Introduction

Flavonoids are important secondary metabolites in the plant kingdom. Individual plant species synthesize a variety of flavonoid compounds that function in providing flower pigmentation to attract pollinators, in defending plants against pathogens, in acting as signal molecules in plant–microbe interactions, and in protecting plants from UV radiation (Shirley 1996, 2001; Dixon and Steele 1999).

The biosynthetic pathway of flavonoids is well established and many of the structural and some of the regulatory genes have been cloned in several model plants including maize, petunia and snapdragon (Holton and Cornish 1995). Anthocyanins and flavonols are both derived from dihydroflavonols (dihydrokaempferol, dihydroquercetin, and dihydromyricetin). The B-ring of dihydrokaempferol can be hydroxylated at the 3′ position by flavonoid 3′-hydroxylase (F3′H), or at the 3′ and 5′ positions by flavonoid 3′5′-hydroxylase (F3′5′H) to produce dihydroquercetin and dihydromyricetin, respectively. Flavonols are formed from dihydroflavonols, as illustrated in Fig. 1, by the introduction of a double bond between C-2 and C-3 through the action of flavonol synthase (FLS) (Forkmann 1991).

Fig. 1
figure 1

Schematic diagram of flavonol biosynthesis from dihydroflavonols. Enzymes involved in each step are indicated as follows: F3′H flavonoid 3′-hydroxylase, F3′5′H flavonoid 3′5′-hydroxylase and FLS flavonol synthase

Five genes (T, W1, I, R and O) control the color of the seed coat, five genes (W1, W3, W4, Wm and Wp) control flower color, and two genes (T and Td) control the color of the pubescence in soybean (reviewed by Palmer et al. 2004). RFLP marker of dihydroflavonol reductase co-segregated with flower color variants in an F2 population segregating for W3 locus (Fasoula et al. 1995). Genotypes at I locus were associated with chalcone synthase gene duplications (Todd and Vodkin 1996). Accordingly, W3 and I are presumed to encode dihydroflavonol reductase and chalcone synthase, respectively. Chromatographic experiments indicated that alleles at T and W1 loci are associated with 3′- and 3′5′-B-ring hydroxylation, respectively. The results suggested that T and W1 encode F3′H and F3′5′H, respectively (Buttery and Buzzell 1973; Buzzell and Buttery 1982). Toda et al. (2002) cloned and characterized the F3′H cDNA from a pair of near-isogenic lines for the T locus, To7B (TT, tawny pubescence) and To7G (tt, gray pubescence). Sequence analysis revealed that they differed by a single-base deletion of C in the coding region of To7G. The deletion generated a truncated polypeptide lacking the GGEK consensus sequence of F3′H gene and the heme-binding domain resulting in non-functional protein. Further, cDNA microarray suggested that Wp corresponds to flavanone 3-hydroxylase gene (Zabala and Vodkin 2005).

Magenta flower mutant of Harosoy (Harosoy-wm, T235) was found at Urbana, IL in 1957 (Fig. 2, Buzzell and Buttery 1977). They found that a single-gene controlled the phenotype and designated the gene as Wm. They found a linkage between W1 and Wm with a recombination value of 2.2 ± 0.7%. The two loci were assigned to the middle region of molecular linkage group F (Cregan et al. 1999). Buzzell and Buttery (1977) further found that the recessive allele wm is associated with low levels of flavonol glycosides in the leaves and flowers. It suggests that Wm is responsible for the production of flavonol and may encode either FLS or a regulatory gene controlling expression of FLS gene. This study was conducted to isolate and characterize the FLS gene of soybean, and to investigate the relationship between FLS gene and the Wm locus.

Fig. 2
figure 2

Flowers of Harosoy (purple, genotype WmWm) and Harosoy-wm (magenta, genotype wmwm)

Materials and methods

Plant materials

Purple flower cultivars, Clark and Harosoy (WmWm), and Harosoy-wm (wmwm) were used. Seeds of Harosoy-wm were provided by the USDA Soybean Germplasm Collections. Plants were grown in fields supplemented with a synthetic fertilizer (3.0 N–10.0P2O5–10.0 K2O kg/10a) at National Institute of Crop Science, Tsukuba, Japan. For genetic analysis, Harosoy-wm was crossed with Clark in 2004. Hybridity of F1 plants was ascertained by purple flower color. One hundred and fourteen plants of the F2 population were grown in fields in 2005.

