Two Independent Spontaneous Mutations Related to Anthocyanin-less Flower Coloration in Matthiola incana Cultivars

Abstract

Matthiola incana is a popular winter-flowering plant, and white is considered a valuable flower color for marketed cultivars. In this study, we aimed to identify the genes responsible for white flower coloration in six commercial cultivars of M. incana used as cut and potted flowers. The expression levels of chalcone synthase, flavanone 3-hydroxylase, flavonoid 3'-hydroxylase, dihydroflavonol 4-reductase, anthocyanidin synthase (ANS), and anthocyanidin 3-O-glucosyltransferase in the petals of ‘Kiss me White’ and ‘Pygmy White’ were 0%–48% lower than those in the purple flower ‘Vintage Lavender’, whereas the expression level of basic helix-loop-helix 2 (bHLH2) was two-fold higher. Significantly reduced expression levels of ANS were also detected in four other white flower cultivars: ‘Vintage White’, ‘Iron White’, ‘White Wonder No. 2’, and ‘Quartet White’. All investigated white flower cultivars had a single nucleotide deletion in the first exon of ANS, which we designated as ans-1. This generates a frameshift mutation and a nonsense codon. In addition to ans-1, ‘Kiss me White’ and ‘Pygmy White’ have a 481-bp insertion within bHLH2. This insertion has features of hAT-type transposable elements and was designated as dTmi1. All white flower cultivars contain the ans-1 mutation, whereas ‘Kiss me White’ and ‘Pygmy White’ are double mutants containing both bhlh2^dTmi1 and ans-1. ‘Kiss me Yellow’, which accumulates carotenoids, but not anthocyanins, in its petals possesses the bhlh2^dTmi1 allele, but not the ans-1 allele. Therefore, either bhlh2^dTmi1 alone or ans-1 alone can lead to a deficiency in anthocyanin production in commercial cultivars of M. incana. We also developed co-dominant DNA markers that can distinguish between wild-type and mutant alleles of both bHLH2 and ANS. In combination with other previously developed markers that can distinguish between single- and double-flowered individuals, these markers will be useful for nursery plant management and breeding of commercial M. incana.

Introduction

Matthiola incana (L.) R. Br., commonly known as garden stock or common stock, is a popular winter-flowering plant in Japan. It is used mainly as cut flowers (standard- and spray-types) and as potted flowers. The flowers of M. incana cultivars have a wide range of colors, from pure white to purple, pale purple, blue, red, and pale yellow (Dole and Wilkins, 2005). Anthocyanins, particularly pelargonidin and cyanidin derivatives, are responsible for these various M. incana flower colors, while delphinidin derivatives have not been detected in its commercial cultivars (Tatsuzawa et al., 2012). A white flower color is a result of anthocyanin deficiency, and white flower cultivars of M. incana accumulate kaempferol derivatives in their petals (Tatsuzawa et al., 2014).

M. incana has been subjected to advanced genetic studies since the early 20^th century (Saunders, 1928). Genes C and R have long been recognized as essential for the production of anthocyanins in M. incana flowers. Thus, a homozygous state in either c or r results in a white flower phenotype (Crane and Lawrence, 1947; Saunders, 1928). In the absence of anthocyanin, flowers may exhibit three color forms: white, sulfur-white, and cream, with the latter two deriving their yellowish color from the presence of carotenoid (Crane and Lawrence, 1947). These different shades of white are a result of differences in a single gene, and the genotypes of white, sulfur-white, and cream flower colors correspond to WW, Ww, and ww, respectively (Crane and Lawrence, 1947). M. incana has been used as a suitable flower to characterize the flavonoid/anthocyanin biosynthetic pathway, resulting in the establishment of various nearly isogenic lines (Seyffert, 1971), from which three white flower mutant lines, designated as e, f, and g, have been isolated (Hemleben et al., 2004). The f mutant is deficient in chalcone synthase (CHS) activity (Hemleben et al., 2004; Spribille and Forkmann, 1981). The white flower and glabrous e mutant has a single nucleotide change within the TRANSPARENT TESTA GLABRA1 (MiTTG1) gene, leading to an amino acid substitution located in the WD motif (W158R) (Dressel and Hemleben, 2009). In the white flower g mutant, anthocyanin biosynthesis can be activated by two Arabidopsis basic helix-loop-helix (bHLH) genes: MYC-146 (ENHANCER OF GLABRA 3 (EGL3)) and GLABRA 3 (GL3) (Ramsay et al., 2003). Thus, it has been speculated that the g mutant is deficient in an ortholog of the bHLH transcription factor for anthocyanin biosynthesis; however, the bHLH genes that are responsible for the white flower coloration of M. incana have not yet been identified. In addition, it remains to be determined whether other mutations besides CHS, bHLH, and TTG1 are present in commercial M. incana white flower cultivars.

The structural genes involved in anthocyanin biosynthesis in M. incana have been reported previously (Chen et al., 2018; Nuraini et al., 2020; Seitz et al., 2006). More recently, using the M. incana cultivar ‘Vintage Lavender’, we identified the genes MiMYB1, MibHLH2, and MiWDR1 (=MiTTG1) as components of the anthocyanin biosynthesis-activator of the MYB-bHLH-WDR (MBW) complex (Nuraini et al., 2020). MiMYB1, MibHLH2, and MiWDR1 are the orthologs of Arabidopsis PRODUCTION OF ANTHOCYANIN PIGMENTS 1 (PAP1), EGL3, and TTG1 (Nuraini et al., 2020), respectively. The petal-specific expression of MiMYB1 correlates well with anthocyanin accumulation in petals, and the ectopic expression of MiMYB1 by the turnip mosaic virus (TuMV) can induce anthocyanin accumulation in leaves. Moreover, we demonstrated that the other bHLH protein, MibHLH1, which is an ortholog of Arabidopsis TRANSPARENT TESTA 8 (TT8), may be an enhancer of anthocyanin biosynthesis in M. incana petals (Nuraini et al., 2020).

