of all the four major anthocyanins that are common in purple flowers (malvidin 3,5-di-O-glucoside, petunidin 3,5-di-O-glucoside, delphinidin 3,5-di-O-glucoside, and delphinidin 3-O-glucoside) and also small traces of 5′-unsubstituted derivatives of these anthocyanins, which suggests that a mutation in the F3′5′H gene may lead to a reduction of the F3′5′H enzymatic activity and an increase of the F3′H enzymatic activity in the w1-lp mutant line (Takahashi et al. 2010).

Recently, Takahashi et al. (2012) found a single plant with purple and white variegated flowers in a G. soja accession (B00146). This single plant with variegated flowers (B00146-m) developed progenies with white flowers (B00146-w) as well as with purple flowers (B00146-r). The w1-m allele of variegated flowers is allelic to the W1 locus, and sequence analysis of the F3′5′H gene in B00146-m showed the insertion of the Tgs1 transposon (CACTA family) in the first exon. Excision of this active transposon led to the development of revertant (purple) lines and mutant (white) lines. In B00146-w, CA insertion in the F3′5′H gene as a footprint of Tgs1 at the transposon insertion site led to a truncated polypeptide with 59 amino acids, which might cause the complete loss of its function. As in G. max, G. soja also displayed white flower when W1 locus was dysfunctional, which revealed that W1 is the most important locus for flower color production in soybean. In addition, W1 locus has a pleiotropic effect on flower and hypocotyl colors. Cultivars with the dominant W1 allele produce purple flowers and purple hypocotyls, whereas cultivars with the recessive w1 allele produce white flowers and green hypocotyls (Takahashi and Fukuyama 1919).

DFR genes at the W3 and W4 loci

The DFR enzyme is known to catalyze the production of leucoanthocyanidins from dihydroflavonols. DFR is an essential enzyme, to step forward, the synthesis of anthocyanins rather than flavonols in the biosynthetic pathway. In soybean, the genes encoding DFR enzymes co-segregate with the two loci, W3 and W4, and they function (or interact) epistatically to each other in W1 background (Fasoula et al. 1995; Xu and Palmer 2005). Soybean accessions show different flower colors depending on the allelic variations of W3 and W4. For example, the genotype W1W3W4 produces dark purple flowers, W1w3w4 genotype produces near white flowers, W1w3W4 genotype produces purple flowers, and W1W3w4 produces purple throat flowers (Hartwig and Hinson 1962; Yan et al. 2014).

To identify the gene that is related to the W3 locus, Fasoula et al. (1995) carried out a restriction fragment length polymorphism analysis with 23 restriction enzymes using a DFR probe, which contained a complete partial sequence of DFR1 (Wang et al. 1994). Out of them, the enzyme HaeIII showed polymorphism in the DFR1 gene between the purple throat (W1W3w4) and the near white (W1w3w4) flowers. Furthermore, recombination analysis of a population developed by a cross between a purple-throat-flower line (W3) and a near-white-flower line (w3) showed a complete co-segregation and the absence of recombination between the W3 and w3 alleles. Fasoula et al. (1995) suggested that the phenotype of purple throat flowers in soybean is possibly due to the undetermined excision of some transposable elements, which cause changes in the expression pattern of the DFR1 gene.

A recent sequence homology search and microsatellite marker analysis also showed that DFR1 is between Sat_287 and Satt467 on MLG B2 (chromosome 14) indicating that DFR1, among the homologous DFR genes, is most likely associated with the W3 locus (Yang et al. 2010). However, the W3 locus has not been fully characterized yet.

The W4 locus is also known to encode DFR in the anthocyanin biosynthesis pathway and was mapped between Satt386 and Sct_137 on MLG D2 (chromosome 17). The DFR2 gene was identified as a candidate gene for the W4 locus (Xu and Palmer 2005). Mutations in the W4 locus with W1w3 background results in different types of pigment accumulation and color patterning in flower petals. Five mutant alleles, w4, w4-m, w4-dp, w4-p, and w4-lp, have been identified in the W4 locus that produce near white, variegated, dilute purple, pale purple, and light purple flower color, respectively (Palmer et al. 1989; Groose and Palmer 1991; Xu and Palmer 2005; Xu et al. 2010; Yan et al. 2014).