RNA extraction

Total RNA was extracted from frozen trifoliolate leaves (5 g) sampled at R3 stage (Fehr et al. 1971) and from frozen banner flower petals (500 mg) of Clark, Harosoy and Harosoy-wm using the TRIZOL Reagent (Invitrogen) according to the manufacturer’s instructions.

cDNA cloning

Three reverse primers, gmfls-GSP1 (5′-CACTCCTATTAGGTGCTTGCA-3′), gmfls-GSP2 (5′-AGGCCACACAATGTGGAACAA-3′) and gmfls-GSP3 (5′-ATGATCCACCCACCCTTTCTT-3′) were constructed for 5′ RACE. 2.5 μg of the total RNA from flower petals of Harosoy was reverse transcribed using gmfls-GSP1 as a primer. After purification and terminal transferase reaction, the resulting TdT product was used as a template in first-round PCR in the presence of the Abridged Anchor Primer (5′ RACE System ver. 2.0, Gibco-BRL) and gmfls-GSP2. A 2 min denaturation at 94°C was followed by 35 cycles of 30 s denaturation at 94°C, 30 s annealing at 55°C, and 1 min extension at 72°C. A final 5 min extension at 72°C completed the program. The PCR was performed in an Applied Biosystems 9700 thermal cycler. For the second round PCR, 5 μl of the diluted first-round PCR product was reamplified with AUAP and gmfls-GSP3. The PCR product with molecular size of about 500 bp was cloned into pCR 2.1 vector (Invitrogen) and sequenced. A putative full-length clone of about 1,200 bp was obtained from Harosoy and Harosoy-wm by PCR using specific primers at the extreme 5′ end (5′-GAAATCAACACACACAAACACAACAAC-3′) and 3′ end (5′-CTACTTCCCACAAGTCACACTTTC-3′), cloned in pCR 2.1, and sequenced.

DNA sequencing

Nucleotide sequences of both strands were determined with BigDye terminator cycle method using an ABI3100 Genetic Analyzer (Applied Biosystems). Nucleotide and amino acid sequences were analyzed with the BLAST program (Altschul et al. 1997).

Expression in E. coli cells and enzyme assay

The entire coding region of gmfls1 was cloned from Harosoy and Harosoy-wm by RT-PCR using the above 5′ primer and a reverse primer (GTGCCCAGAATTCTCATTGAGGGATCTTATTG) that included an additional EcoRI site. The amplified bands were digested at the intrinsic NdeI and the introduced EcoRI sites, and were inserted into a pET21a vector (Novagen). E. coli RosettaTM2 (DE3) was transformed with the vectors harboring gmfls1 or an empty vector. Expression of FLS proteins was induced by adding 1 mM isopropyl-thio-b-d-galactoside to 50 ml of overnight culture for 3 h at 28°C. Cells were harvested by centrifugation, washed with 0.1 M potassium-phosphate (pH 7.0) containing 10% sucrose and 1.4 mM 2-mercaptoethanol and collected by centrifugation.

The cell pellets were suspended in 10 ml of 0.1 M potassium phosphate (pH 7.0) containing 10% sucrose and 1.4 mM 2-mercaptoethanol, disrupted by sonication and then centrifuged. The supernatant was used as a crude enzyme extract for the following FLS assay. A 500 μl reaction mixture, consisting 0.1 M potassium-phosphate (pH 7.0), 10% sucrose, 1.4 mM 2-mercaptoethanol, 250 μM 2-oxoglutarate, 5 mM ascorbic acid, 50 μM ferrous sulfate, 10 μg (±)-taxifolin (dihydroquercetin; Sigma) and 50 μl of crude enzyme extract, was incubated for 1 h at 28°C. The reaction was terminated by extraction with 500 μl of ethyl acetate, and the extract was analyzed by HPLC-photodiode array spectrophotometry. HPLC was performed using a TSK-GEL ODS 80TM column (4.6 × 150 mm; TOSOH). Mobile phase solution A was MeOH and solution B consisted of distilled water containing 0.1% phosphoric acid. A gradient program was used for separation at the flow rate of 1 ml/min at 40°C. The gradient program was as follows: 0–15 min, 40–70% (v/v) A; 15–16 min, 70–100% (v/v) A; 16–17 min, 100% (v/v) A; 17–18 min, 100–40% (v/v) A; 18–24 min, 40% (v/v) A. The chromatograms were recorded at the maximum absorbance in the range of 230–400 nm. Quercetin was identified based on the retention time and its UV-Vis spectrum was compared with the standard samples.