M. incana may be either single- or double-flowered, and most cultivars are grown double-flowered individuals because this form is more commercially advantageous (Nakatsuka and Koishi, 2018). However, double-flowered M. incana do not produce seeds as they lack reproductive organs. Generally, the eversporting, single-flower trait that is associated strongly with pollen sterility is selected for M. incana cultivars, resulting in 50%–60% double-flowered individuals (Dole and Wilkins, 2005). Double-flowered seedlings of eversporting can be selected based on distinct leaf characteristics such as serrated leaves, pale yellow leaves, and round and large cotyledons because the double flower trait is linked strongly to the visible traits of the seedlings (Ecker et al., 1993). The seed testa color in M. incana is usually blackish, but that in certain white flower cultivars is yellow (Dole and Wilkins, 2005). Among eversporting lines of some cultivars, the double flower trait is strongly linked to yellow seed color. Therefore, by purchasing only yellow seeds, growers can select mostly double flowered individuals, and these cultivars are commonly known as “all-double” (Fujita, 1994).

White is a commercially important flower color, and several studies have investigated the molecular mechanisms that generate white flowers. In the Japanese morning glory (Ipomoea nil), a frameshift mutation in InMYB1 (anthocyanin biosynthetic regulatory gene) produces white flowers with red stems and colored seeds (Morita et al., 2006). White flower cultivars of the Japanese gentian (Gentiana triflora) have either a mutation in the anthocyanidin synthase gene (ANS) (Nakatsuka et al., 2005b) or a mutation in the transposable element inserted into GtMYB3 (anthocyanin biosynthetic regulatory gene) (Nakatsuka et al., 2008). In Eustoma grandiflorum, a 94-bp deletion in ANS leads to white coloration (Shimizu et al., 2011). In the white flower cultivar Torenia hybrida ‘Crown White’, a retrotransposable element is inserted into a 5' region upstream of the flavanone 3-hydroxylase gene (F3H), which suppresses F3H promoter activity (Nishihara et al., 2014). White-colored Lilium speciosum has two independent mutations (Suzuki et al., 2015). The first is an amino acid substitution in the R2 repeat of LsMYB, which leads to reduced transcription of the anthocyanin biosynthetic gene and no tepal pigmentation. The second is a nonsense mutation of the dihydroflavonol 4-reductase gene (DFR), which results in the absence of anthocyanin accumulation in tepals and anthers (Suzuki et al., 2015).

Previous studies have demonstrated that the white flower trait is associated with double flowers, pollen sterility, glabrous leaves, and seed color (Dressel and Hemleben, 2009; Hemleben et al., 2004; Saunders, 1928). To identify the genes responsible for the white flower coloration in M. incana cultivars, we selected six representative white flower M. incana cultivars possessing various genetic characteristics in this study (Table 1). ‘Iron White’ and ‘Quartet White’ are typically produced as standard- and spray-type cut flower cultivars in Japan, respectively, and they have hairy leaves. ‘White Wonder No. 2’ is an all-double-type cut flower cultivar with white flowers and yellow seed color loci that are linked strongly to both loci with regard to double-flower and pollen sterility (Fujita, 1994). ‘Pygmy White’ is a potted flower cultivar that possesses serrate and glabrous leaves (Fujita, 1994). ‘Kiss me White’ is a derivative of ‘Pygmy White’, but does not possess the serrated leaf trait. ‘Vintage White’ is a potted flower cultivar that has hairy leaves. We hypothesized that the genes responsible for anthocyanin-less phenotypes differ among cultivars with hairy leaves, those lacking hairy leaves, and all-double cultivars. Here, we identified the two independent genes responsible for anthocyanin-less coloration in commercial M. incana cultivars.

Table 1

Features of Matthiola incana commacal white flower cultivars used in this study.

Materials and Methods

Plant materials

We purchased seeds of ‘Vintage Lavender’ and ‘Vintage White’ from M&B flora (Hokuto, Japan), and ‘Kiss me White’, ‘Pygmy White’, ‘Iron White’, ‘White Wonder No. 2’, ‘Quartet White’, ‘Kiss me Blue’, ‘Kiss me Rose’, ‘Kiss me Cherry’, ‘Kiss me Salmon’, and ‘Kiss me Yellow’ from Takii Seed (Kyoto, Japan). All plants were grown in a greenhouse at Shizuoka University (Shizuoka, Japan) under natural conditions from September 2019 to February 2020. The samples were collected from petals at the S3 and S4 stages of flower development as described previously (Nuraini et al., 2020) and from young leaves. After collection, all samples were frozen immediately using liquid nitrogen and stored at −80°C until use.

Flavonoid analysis

Anthocyanin and flavonol analyses were performed as described previously (Nuraini et al., 2020). For anthocyanins, 200 mg of petal samples at stage 4 of flower development were ground and extracted with 40% methanol and 10% acetic acid for 24 h at 4°C. For flavonols, one gram of petal samples collected from each cultivar was ground using liquid nitrogen and extracted with 5 mL of 80% methanol for 24 h at 4°C. To convert flavonoids to aglycone by hydrolysis, 4.5 mL of 3 M hydrochloric acid was added to 500 μL of extracts, and the mixture was boiled for 1 h. Flavonol aglycones were extracted using ethyl acetate, and the resulting extracts were dried and then resuspended with 500 μL of 80% methanol. The anthocyanin and flavonol aglycone extracts were filtered through a 0.22-μm polytetrafluoroethylene syringe filter (Osaka Chemical, Osaka, Japan). Flavonoid analyses were conducted using the Agilent Infinity 1290 High Performance Liquid Chromatography system (Agilent Technologies, Santa Clara, CA, USA) and Agilent 1200 Diode-Array Detector SL (Agilent Technologies), with the reverse-phase column YMC-Pack Pro C18 RS (4.6 × 150 mm, 5 μm; YMC, Kyoto, Japan). Solvent A [1.5% (v/v) phosphoric acid] and solvent B [1.5% (v/v) phosphoric acid, 20% (v/v) acetic acid and 25% (v/v) acetonitrile] were used for elution at 40°C at a flow rate of 0.8 mL·min⁻¹. Elution was performed with 20%–85% of solvent B for 40 min under gradient elution conditions. Anthocyanins and flavonols were monitored based on their absorbance levels at 520 and 360 nm, respectively. Flavonol concentrations were calculated from a standard curve using authentic standards consisting of kaempferol, quercetin, and myricetin (Extrasynthese, Lyon, France). Five biological replicates were used for flavonoid analyses.