The complete loss of function of the W4 locus (w4 recessive allele) in different lines (Clark-w4, 222-A-3, and kw4) develops only near white flowers, irrespective of the mutation type (Yan et al. 2014). Single-base substitution at the 5′ splice site of the fourth intron of DFR2 in Clark-w4 led to the retention of the fourth intron in the transcript. A single-base deletion in 222-A-3 resulted in a truncated polypeptide of only 24 amino acids that lacked the NADPH binding domain, which is important for the DFR2 function. In the wild soybean mutant kw4, no expression of DFR2 in flower petals resulted in substantial reduction of anthocyanins, possibly due to the deletion of a 367-bp fragment in the third intron of DFR2 or a nucleotide polymorphism in the promoter region (Yan et al. 2014). Further experiments are required to clarify this mechanism.

Analysis of the DFR2 gene from another mutant w4-m showed the insertion of an active transposon (Tgm9) in the second intron (Xu et al. 2010). Excision of this active transposon in some somatic cells during flower development resulted in the production of variegated flowers, and this kind of transposon excision also caused stable revertants. Two stable revertants, w4-dp (W1w3w4-dp) and w4-p (W1w3w4-p), were developed from the w4-m line, and they produced dilute purple and pale purple flowers, respectively (Palmer and Groose 1993; Xu and Palmer 2005). In the w4-dp and w4-p mutants, excision of the Tgm9 transposon leaves 4- and 0-bp footprints, respectively. Moreover, the excised fragment of Tgm9 incorporates into the promoter region at 1043 and 1034 bp of the upstream transcription start site in the w4-dp and w4-p alleles, respectively (Xu et al. 2010). Expression analysis showed a very low expression level of DFR2 in the w4 and w4-dp mutants, but a higher expression level in the w4-p mutant than in the W4 dominant line. Flower petals also showed the presence of unique dihydroflavonols in the w4-p mutant (Yan et al. 2014).

Recently, analysis of an EMS-induced mutant (W1w3w4-lp) with light purple flowers generated from a purple-flower soybean line showed that the w4-lp was responsible for the light purple flowers (Yan et al. 2014). In the w4-lp mutant, a single-base substitution in DFR2 resulted in an amino acid change from arginine to histidine (position 39) and led to low expression levels of DFR2 (Yan et al. 2014). Further experiments may be required to clarify the reasons of low transcript abundance and study the functional activity of DFR2 in the w4-lp mutant.

It is clear that mutations of DFR2 with w3 background leads to unique flavonoid compositions and different shades of purple, such as dilute purple, pale purple, and light purple, depending on the expression level of DFR2 or the activity of the enzyme. However, a comprehensive study of the promoter is necessary to identify the cis-regulated expression of DFR2.

F3H gene at the Wp locus

The F3H enzymes catalyze an early and very important step in anthocyanin metabolism, in which flavonones are converted to dihydroflavonol (dihydrokaempferol). Genetic analysis showed that the Wp locus corresponds to F3H, and the homozygous recessive wp allele in W1 background develops pink flowers in soybean (Stephens and Nickell 1992). The Wp locus was mapped between SL007 and Satt216 on MLG D1b (chromosome 2) (Yang et al. 2010), and the F3H1 gene was identified as a candidate gene for the Wp locus (Zabala and Vodkin 2005). A mutant (wp allele) with pink flowers was derived from a mutable line with a high rate of instability in flower color (variegated flowers) (Johnson et al. 1998; Zabala and Vodkin 2005).

Analysis of the F3H1 gene from the wp line showed the insertion of an active transposon (Tgm-Express1) in the second intron, and expression analysis showed very low expression levels in the wp mutants compared to the Wp dominant lines. This active transposon also created a variety of chimerical transcripts with varying open-reading frames in the wp mutants. About twelve distinct chimeric transcripts and one non-chimeric transcript were identified in the wp mutants by Zabala and Vodkin (2007b). The non-chimeric transcript of the wp mutant was identical to the normal transcript of the Wp allele, even though the F3H1 transcripts with typical sizes were found in very low amounts in the wp mutant, suggesting that they are still sufficient for synthesizing anthocyanin pigments to develop pink flowers (Zabala and Vodkin 2007b).