DNA blot analysis

Genomic DNA of Clark, Harosoy and Harosoy-wm was isolated from trifoliolate leaves by the CTAB method (Murray and Thompson 1980), and digested with three restriction enzymes, EcoRI, HindIII and BglII. The digested DNA of 5 μg per lane was separated by electrophoresis, and blotted on to nylon membranes (Hybond-N+, Amersham Pharmacia Biotech). Southern hybridization was performed with gmfls1 probes labeled by the ECL direct labeling system (Amersham Pharmacia Biotech), and hybridization signals were detected by chemiluminescence on X-ray films.

CAPS analysis

A pair of PCR primers (5′-GAAATCAACACACACAAACACAACAAC, 5′-CAAAGAACATTTTCCCAACACTTTGCA) was designed to amplify a band including the single-base deletion of G found in Harosoy-wm. The deletion was expected to generate a BglII site (AGATCT) in the amplified band of Harosoy-wm. The PCR mixture contained 30 ng of genomic DNA, 5 pmol of each primer, 10 pmol of nucleotides, and 1 unit of ExTaq in 1× ExTaq Buffer supplied by the manufacturer (TAKARA BIO) in a total volume of 50 μl. After an initial 30 s denaturation at 94°C, there were 30 cycles of 30 s denaturation at 94°C, 1 min annealing at 57°C, and 1 min extension at 72°C. A final 7 min extension at 72°C completed the program. The amplified products were digested with BglII and the digests were separated on 8% polyacrylamide gel in 1× TBE buffer (90 mM Tris–borate, 2 mM EDTA, pH 8.0). After electrophoresis, the gel was stained with ethidium bromide and the DNA fragments were visualized under UV light.

SSR analysis and linkage mapping

The PCR mixture for SSR analysis contained 1 μl of the extracted DNA that was diluted by ten times in TE-buffer, 5 pmol of primers, 10 pmol of nucleotides, and 1 unit of ExTaq in 1× ExTaq Buffer in a total volume of 20 μl. A 12 min denaturation at 94°C was followed by 40 cycles of 30 s denaturation at 94°C, 30 s annealing at 50°C, and 30 s extension at 68°C. A final 5 min extension at 68°C completed the program. The PCR products were separated on 8% acrylamide gels in 1× TBE buffer, and the fragments were visualized by ethidium bromide staining. A linkage map was constructed using MAPMAKER/ EXP. ver. 3.0 (Lander et al. 1987).

Semi-quantitative RT-PCR analysis

Semi-quantitative RT-PCR was conducted by reverse-transcribing 5 μg of total RNA using the Superscript III First-Strand Synthesis System (Invitrogen) and oligo d(T) primer according to the manufacturer’s instruction. To test the transcription level of gmfls1, PCR reactions were carried out for 15 or 22 cycles of 94°C 1 min/57°C 1 min/72°C 30 s in a volume of 20 μl, using 1 μl of cDNA. The primers were 5′-ATGGAGGTGCTAAGGGTGCAAA-3′ (forward) and 5′-ATGATCCACCCACCCTTTCTT-3′ (reverse). SAc1, a soybean actin gene (Shah et al. 1983), was used as a control. PCR reactions were carried out for 30 cycles of 94°C 1 min/57°C 1 min/72°C 30 s in a volume of 20 μl, using 1 μl of cDNA. The primers were 5′-CTGGGGATGGTGTCAGCCACAC-3′ (forward) and 5′-CACCGAACTTTCTCTCGGAAGGTG-3′ (reverse). PCR products were loaded on a 2% agarose gel, stained by ethidium bromide and visualized under UV light.