Gene expression analysis of anthocyanin biosynthetic genes

Total RNA was isolated from the petals of each cultivar at the S3 stage of flower development using RNAiso Plus (Takara Bio, Otsu, Japan) and purified using lithium acetate precipitation as described previously (Nuraini et al., 2020). Genomic DNA was eliminated and then cDNA synthesized using a PrimeScript RT Reagent Kit with gDNA Eraser (Takara Bio). RT-qPCR was performed for each gene using a Thermal Cycler Dice Real Time System (TP850; Takara Bio) and TB Green Fast qPCR mix (Takara Bio). The reaction mixture (10 μL) comprised 1× Master Mix, 0.2 μM of each primer, and 1 μL of template cDNA. The cycling conditions were as follows: 95°C for 20 s, followed by 40 cycles of 95°C for 1 s and 60°C for 20 s. The primer sets used to quantify M. incana anthocyanin biosynthetic genes have been described previously (Nuraini et al., 2020). The specificity of each amplification reaction was checked by a dissociation analysis step added after the cycle reaction. Data were analyzed by second derivative maximum methods using Thermal Cycler Dice Real Time System II version 5.0 software (Takara Bio). The transcript level of each gene was calculated relative to that of the reference gene for actin encoded by MiACT2 (LC496063) in M. incana. RT-qPCR was performed using five biological replicates.

Genomic PCR analysis of anthocyanin biosynthesis-related genes

Genomic DNA was isolated from 10 g of young leaves using the cetyl trimethylammonium bromide method as described previously (Nakatsuka et al., 2005a), with some modifications. In addition, genomic fragments from CHS, DFR, ANS, 3GT, MYB1, bHLH2, and WDR1 in 12 cultivars were amplified. The primer sets used are shown in Table S1. Genomic fragments from bHLH2 were also amplified using bHLH2-U881 (5'-CGC TTG CTG GAA AGG TTG CTT GTG ATG C-3') and bHLH2-L1405 (5'-ACT CCA ATT CTT GAA CCC GCC TTT GCA GT-3'). The PCR mixture (50 μL) contained 100 ng of genomic DNA, 1× Ex Taq buffer, 200 μM dNTPs, 0.4 μM of each primer, and 0.25 units of Ex Taq Hot Start (Takara Bio). The reaction conditions were as follows: denaturation at 94°C for 2 min, followed by 35 cycles of 95°C for 20 s, 60°C for 40 s, and 72°C for 2 min, followed by extension at 72°C for 10 min. PCR products were separated on 1.0% or 1.5% agarose gels and then stained with Midori-Green Advance (Nippon Genetics, Tokyo, Japan). The amplified fragments were subcloned into a pGEM-T Easy vector system (Promega, Madison, WI, USA). The sequences of all constructs were confirmed by DNA sequencing (Fasmac, Atsugi, Japan).

To investigate the presence or absence of nucleotide polymorphisms among white flower cultivars and ‘Vintage Lavender’, we conducted a T7 endonuclease I assay using the genomic PCR fragments described above. Reaction mixtures (19 μL) comprised 2.5 μL of each PCR product, 2 μL of 10× NEBuffer2, and 12 μL of sterile water. PCR was performed under the following conditions: 95°C for 5 min, followed annealing at 95°C–85°C (−2°C/sec) and at 85°C–25°C (−0.1°C/sec). Annealed PCR products were incubated at 37°C for 15 min after the addition of 1 μL of T7 endonuclease I (10 units; New England Biolab Japan Inc., Tokyo, Japan). The reaction products were separated on 1% agarose gels, and then stained with Midori-Green Advance DNA stain.

Development of a PIRA-PCR marker for the ANS mutation

To distinguish between ANS and ans-1 alleles, we used primer-introduced restriction analysis (PIRA)-PCR (Haliassos et al., 1989; Neff et al., 1998). PCR was performed as described above using the following PIRA-PCR primer set designed to distinguish the single nucleotide deletion of ans-1: ANS_PIRA-U441, 5'-TTA TCC TCA AGA CAA GAG AGA TCT CTC ACT TTG GCC CAA AAC AGC-3' (underlined guanine is substituted for cytosine) and ANS_PIRA-L749, 5'-CAG GCT GAG GAC ATT TTG GGT AAT-3'. Ten microliters of the PCR mixture was added to 10 μL of Alu I reaction mixture containing 1× CutSmart buffer and 5 units of Alu I (New England BioLab Japan); the mixture was incubated at 37°C for 1 h. The reaction products were separated on 2% agarose gel and then stained as shown above.

Southern blot analysis of bHLH2 and ANS

To determine the number of copies of bHLH2 and ANS in the M. incana genome, we performed Southern blot analysis as described previously (Nakatsuka et al., 2005a). Ten micrograms of genomic DNA were digested by Hind III and then separated on a 0.7% agarose gel. The separated DNAs were transferred to a Hybond-N membrane (GE healthcare Life Science, Little Chalfont, England). The open reading frames (ORFs) of bHLH2 and ANS were labeled using the DIG-High Prime DNA Labeling and Detection starter Kit II (Sigma-Aldrich Japan, Tokyo, Japan) and used as probes.

Results

Flavonoid analysis in white flower cultivars

No peaks were detected at 520 nm by HPLC analysis in all six white flower cultivars, which implied that there was no accumulation of anthocyanin pigments in their petals (Figs. 1A and S1). In contrast, post-hydrolysis measurements of flavonol concentrations in the petals of each white flower cultivar showed that flavonol derivatives accumulated in the petals of all six white flower cultivars at levels 2.4- to 3.6-fold higher than that in ‘Vintage Lavender’, but there was no significant difference among the values (Fig. 1B). These results indicate that the white flower cultivars are deficient in anthocyanin biosynthesis, but not in flavonol biosynthesis.

Fig. 1

Phenotypes and flavonol concentrations of white-flowered M. incana cultivars. A: VW, ‘Vintage White’; KW, ‘Kiss me White’; PW, ‘Pygmy White’; IW, ‘Iron White’; WW, ‘White Wonder No. 2’; QW, ‘Quartet White’; VL, ‘Vintage Lavender’ (control). Bar = 5 cm. B: Flavonol accumulation in petals of seven cultivars at the anthesis stage (S4). Flavonoids extracted from each sample were converted to aglycones by hydrolysis and quantified using kaempferol as a standard. Values are expressed as mean ± standard error (n = 5). There was no difference among any values by the Tukey–Kramer method (P < 0.05).