Other anthocyanin biosynthesis pathway enzymes

The enzyme CHS is the first key enzyme in anthocyanin biosynthesis pathway, which catalyzes the production of chalcone that acts as the precursor of flavonoid biosynthesis. In petunia and tobacco, the CHS genes have been isolated and characterized for flower color and pigmentation (Koes et al. 1986; Wang et al. 2006). In soybean, the CHS repeats and I locus were mapped between A454.p2 and GMNOD2B on MLG A2 (chromosome 8) (Yang et al. 2010). The CHS gene family has been studied extensively for seed coat color (Tuteja et al. 2004, 2009).

The F3′H enzyme is involved in the conversion of dihydrokaempferol to dihydroquercetin. Genetic analysis showed that the F3′H gene was identified as the T locus (Toda et al. 2002; Zabala and Vodkin 2003). The T locus was mapped between Satt286 and Satt365 on MLG C2 (chromosome 6) (Yang et al. 2010) and controls seed coat/hilum color in soybean (Toda et al. 2002).

The enzyme ANS is essential for the oxidation of leucoanthocyanidins to anthocyanins, and there is a high sequence similarity between the genes encoding these enzymes. Among the ANS genes (ANS1, ANS2, and ANS3) in soybean, ANS2 and ANS3 have identical sequences, and they are likely from the same gene. ANS1 and ANS2 were mapped between BE806308 and Sat_272 on MLG B1 (chromosome 11) and between Sat_414 and Satt129 on MLG D1a (chromosome 1), respectively (Yang et al. 2010).

UFGT catalyzes the glycosylation of anthocyanin to colored anthocyanin 3-O-glucosides by transferring glucose moiety from uridine diphosphate-glucose to C-3 hydroxyl group of anthocyanin (Buzzell et al. 1987; Todd and Vodkin 1993). The putative genes UGT78K1 and UGT78K2 were identified by homology search using BLASTn and characterized as the UFGT genes that control seed coat pigment (Kovinich et al. 2010, 2011). However, UFGT has not been precisely mapped in soybean linkage map.

In addition, the transcript level of the UFGT and ANS genes were regulated by the R locus, which encodes the R2R3 MYB transcription factor. The R locus was mapped between BARCSOYSSR_09_1489 and BARCSOYSSR_09_1506 on MLG K (chromosome 9) (Gillman et al. 2011). When the R2R3 MYB gene is overexpressed, the UFGT (UGT78K1 and UGT78K2) and ANS (ANS2/ANS3) genes are also up-regulated and produce black seed coat/hilum (Gillman et al. 2011; Kovinich et al. 2011). In contrast, loss of function of the R2R3 MYB gene causes a down-regulation in both the UFGT and ANS genes and produces brown seed coat/hilum. It remains to be answered whether the MYB transcriptional factor directly promotes the transcriptional activation of the UFGT and ANS genes.

In soybean, all four enzymes discussed in this review have been studied extensively for their involvement in seed coat/hilum color, but not in flower pigmentation. However, genetic regulation of these enzymes with flower color are well established in other plant species, such as in petunia and Arabidopsis (To and Wang 2006); hence it is likely that these enzymes play an important role in soybean flower pigmentation too. These enzymes and their corresponding and regulating loci need to be addressed for flower color in soybean.

Additional factors affecting flower pigmentation

Co-pigmentation by the FLS gene at the Wm locus

Flavonols (myricetin, kaempferol, and quercetin) are yellow-color components, and they act as either pigments or co-pigments to anthocyanins (Harborne 1967); however, in soybean, flavonols act only as co-pigments (Takahashi et al. 2007). In flavonoid biosynthesis, both anthocyanin and flavonols are derived from dihydroflavonols by the establishment of a double bond between C-2 and C-3 positions by the action of FLS (Forkmann 1991). Analysis of a mutant with magenta flowers from Harosoy (Harosoy-wm) showed that the Wm locus controls magenta color (Buzzell et al. 1977). The recessive wm allele at the Wm locus was found to be associated with low levels of flavonol glucoside synthesis in flowers and leaves, which suggested that Wm is responsible for the production of flavonol and probably encodes FLS. The Wm locus was mapped between Satt252 and Satt425 on MLG F (chromosome 13) (Takahashi et al. 2007).