Analysis of flavonoids in flower petals

Two hundred milligrams of banner petals were collected from field-grown Harosoy and Harosoy-wm in 2 ml of MeOH containing 0.1% HCl (for anthocyanin) or absolute MeOH (for flavonol) in three replications in 2003 and 2004. Crude extracts of flower petals were analyzed with Shimadzu HPLC systems using Shim-pack CLC-ODS [I.D. 6.0 mm × 150 mm (Shimadzu)] at a flow rate of 1.0 ml/min. Detection wavelength was 190–700 nm and eluents were MeCN/HOAc/H2O/H3BO3 (6:8:83:3) for anthocyanins and MeCN/H2O/H3BO3 (22:78:0.2) for flavonols. Amount of flavonoids was evaluated by their peak areas in HPLC chromatogram.

Results

Isolation of a flavonol synthase cDNA clone from soybean

A soybean EST clone that is homologous to FLS gene (GenBank accession number: BE059354) was purchased from the American Type Culture Collection (ATCC), Manassas, VA. Sequence analysis revealed that the clone was a chimera of ubiquitin and a truncated FLS gene lacking the 5′-region. 5′ RACE was performed to obtain the entire cDNA. A DNA fragment with approximate size of 500 bp was cloned and sequenced. A putative full-length cDNA of about 1.2 kb, gmfls1 was obtained by PCR using primers specific to the extreme 5′ and 3′ ends. Sequence analysis revealed that it contained 1,208 nucleotides (Fig. 3). The polypeptide (GmFLS1) of 334 amino acids was deduced from the DNA sequence based on an open reading frame that stretched from the putative 5′ end of the cDNA to nucleotide position 1,031. The apparent molecular mass and isoelectric point of GmFLS1 were 38,150 and 5.78, respectively. GmFLS1 had amino acid homology with the previously reported FLS from Eustoma grandiflorum (EgFLS, AF240764, 71.6%), Malus domestica (MdFLS, AY965343, 71.5%), Rosa hybrid (RhFLS, AB038247, 71.0%), Petunia hybrida (PhFLS, Z22543, 69.6%), Solanum tuberosum (StFLS, X92178, 67.2%), and Arabidopsis thaliana (AtFLS, U84259, 59.0%) (Fig. 4). Conserved dioxygenase domains A and B (Holton et al. 1993) were found in the deduced polypeptide.

Fig. 3
figure 3

(A) The cDNA and deduced amino acid sequence of gmfls1 (DDBJ accession number: AB246668) of Harosoy with purple flower (WmWm). The entire sequence shown is 1,208 bases long and contains a putative open reading frame of 334 amino acids shown below the nucleotide sequence. The conserved dioxygenase domains A and B are underlined. A guanine residue at the nucleotide position 75 (shown in bold) was deleted in Harosoy-wm (wmwm). (B) Comparison of cDNA and deduced amino acid sequences of gmfls1 from Harosoy and Harosoy-wm. Deletion of a guanine residue (shown in bold) altered the subsequent reading frame, changed the following 22 amino acids, and prematurely terminated the transcription of Harosoy-wm. The conserved dioxygenase domains A and B are absent in Harosoy-wm. The deletion generated a BglII site in Harosoy-wm (underlined)

Fig. 4
figure 4

Amino acid sequence alignment for FLS cDNAs from Rosa hybrid (RhFLS, AB038247), Malus domestica (MdFLS, AY965343), Glycine max (GmFLS, AB246668), Petunia hybrida (PhFLS, Z22543), Solanum tuberosum (StFLS, X92178), Eustoma grandiflorum (EgFLS, AF240764), and Arabidopsis thaliana (AtFLS, U84259). Amino acids common to all the FLS genes are in reverse type

A deletion mutation occurs in the mutant with a recessive wm allele

Sequence analysis of FLS gene from Harosoy and Harosoy-wm revealed that a single-base deletion of G occurs at the nucleotide position 75 in Harosoy-wm (Fig. 3). Amino acids were identical between Harosoy and Harosoy-wm until amino acid number 15. However, the deletion changed the subsequent reading frame in Harosoy-wm, altered the following 22 amino acids, and the polypeptide was terminated thereafter. The truncated polypeptide consisted of only 37 amino acids and lacks the dioxygenase domains A and B.