Comparative expression analysis of anthocyanin biosynthetic genes

To compare the different expression levels of anthocyanin biosynthesis-related genes among white flower cultivars and ‘Vintage Lavender’, we determined the expressions of these genes using petals at the S3 stage of flower development, when they are expressed strongly (Nuraini et al., 2020).

RT-qPCR showed that the expression levels of CHS, F3H, F3'H, DFR, ANS, 3GT, and FLS1 in the petals of ‘Kiss me White’ and ‘Pygmy White’ were suppressed by 0%–48% compared with those in ‘Vintage Lavender’, although there was no significant difference in the expression levels of some genes (Fig. 2). Furthermore, the expression levels of bHLH2 in both white flower cultivars were two-fold higher than those in ‘Vintage Lavender’. The expression levels of ANS in ‘Vintage White’, ‘Iron White’, ‘White Wonder No. 2’, and ‘Quartet White’ were significantly lower than those in ‘Vintage Lavender’ (Fig. 2). The expression level of WDR1 in all white flower cultivars was also significantly lower than that in ‘Vintage Lavender’ (Fig. 2). The expression level of FLS1 in all white flower cultivars, except ‘Vintage White’ and ‘Iron White’, was lower than that in ‘Vintage Lavender’ (Fig. 2). The expression levels of CHS, CHI, F3H, DFR, 3GT, and bHLH1 were higher in ‘Vintage White’ than in ‘Vintage Lavender’. The expression level of F3'H was lower in all white flower cultivars except ‘White Wonder No. 2’ (Fig. 2).

Fig. 2

Relative expression levels of anthocyanin biosynthetic genes in petals of white-flowered cultivars. The expression levels of chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), flavonoid 3'-hydroxylase (F3'H), dihydroflavonol 4-reductase (DFR), anthocyanidin synthase (ANS), anthocyanidin 3-O-glucosyltransferase (3GT), flavonol synthase 1 (FLS1), MYB1, basic helix-loop-helix 1 (bHLH1), bHLH2, and WD repeat 1 (WDR1) were normalized relative to ACT2. Values are expressed as mean ± standard error (n = 5). Different letters indicate significant difference by the Tukey-Kramer method (P < 0.05). VL, ‘Vintage Lavender’; VW, ‘Vintage White’; KW, ‘Kiss me White’; PW, ‘Pygmy White’; IW, ‘Iron White’; WW, ‘White Wonder No. 2’; QW, ‘Quartet White’.

Based on expression analysis, we classified the white flower cultivars into two groups. Group 1, comprising ‘Kiss me White’ and ‘Pygmy White’, was characterized by suppressed expression of most anthocyanin biosynthetic structural genes. Here, we assumed that mutations in transcriptional factors were responsible for anthocyanin deficiency. In group 2, comprising cultivars ‘Vintage White’, ‘Iron White’, ‘White Wonder No. 2’, and ‘Quartet White’, the expression levels of most anthocyanin biosynthetic structural and regulatory genes were not suppressed and were often higher compared with those in ‘Vintage Lavender’. Therefore, we hypothesized that the anthocyanin deficiency was a result of functional deficiency in certain structural genes in group 2.

Identification of bHLH2 mutation in ‘Kiss me White’ and ‘Pygmy White’

We hypothesized that group 1 white flower cultivars have a mutation in either MYB1, bHLH2, or WDR1, which are components of the MBW complex (Nuraini et al., 2020). The expression level of bHLH2 in ‘Kiss me White’ and ‘Pygmy White’ was increased significantly compared with that in the other four white flower cultivars, and in the latter, it was two times higher than that in ‘Vintage Lavender’ (Fig. 2). On the other hand, the expression level of MYB1 in ‘Kiss me White’ and ‘Pygmy White’ did not differ significantly from that in other cultivars (Fig. 2). The expression level of WDR1 in all white flower cultivars was significantly lower than that in ‘Vintage Lavender’. Genomic fragments of MYB1, bHLH2, and WDR1 were amplified from the genomic DNA of each cultivar. The lengths of the 1.3-kb and 1.0-kb fragments amplified from MYB1 and WDR1 were the same in ‘Kiss me White’ and ‘Pygmy White’ as in ‘Vintage Lavender’ and other white flower cultivars (data not shown). In contrast, the 5-kb amplified fragment from the bHLH2 genome in ‘Kiss me White’ and ‘Pygmy White’ was longer than the 4.5-kb fragment in other cultivars (data not shown). To find an insertion in the bHLH2 genome of both white flower cultivars, PCR was performed using several primer combinations. Approximately 500-bp long fragments were amplified using bHLH2-U881 and bHLH2-L1405 primers in both ‘Kiss me White’ and ‘Pygmy White’ compared with 700-bp fragments in other cultivars (Fig. 3A).

Fig. 3

Genomic PCR analysis of bHLH2 in white-flowered cultivars. A: Amplified fragments of bHLH2 using genomic DNAs from seven cultivars including ‘Vintage Lavender (VL)’, ‘Vintage White (VW)’, ‘Kiss me White (KW)’, ‘Pygmy White (PW)’, ‘Iron White (IW)’, ‘White Wonder No. 2 (WW)’, and ‘Quartet White (QW)’. ‘+’ represents anthocyanin-accumulating cultivars, whereas ‘−’ represents cultivars that did not accumulate anthocyanin in their petals. Genomic PCR was performed using bHLH2-U881 and bHLH2-L1405 primers. Amplified fragments were separated on 2.0% agarose gels. The 0.7-kb and 1.2-kb fragments correlate to bHLH2 and bhlh2^dTmi1 alleles, respectively. B: Genome structure of bHLH2 in ‘Vintage Lavender’ (accession number LC557498), ‘Kiss me White’ and ‘Pygmy White’ (LC557495). Empty and full boxes indicate untranslated regions and CDS within exons, respectively. The numbers below boxes and above lines indicate the number of base pairs of exons and introns, respectively. Arrows indicate the positions of bHLH2-U881 and bHLH2-L1405 primers. The inverted triangle indicates the insertion position of dTmi1 transposable element. C: Sequences of the dTmi1 transposable element. TSD is an 8 bp of target site duplication. Arrows indicate 13-bp inverted repeats. dTmi1 exhibits typical features of hAT transposable elements. D: Amplified fragments of bHLH2 in six cultivars of the ‘Kiss me series’ including ‘Kiss me Blue (KB)’, ‘Kiss me Rose (KR)’, ‘Kiss me Cherry (KC)’, ‘Kiss me Salmon (KS)’, ‘Kiss me Yellow (KY)’, and ‘Kiss me White (KW)’.