Takahashi et al. (2007) identified a candidate gene, gmfls1, in the Wm locus, which encodes a 334-amino-acid-long polypeptide, GmFLS1, consisting of conserved dioxygenase domains (A and B), and this gene showed homology with previously reported FLS genes from other plant species. Bacterial heterologous expression assay showed that GmFLS1 of Harosoy (Wm allele) has the activity of FLS, whereas a single-base deletion in the wm allele resulted in a truncated polypeptide and devoid of FLS activity. Thus, co-pigmentation between anthocyanin and flavonol glucosides may contribute to purple flowers in soybeans along with the Wm alleles, whereas the recessive wm allele substantially reduces the content of flavonol glucosides, and it may inhibit co-pigmentation, resulting in magenta flowers (Takahashi et al. 2007).

MYB transcription factor gene at the W2 locus and regulation of vacuolar pH

MYB transcription factors represent a group of proteins that consist of a conserved MYB DNA-binding domain. In plants, an MYB-protein subfamily is illustrated as the R2R3-type MYB domain and the R2R3-type MYB genes control many paths of plant secondary metabolism (Stracke et al. 2001).

The purple-blue-flower landrace, Nezumisaya, was identified, and a complementation test showed that purple blue color was controlled by the W2 locus, which was mapped between Satt318 and Satt020 on MLG B2 (chromosome 14) (Takahashi et al. 2008). Flavonoid analysis of flower petals showed that the alleles of W1, W3, W4, Wm, and Wp loci affect the structure and/or amount of flavonoids (Iwashina et al. 2007, 2008; Takahashi et al. 2010). In contrast, flavonoids in the purple blue flowers were similar to that of purple flowers, which suggest that quantitative or structural differences of anthocyanins or co-pigmentation were not responsible for the purple blue flowers (Iwashina et al. 2008).

To identify the factor affecting the color change in the purple blue flowers, Takahashi et al. (2008) investigated the physiological basis of flower color. An increase in pH has a bluing effect in flower petals, while a decrease in pH causes a reddening effect (To and Wang 2006). In accordance, the sap extract of standard petal from purple flowers showed pH values from 5.73 to 5.77, whereas purple blue flowers had pH values from 6.07 to 6.10, which suggests that the recessive allele w2 may be responsible for the acidification of flower petals and produces purple blue flowers (Takahashi et al. 2008). Similarly, in petunia, when mutation occurs in the PH genes (PH1 to PH7), plants produce blue flowers and show increased pH in petal sap extract (deVlaming et al. 1983; Chuck et al. 1993).

To ascertain how the transcription factor controls the vacuolar pH of flower petals, Takahashi et al. (2011) identified a candidate gene, GmMYB-G20-1, which encodes an MYB transcription factor. This gene has a 53.7 % amino-acid sequence similarity with the PH4 gene of petunia, which controls the vacuolar pH and blue color. GmMYB-G20-1 of purple flower (W2 allele) encodes a 361-amino-acid-long polypeptide, whereas GmMYB-G20-1 of purple blue flower mutants (w2 allele) has a nonsense mutation in the MYB domain, resulting in a truncated polypeptide (Takahashi et al. 2011). Furthermore, virus-induced gene silencing in Harosoy (purple flowers) revealed that the silencing of GmMYB-G20-1 changes the flower color from purple to gray or blue. These results showed that GmMYB-G20-1 corresponds to the W2 locus and controls the proteins involved in pH regulation of petals as a transcription factor (Takahashi et al. 2013).

Conclusion

Flower color and pigmentation have been recognized as a potential tool for elucidating the basis of genetics and biochemistry. Flower color is developed by the accumulation of flavonoids (including anthocyanins), carotenoids, and betalains. Anthocyanin pigments act as major components in flower color development. The structural genes, which control anthocyanin and flavonol biosynthesis pathway, have been studied to some extent. Mutations in these genes create variations in soybean flower color. In addition, other factors, such as environment, regulatory genes, co-pigments, and vacuolar pH also play an important role in flower coloration. However, the factors influencing biosynthesis and regulation of anthocyanins are still unclear in soybean. Therefore, detailed studies are necessary to understand the whole mechanism of flower coloration. Furthermore, the available knowledge of active transposable elements, such as Tgs1, Tgm9, and Tgm-Express1, may provide a path for the utilization of these active transposons as tagging tools in soybean.