FLS activity

The HPLC chromatogram revealed that extracts of E. coli cells expressing Harosoy gmfls1 catalyzed dihydroquercetin into quercetin, indicating that gmfls1 encodes FLS (Fig. 5). In contrast, trace amounts of quercetin was detected in extracts of cells transformed with gmfls1 of Harosoy-wm or empty vector, and in the substrate (dihydroquercetin), indicating that GmFLS1 of Harosoy-wm has no FLS activity.

Fig. 5
figure 5

HPLC chromatogram of the FLS enzyme assay of E. coli cell extracts expressing FLS cDNAs (gmfls1) using dihydroquercetin as a substrate. FLS cDNA of Harosoy or Harosoy-wm was cloned into pET21a and transformed. E. coli cell extracts transformed with empty vector were also subjected to FLS assay as a negative control. The cell extracts expressing gmfls1 of Harosoy catalyzed the formation of quercetin from dihydroquercetin. In contrast, trace amount of quercetin was detected in extracts of cells transformed with gmfls1 of Harosoy-wm or empty vector, and in the substrate

DNA blot analysis

Genomic DNA blots of Clark, Harosoy and Harosoy-wm were analyzed using the entire region of gmfls1 as a probe to estimate the copy number. Figure 6 shows the hybridization patterns after digestion with EcoRI, HindIII, and BglII. Three to four bands with various hybridizing intensities were observed. Polymorphism was observed between Harosoy and Harosoy-wm after BglII digestion probably due to the single-base deletion of G in Harosoy-wm.

Fig. 6
figure 6

Restriction fragment length polymorphisms among Clark, Harosoy and Harosoy-wm using gmfls1 cDNA as probe. Genomic DNA was digested with restriction enzymes, EcoRI, HindIII and BglII, separated by electrophoresis of a 0.8% agarose gel, transferred to nylon membrane, and hybridized with gmfls1. Lanes 1 and 11 molecular marker λ/HindIII, 2 Clark (EcoRI digest), 3 Harosoy (EcoRI digest), 4 Harosoy-wm (EcoRI digest), 5 Clark (HindIII digest), 6 Harosoy (HindIII digest), 7 Harosoy-wm (HindIII digest), 8 Clark (BglII digest), 9 Harosoy (BglII digest), and 10 Harosoy-wm (BglII digest). The migration of size markers is shown to the right of the blots

CAPS analysis

CAPS analysis was conducted to detect the single-base deletion of G found in Harosoy-wm (Fig. 7). The deletion introduced BglII site in Harosoy-wm resulting in the two fragments (234 and 71 bp) after BglII digestion. In contrast, the amplified band of Harosoy was not digested (306 bp) because of the lack of the restriction site (Fig. 3).

Fig. 7
figure 7

Results of CAPS analysis of GmFLS in Clark, Harosoy and Harosoy-wm. PCR products were digested with BglII and separated on a 8% polyacrylamide gel. φ molecular marker φx174/HaeIII, Lanes 1 Clark (undigested), 2 Harosoy (undigested), 3 Harosoy-wm (undigested), 4 Clark (digested), 5 Harosoy (digested), and 6 Harosoy-wm (digested). The single-base deletion of G in Harosoy-wm introduced BglII site resulting in the two fragments (234 and 71 bp) after BglII digestion. In contrast, amplified bands of Clark or Harosoy were not digested (306 bp) because of the lack of the restriction site. The migration of size markers is shown to the left of the gel

Co-segregation of deletion mutation with flower color

Flower color of the F2 population segregated into 90 purple and 24 magenta flower plants. It fitted to a 3:1 ratio (χ 2 = 0.95, 0.3 < P < 0.5) similar to the report of Buzzell and Buttery (1977). To test whether the deletion mutation co-segregated with flower color, genomic DNA from randomly selected 89 F2 plants was subjected to CAPS analysis. Magenta flower plants had the 234 and 71 bp bands, whereas purple flower plants had fragment of either only the 306 bp band or all of the three bands (Fig. 8 and Table 1). The former and latter were presumed to be homozygotes (WmWm) and heterozygotes (Wmwm) at the Wm locus, respectively.