The 1.2-kb fragments amplified from bHLH2 in ‘Kiss me White’ and ‘Pygmy White’ were subcloned and sequenced. Sequences from both cultivars had a 481-bp insertion sequence in the 6^th exon of bHLH2 (Fig. 3B). This insertion had an 8-bp target site duplication (TSD) and 13-bp inverted repeats (IRs) at both ends, which are features that correspond to hAT-type transposable elements. Here, we designated this as dTmi1 (Fig. 3C). This insertion does not contain protein-coding sequences; thus, dTmi1 is likely a non-autonomous transposable element.

Southern blot analysis showed that bHLH2 exists as a single copy within the M. incana genome and that the gene is longer in ‘Kiss me White’ and ‘Pygmy White’ due to the dTmi1 insertion (Figs. S2A and B).

To determine the relationship between the bhlh2^dTmi1 allele and flower coloration in the ‘Kiss me’ series of cultivars, we performed genomic PCR for bHLH2 in various cultivars of the ‘Kiss me’ series with different flower colors. All cultivars had the glabrous leaf phenotype. ‘Kiss me Blue’, ‘Kiss me Rose’, ‘Kiss me Salmon’, and ‘Kiss me Cherry’ accumulated anthocyanin pigments in their petals, whereas ‘Kiss me Yellow’ did not accumulate anthocyanin pigments (Fig. S3). Carotenoid pigment was responsible for the cream flower coloration in ‘Kiss me Yellow’ (data not shown). Among the six cultivars with different flower colors, a 1.2-kb fragment corresponding to the bhlh2^dTmi1 allele was amplified in ‘Kiss me Yellow’ and ‘Kiss me White’. Hence, the bhlh2^dTmi1 allele correlates well with anthocyanin deficiency in the petals of various cultivars. The enhanced expression level of bHLH2 and the suppressed expression levels of CHS, F3H, DFR, ANS, and 3GT were also detected in not only ‘Kiss me White’, but also ‘Kiss me Yellow’ (Fig. 4).

Fig. 4

Relative expression levels of anthocyanin biosynthetic genes in petals of the ‘Kiss me series’. The expression levels of chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), dihydroflavonol 4-reductase (DFR), anthocyanidin synthase (ANS), anthocyanidin 3-O-glucosyltransferase (3GT), MYB1, basic helix-loop-helix 2 (bHLH2), and WD repeat 1 (WDR1) were normalized relative to ACT2. Values are expressed as mean ± standard error (n = 5). Different letters indicate significant difference by the Tukey–Kramer method (P < 0.05). n.s. indicates no significant difference among all cultivars. KB, ‘Kiss me Blue’; KR, ‘Kiss me Rose’; KC, ‘Kiss me Cherry’; KS, ‘Kiss me Salmon’; KY, ‘Kiss me Yellow’; KW, ‘Kiss me White’.

Identification of the ans-1 mutation in white flower cultivars in group 2

The group 2 cultivars ‘Vintage White’, ‘Iron White’, ‘White Wonder No. 2’, and ‘Quartet White’ did not contain any mutations in bHLH2 (Fig. 3A). In these cultivars, we assumed that mutations in either DFR, ANS, or 3GT are responsible for anthocyanin deficiency. We amplified fragments approximately 1.5, 1.2, and 1.5 kb in length from genomic regions corresponding to the DFR, ANS, and 3GT ORFs, respectively (Figs. S4A, S5A, and S6A). All seven cultivars yielded PCR products of the same fragment length. To investigate whether mismatches of DFR, ANS, and 3GT genomic sequences are present between ‘Vintage Lavender’ and each white flower cultivar, we mixed PCR products and then treated them with T7 endonuclease I (Figs. S4B, S5B, and S6B). PCR products of ANS had two partially digested fragments (ca. 500 and 700 bp) in all white flower cultivars, but the fragments were not detected in ‘Vintage Lavender’ (Fig. S5B). Conversely, the digested pattern of DFR and 3GT did not differ between ‘Vintage Lavender’ and all white flower cultivars (Figs. S4B and S6B). Therefore, ANS genomic fragments from white flower cultivars were subsequently subcloned and sequenced.

ANS sequences from white flower cultivars and ‘Vintage Lavender’ had single-nucleotide polymorphisms, including deletions, only in the white flower cultivars, of a single nucleotide at the 486^th position from the start codon of ANS (Fig. 5A). This deletion exists in the 1^st exon of ANS and leads to a frameshift mutation, resulting in a nonsense codon in the 3' position immediately after the deletion. We designated this mutated ANS allele in the white flower cultivars as ans-1.

Fig. 5

Genomic sequences of ANS in white-flowered cultivars. A: Genomic structure of ANS in colored- (accession number LC557496) and white-flower cultivars (LC557497). A single nucleotide at the 486^th position from the initiation codon of the ANS gene is deleted in white cultivars. ‘−’ indicates the deletion of ans-1, resulting in frameshift and nonsense mutations. Full boxes indicate untranslated regions and CDS within exons. The numbers above boxes and lines indicate the number of base pairs of exons and introns, respectively. Upper and lower case exhibit exon and intron sequences, respectively. B: PIRA-PCR analysis for ANS alleles in seven cultivars including ‘Vintage Lavender (VL)’, ‘Vintage White (VW)’, ‘Kiss me White (KW)’, ‘Pygmy White (PW)’, ‘Iron White (IW)’, ‘White Wonder No. 2 (WW)’, and ‘Quartet White (QW)’. ‘+’ represents anthocyanin-accumulating cultivars, whereas ‘−’ represents cultivars that do not accumulate anthocyanin in their petals. Genomic PCR was performed using ANS-PIRA-U441 and ANS-PIRA-L749 primers, and the amplified fragments were digested with Alu I. Afterwards, fragments were separated on 2.0% agarose gels. The 270-bp and 300-bp fragments correlate to the ANS and ans-1 alleles, respectively. C: PIRA-PCR analysis for ANS alleles in six cultivars of the ‘Kiss me series’ including ‘Kiss me Blue (KB)’, ‘Kiss me Rose (KR)’, ‘Kiss me Cherry (KC)’, ‘Kiss me Salmon (KS)’, ‘Kiss me Yellow (KY)’, and ‘Kiss me White (KW)’.