Fig. 8
figure 8

Results of CAPS analysis of a F2 population derived from a cross between Harosoy-wm and Clark. PCR products were digested with BglII and separated on a 8% polyacrylamide gel. φ molecular marker φx174/HaeIII, P1 Clark, P2 Harosoy-wm, P F2 plants with purple flower and M F2 plants with magenta flower. The CAPS markers co-segregated with the genotypes at Wm locus; F2 plants with magenta flower had 234 and 71 bp bands, whereas F2 plants with purple flower had fragment of either only 306 bp band or all of the three bands. The former and latter were presumed to be homozygotes (WmWm) and heterozygotes (Wmwm) at the Wm locus, respectively. The migration of size markers is shown to the left of the gel

Table 1 Association between flower color and band patterns in CAPS analysis in the F2 population derived from a cross between Harosoy-wm with magenta flower and Clark with purple flower

Linkage mapping

For linkage mapping, SSR markers that are presumed to locate around the Wm locus in molecular linkage group F, Satt193, Satt252, Satt193, Satt423, Satt425 and Sat_039 were used for analysis. Among the SSR markers, Satt269, Satt252, Satt193 and Satt425 exhibited a good separation in the F2 population. Genotypes of the SSR markers and the CAPS marker were used for linkage analysis. Linkage mapping revealed that the Wm locus and gmfls1 were mapped at similar position in the molecular linkage group F (between Satt252 and Satt425) in accordance with the previous report (Fig. 9) (Cregan et al. 1999).

Fig. 9
figure 9

Linkage mapping of gmfls1 and Wm locus using 89 plants of a F2 population developed by a cross between Harosoy-wm and Clark. The name of linkage group is indicated at the top. Distances of markers from the top of the linkage group are shown on the left. The prefix Satt indicates SSR markers. gmfls1 and Wm locus were mapped at similar position in the molecular linkage group F

Expression of FLS gene

Because RNA blot analysis with labeled gmfls1 could not detect any signal, semi-quantitative RT-PCR analysis was performed. Figure 10 illustrates the results of RT-PCR analysis using total RNA from trifoliolate leaves (R3 stage) and flower petals of Harosoy and Harosoy-wm. Amount of RT-PCR products in flower petals were substantially higher than that in leaves. Further, RT-PCR products of Harosoy-wm were lower than Harosoy in both leaves and flower petals.

Fig. 10
figure 10

RT-PCR analysis of RNA isolated from leaves and flower petals of Harosoy and Harosoy-wm. Specific primers for gmfls1 of Harosoy (upper and middle panels) and actin control (lower panel) generated fragments of 383 and 552 bp, respectively. φ molecular marker φx174/HaeIII, Lanes 1 Harosoy (leaf), 2 Harosoy-wm (leaf), 3 Harosoy (flower petal), and 4 Harosoy-wm (flower petal). PCR reaction was performed for 22 (upper panel) and 15 cycles (middle panel) to evaluate FLS gene expression and 30 cycles for actin control. The amount of RT-PCR products using the gmfls1 primer in leaves was lower than in flower petals. Less PCR product was generated from Harosoy-wm than from Harosoy. The migration of size markers is shown to the left of the gel

Amount of flavonoids in flower petals

Amount of anthocyanins in banner petals was similar between Harosoy and Harosoy-wm in 2003 and 2004 (Table 2). In contrast, amount of flavonols in Harosoy-wm was significantly lower (17% in 2003; 9% in 2004) compared to Harosoy.

Table 2 Contents of anthocyanins and flavonols represented by peak areas in HPLC chromatogram ± SE (×103) in flower petals of soybean in 2003 and 2004 at Tsukuba, Japan

Discussion

Flavonol glycosides (kaempferol and quercetin glycosides) are the most abundant flavonoids in soybean leaves (Buttery and Buzzell 1973). We cloned a putative full-length cDNA of soybean FLS gene responsible for conversion from dihydroflavonols to flavonols, gmfls1. gmfls1 is 1,208 bp long and it is presumed to encode a polypeptide, GmFLS1, with 334 amino acids. GmFLS1 has conserved dioxygenase domains A and B and exhibits a significant homology with previously reported FLS. The GmFLS1 of Harosoy had FLS activity as shown by bacterial heterologous expression assay. Genomic Southern analysis revealed that soybean contains three to four copies of FLS gene in the genome. To our knowledge, this is the first report on the isolation of full-length cDNA of FLS from leguminous plants.

The single-base deletion in Harosoy-wm probably shortened the polypeptide to only 37 amino acids. The truncated polypeptide lacks the conserved dioxygenase domains A and B. Bacterial expression assay confirmed that GmFLS1 of Harosoy-wm has no FLS activity.

The CAPS marker that was designed to detect the single-base deletion co-segregated with alleles at Wm locus in a F2 population segregating for the locus. The CAPS marker was mapped at the expected position in the molecular linkage group F. These results strongly suggest that Wm encodes FLS and the single-base deletion produced a truncated polypeptide lacking FLS activity, resulting in the substantial reduction of flavonols in flower petals and leaves. Flavonoid analysis revealed that amount of flavonols in flower petals of Harosoy-wm was 9–17% compared to Harosoy. It suggests that FLS encoded by other FLS gene member(s) may also have FLS activity.

Semi-quantitative RT-PCR analysis revealed that abundance of transcript in Harosoy-wm was lower than Harosoy in both leaves and flower petals. Nonsense-mediated mRNA decay (NMD) is a mRNA surveillance widespread among eukaryotes that detects and eliminates mRNAs harboring premature translation termination codons, thereby preventing the synthesis of truncated and potentially harmful proteins (Baker and Parker 2004; Conti and Izaurralde 2005). NMD may be responsible for the decrease in transcripts in the mutant similar to the F3′H gene in soybean (Toda et al. 2002).

Co-pigmentation of anthocyanins with tannins, flavones and flavonol glycosides can have major effects on flower color (Scott-Moncrieff 1936). When compared over a range of physiological pH values, co-pigmented anthocyanins are usually bluer than anthocyanins alone. The recessive wm allele substantially reduces flavonol glycoside content and it may inhibit co-pigmentation between anthocyanins and flavonol resulting in less bluer (magenta) flower color (Fig. 2) similar to the FLS mutant in petunia (Holton et al. 1993). Co-pigmentation between anthocyanins and flavonol glycosides may contribute to the purple flower color in soybeans having Wm allele.

Recessive wm allele reduces photosynthetic rate and seed yield, and causes earlier leaf senescence (Buzzell and Buttery 1977). Either low levels of flavonol glycosides or high levels of their precursors may be responsible for the deleterious effects. Flavonols are required for pollen tube growth, UV-protection, IAA metabolism or transport, and defense (Stafford 1990). In soybean, additional functions of flavonol glycosides have been reported. Four flavonol glycoside alleles, Fg1 (β(1-6)-glucoside present), Fg2 (α(1-6)-rhamnoside present), Fg3 (β(1-2)-glucoside present), and Fg4 (α(1-2)-rhamnoside present), have been identified in soybean (Buzzell and Buttery 1974). Plants with the Fg1 Fg3 alleles have a lower rate of photosynthesis, lower leaf chlorophyll concentration, lower leaf weight, and lower seed yield (Buttery and Buzzell 1976). Further, Fg1 and Fg3 control waviness of leaf margins in soybean (Buzzell and Buttery 1998). However, in the presence of brown pubescence (TT), wavy leaf is not expressed (Rode and Bernard 1975). Further investigations are necessary to determine the mechanisms how flavonol glycosides in soybean affect a wide variety of characters such as photosynthesis, seed yield, leaf weight, leaf senescence, leaf chlorophyll concentration and leaf waviness.