Southern blot analysis revealed an approximately 7.0-kb fragment in all cultivars (Fig. S2C). No Hind III-recognizing site was present in the ANS genomic region (Fig. S2A). Therefore, these findings suggest that ANS exists as a single copy in the M. incana genome and that no genome rearrangement has occurred in any of the white flower cultivars.

All white flower cultivars belonging to group 2 exhibited significantly suppressed levels of WDR1 (Fig. 2). However, we were unable to find any mutation within each respective WDR1 gene sequence (data not shown).

Development of a marker to distinguish between ANS and ans-1 alleles

A single nucleotide deletion in the 1^st exon of ANS induced a frameshift mutation, resulting in the ans-1 allele (Fig. 5A). The ans-1 allele is indistinguishable from the ANS allele based on their respective amplified fragment lengths, unlike the situation between bHLH2 and bhlh2^dTmi1. We thus employed PIRA-PCR (Haliassos et al., 1989; Neff et al., 1998) to generate a specific marker for each ANS allele. We designed a PIRA-PCR marker to distinguish between the ANS and the ans-1 alleles, and the appropriate mismatch was introduced into the ANS_PIRA-U441 primer. Although the ANS_PIRA-U441 and ANS_PIRA-U749 primers amplified 300-bp fragments in all cultivars, the recognition sequence of the restriction enzyme Alu I (AG|CT) was present only in the ANS allele. After digestion with Alu I, a 270-bp fragment was generated only in ‘Vintage Lavender’, but not in all the white flower cultivars (Fig. 5B). Similarly, the ANS gene of six cultivars belonging to the ‘Kiss me’ series was genotyped using PIRA-PCR. The 270-bp Alu I fragment, which indicated the ANS allele, was detected in ‘Kiss me Blue’, ‘Kiss me Rose’, ‘Kiss me Cherry’, ‘Kiss me Salmon’, and ‘Kiss me Yellow’, but not in ‘Kiss me White’ (Fig. 5C). Therefore, this PIRA-PCR marker can distinguish between ANS and ans-1 alleles. These results indicate that the loss of anthocyanin in ‘Kiss me Yellow’ is due solely to the mutation in bhlh2^dTmi1, which differs from ‘Kiss me White’, in which a bhlh2^dTmi1 and ans-1 double mutation occurs (Figs. 3B and 5C).

Discussion

We previously investigated the expression profiles of anthocyanin biosynthetic genes in the purple flower cultivar ‘Vintage Lavender’ (Nuraini et al., 2020). In this cultivar, the regulatory complex, which is composed of MYB1, bHLH2, and WDR1, activates the transcription of anthocyanin biosynthetic structural genes, resulting in the accumulation of anthocyanin in the petals of M. incana (Nuraini et al., 2020). In the present study, we aimed to identify the genes responsible for the white-flower coloration in commercial M. incana cultivars.

None of the white flower cultivars accumulated anthocyanin in their petals (Fig. S1), although their flavonol levels tended to be higher than those of ‘Vintage Lavender’ (Fig. 1B). Based on flavonoid analysis, we hypothesized that white flower cultivars have a mutation in any of three anthocyanin biosynthetic structural genes, namely DFR, ANS, and 3GT, or three regulatory genes, namely MYB1, bHLH2, and WDR1. Previous studies have reported that single nucleotide substitutions of CHS and WDR1 were responsible for the white flower color phenotype in f and e mutants, respectively (Dressel and Hemleben, 2009; Hemleben et al., 2004). After confirming the genomic sequences of both CHS and WDR1 in the six white flower cultivars, we did not find any base substitutions in their genomic sequences (data not shown).

In ‘Kiss me White’ and ‘Pygmy White’, the expression levels of CHS, F3H, F3'H, DFR, ANS, and 3GT were suppressed to less than half of those in anthocyanin-accumulated cultivars (Fig. 2). We speculate that in both potted cultivars, white flower coloration arises by the same mutation. ‘Kiss me White’ is a cultivar derived from ‘Pygmy White’ (Plant variety protection, <http://www.hinshu2.maff.go.jp/index.html>). In our study, group 1 was assumed to have a mutation in one of three regulatory genes: MYB1, bHLH2, and WDR1. Among them, the expression level of bHLH2 in ‘Kiss me White’ and ‘Pygmy White’ increased two-fold compared with that in ‘Vintage Lavender’ (Fig. 2). The expression level of MYB1 in these two cultivars did not significantly differ from that in ‘Vintage Lavender’, and there were no mutations in the MYB1 genomic sequence. The expression level of WDR1 in both white flower potted cultivars was suppressed significantly compared with that in ‘Vintage Lavender’, but they did not have any mutations.

Genomic PCR analysis showed that a fragment amplified from bHLH2 in group 1 was approximately 500 bp longer than that in the other cultivars (Fig. 3A). Sequence analysis revealed a 481-bp insert in the 6^th exon of bHLH2 in group 1 cultivars (Fig. 3B). This insertion exhibits features of hAT-type transposable elements (Kunze and Well, 2002), including 8-bp of TSD and 13-bp of IR at both ends. In addition, these transposable elements are assumed to be non-autonomous because no proteins are encoded within their sequences (Fig. 3C). This is the first report identifying a transposable element in M. incana, and we designate this transposable element dTmi1.

In many ornamental flowers, transposable elements contribute to white flower formation. For example, in the Japanese gentian ‘Polarno White’, GtMYB1, one of the anthocyanin regulatory genes, is mutated via insertions of either a hAT-type transposable element (dTgt1) or an uncharacterized sequence (Nakatsuka et al., 2008). In the torenia cultivar ‘Crown White’, an LTR-type retrotransposable element TORE1 is inserted into the 5' region upstream of TfF3H (Nishihara et al., 2014). The bHLH domain of bHLH2, located in the 6^th and 7^th exons (Fig. 3B), is predicted to bind DNA. Regions outside the bHLH domain are also crucial for stabilizing dimerization and DNA-binding specificity (Feller et al., 2011). In group 1 white flower M. incana cultivars, transcripts from the bhlh2^dTmi1 allele did not contain the bHLH domain due to a nonsense mutation within the 481-bp insert (Fig. 3B, C). However, we could not confirm why the expression level of bHLH2 was enhanced in both white flower cultivars. In dahlia, the spontaneous yellow ray floret variety ‘Michael J Yellow (MJY)’ accumulates flavone and butein, but not anthocyanin; moreover, DvCHS1, DvF3H, DvDFR, DvANS, and DvIVS are suppressed (Ohno et al., 2011). DvIVS encodes the bHLH transcription factor for anthocyanin biosynthesis, and in ‘MJY’, a 5.4-kb CACTA transposable element, Tdv1, is inserted into DvIVS. In pale aleurone pigmented maize, the r-mi mutation is caused by a 396-bp Ds transposable element in the bHLH domain of R (Liu et al., 1998). Thus, mutations in the bHLH transcription factor lead to the non-accumulation of anthocyanin in several plants. Furthermore, the downregulation of CHS, F3'H, DFR, ANS, and 3GT in bhlh2^dTmi1 mutant cultivars (Figs. 2 and 4) implies that the activation of these genes is controlled by the MYB1-bHLH2-WDR1 complex.

In addition to mutations in bHLH2, all investigated white flower cultivars had a single nucleotide deletion in the 1^st exon of ANS (Fig. 5). This deletion results in a nonsense codon via a frameshift mutation; therefore, the ans-1 allele functions nonenzymatically (Fig. 5A). White flower coloration in lisianthus also involves an ANS mutation, specifically a 94-bp deletion in exon 2 (Shimizu et al., 2011). Japanese gentian lacks the ANS locus, which is responsible for its white coloration (Nakatsuka et al., 2005b). Therefore, white flower coloration due to a deficiency in ANS may be a conserved trait in many ornamental plants. In the present study, we identified two mutated genes, bhlh2^dTmi1 and ans-1, that were independently responsible for white flower coloration in commercial M. incana cultivars. In Japanese gentian and L. speciosum, two distinct pairs of spontaneous mutations are involved in white flower development, i.e., GtMYB1 and ANS (Nakatsuka et al., 2005b, 2008) and LsMYB12 and DFR (Suzuki et al., 2015), respectively. In three white flower mutant lines of M. incana established by Seyffert (1971), e, f, and g alleles were thought to be WDR1, CHS, and bHLH mutants, respectively (Dressel and Hemleben, 2009; Hemleben et al., 2004; Ramsay et al., 2003). Among the aforementioned mutants, the g mutant was complementary to anthocyanin accumulation when two Arabidopsis bHLH genes (EGL3 and GL3) were introduced transiently into its petals (Ramsay et al., 2003). bHLH2 is an ortholog of Arabidopsis EGL3 and GL3 (Nuraini et al., 2020); therefore, the g mutant may correspond with the bhlh2^dTmi1 mutant identified in this study. On the other hand, Saunders (1928) also defined the R and C genetic loci as being responsible for anthocyanin accumulation in M. incana. Both R and C have been also known to regulate the leaf trichome phenotype (Saunders, 1928). Trichome formation on leaf surfaces is also regulated by the MBW complex in Arabidopsis (Ramsay and Glover, 2005). Therefore, we assumed that R and C may correspond with bHLH2 and WDR1. Furthermore, we posited that ans-1 is a new locus related to white flower coloration in M. incana.

‘Kiss me White’ and ‘Pygmy White’ are bhlh2^dTmi1 and ans-1 double mutants (Figs. 3 and 5). This raises the question of whether these two cultivars need both mutations or a single mutation for white flower coloration. In a collection of isogenic M. incana lines, three white flower lines (e, f, and g mutants) were identified, with the e (WDR1) and g (bHLH2) mutants exhibiting the glabrous phenotype (Seyffert, 1971). The ‘Kiss me’ and ‘Pygmy’ series are potted cultivars that exhibit the glabrous leaf phenotype (Table 1). Therefore, we speculated that bhlh2^dTmi1 contributes to the glabrous leaf phenotype in these cultivars irrespective of anthocyanin production. We additionally investigated the presence of mutant bHLH2 and ANS alleles in other cultivars of the ‘Kiss me’ series that produce anthocyanin or carotenoid pigments and have glabrous leaves (Figs. 3D and 5C). Among cultivars of the ‘Kiss me’ series, bhlh2^dTmi1 alleles were detected in two anthocyanin-less cultivars, ‘Kiss me Yellow’ and ‘Kiss me White’ (Fig. 3D). On the other hand, the ans-1 allele was detected in only ‘Kiss me White’ (Fig. 5C). The expression profiles of anthocyanin structure and regulatory genes in ‘Kiss me Yellow’ were similar to those in ‘Kiss me White’ (Figs. 2 and 4). Although all cultivars of the ‘Kiss me’ series have glabrous leaf phenotypes, this phenotype is not correlated with the presence of bhlh2^dTmi1 (Fig. 3D). Moreover, the petals of ‘Kiss me Yellow’ accumulated carotenoid, but not anthocyanin, pigments. This is correlated with a single mutation designated as bhlh2^dTmi1 (Figs. 3D and S3). Therefore, we suggest that either bhlh2^dTmi1 or ans-1 alleles are enough to induce deficient anthocyanin production in M. incana flowers.

‘White Wonder No. 2’ is an all-double type cultivar and is utilized as a cut flower. This cultivar exhibits a strong linkage between the double flower and white flower loci (Fujita, 1994). Therefore, we hypothesized that the gene or allele responsible for the white flower coloration is different from that in other white-flowered cultivars. Our results revealed that the white flower coloration in ‘White Wonder No. 2’ is associated with the ans-1 mutation (Fig. 5B). However, Southern blot analysis did not detect any genome rearrangement within the ‘White Wonder No. 2’ ANS locus (Fig. S2). Further investigation into the generation of mostly double flowers in cultivars such as ‘White Wonder No. 2’ is necessary to understand this unique phenomenon.

DNA markers for anthocyanin biosynthesis-related genes have been developed for several plants (Matsubara et al., 2006; Nakajima et al., 2005; Nakatsuka et al., 2012; Nishihara et al., 2015; Shimizu et al., 2011). In the present study, we utilized the bHLH-U881 and bHLH2-L1405 primer set to generate PCR products of distinct lengths that are able to distinguish between the bHLH2 and bhlh2^dTmi1 alleles (Figs. 3A and D). The ans-1 allele is the result of a single nucleotide deletion, and it is difficult to distinguish between ans-1 and ANS based solely on PCR. To distinguish between ans-1 and ANS, we employed PIRA-PCR (also known as derived CAPS, dCAPS), a method that does not require specialized and expensive equipment (Haliassos et al., 1989; Nakatsuka et al., 2012; Neff et al., 1998). Alu I digestion of amplified fragments from ans-1 and ANS resulted in the generation of a fragment that was 30-bp shorter in the latter (Fig. 5B, C). The DNA markers we developed for differentiating between bHLH2/bhlh2^dTmi1 and ANS/ans-1 alleles were then applied to genotypes in other cultivars, including the ‘Kiss me’ series. As a result, we identified the genotype of ‘Kiss me Yellow’ as bhlh2^dTmi1 (Figs. 3D and 5C). Southern blot analysis revealed that bHLH2 and ANS exist as single copies in the M. incana genome (Fig. S2). Therefore, we consider the results of genotyping using our DNA markers to be accurate.

We successfully identified two distinct spontaneous mutated alleles, bhlh2^dTmi1 and ans-1, that are of importance and are involved in anthocyanin-less flower coloration in M. incana commercial cultivars. In addition, we also developed DNA markers that can be used to distinguish between the genotypes of both alleles of ANS and bHLH2. Along with other DNA markers that can distinguish between single- and double-flower individuals (Nakatsuka and Koishi, 2018), these DNA markers will be helpful in future nursery plant management and further breeding of commercial M. incana.

Acknowledgements

We thank Ms. Nobue Nakamura from Shizuoka University for technical support.

Literature Cited

•  Chen,  D.,  Y.  Liu,  Q.  Pan,  F. F.  Li,  Q.  Zhang,  X.  Ge and  Z.  Li. 2018. De novo transcriptome assembly, gene expressions and metabolites for flower color variation of two garden species in Brassicaceae. Sci. Hortic. 240: 592–602.
• Crane, M. B. and M. B. Lawrence. 1947. The genetics of garden plants. Macmillan & Co. Ltd. London.
• Dole, J. M. and H. F. Wilkins. 2005. Matthiola. p. 682–687. In: J. M. Dole and H. F. Wilkins (eds.). Floriculture prinples and species. Pearson Prentice Hall. New Jersey.
•  Dressel,  A. and  V.  Hemleben. 2009. Transparent Testa Glabra 1 (TTG1) and TTG1-like genes in Matthiola incana R. Br. and related Brassicaceae and mutation in the WD-40 motif. Plant. Biol. 11: 204–212.
•  Ecker,  R.,  A.  Barzilay and  E.  Osherenko. 1993. Linkage relationships of genes for leaf morphology and double flowering in Matthiola incana. Euphytica 74: 133–136.
•  Feller,  A.,  K.  Machemer,  E. L.  Braun and  E.  Grotewold. 2011. Evolutionary and comparative analysis of MYB and bHLH plant transcription factors. Plant J. 66: 94–116.
• Fujita, M. 1994. Stock (In Japanese). Seibundo Shinkosya, Tokyo.
•  Haliassos,  A.,  J. C.  Chomel,  S.  Grandjouan,  J.  Kruh,  J. C.  Kaplan and  A.  Kitzis. 1989. Detection of minority point mutations by modified PCR technique: a new approach for a sensitive diagnosis of tumor-progression markers. Nucleic Acids Res. 17: 8093–8099.
•  Hemleben,  V.,  A.  Dressel,  B.  Epping,  R.  Lukacin,  S.  Martens and  M.  Austin. 2004. Characterization and structural features of a chalcone synthase mutation in a white-flowering line of Matthiola incana R. Br. (Brassicaceae). Plant Mol. Biol. 55: 455–465.
• Kunze, R. and C. F. Well. 2002. The hAT and CACTA superfamilies of plant transposons. p. 565–610. In: N. L. Craig, R. Craigie, M. Gellert and A. M. Lambowitz (eds.). Mobile DNA II. ASM press. Washington, D. C.
•  Liu,  Y.,  L.  Wang,  J. L.  Kermicle and  S. R.  Wessler. 1998. Molecular consequences of Ds insertion into and excision from the helix-loop-helix domain of the maize R gene. Genetics 150: 1639–1648.
•  Matsubara,  K.,  S.  Chen,  J.  Lee,  H.  Kodama,  H.  Kokubun,  H.  Watanabe and  T.  Ando. 2006. PCR-based markers for the genotype identification of flavonoid 3',5'-hydroxylase genes governing floral anthocyanin biosynthesis in commercial petunias. Breed. Sci. 56: 389–397.
•  Morita,  Y.,  M.  Saitoh,  A.  Hoshino,  E.  Nitasaka and  S.  Iida. 2006. Isolation of cDNAs for R2R3-MYB, bHLH and WDR transcriptional regulators and identification of c and ca mutations conferring white flowers in the Japanese morning glory. Plant Cell Physiol. 47: 457–470.
•  Nakajima,  T.,  K.  Matsubara,  H.  Kodama,  H.  Kokubun,  H.  Watanabe and  T.  Ando. 2005. Insertion and excision of a transposable element governs the red floral phenotype in commercial petunias. Theor. Appl. Genet. 110: 1038–1043.
•  Nakatsuka,  T. and  K.  Koishi. 2018. Molecular characterization of a double-flower mutation in Matthiola incana. Plant Sci. 268: 39–46.
•  Nakatsuka,  T.,  K. S.  Haruta,  C.  Pitaksutheepong,  Y.  Abe,  Y.  Kakizaki,  K.  Yamamoto,  N.  Shimada,  S.  Yamamura and  M.  Nishihara. 2008. Identification and characterization of R2R3-MYB and bHLH transcription factors regulating anthocyanin biosynthesis in gentian flowers. Plant Cell Physiol. 49: 1818–1829.
•  Nakatsuka,  T.,  M.  Nishihara,  K.  Mishiba and  S.  Yamamura. 2005a. Temporal expression of flavonoid biosynthesis-related genes regulates flower pigmentation in gentian plants. Plant Sci. 168: 1309–1318.
•  Nakatsuka,  T.,  M.  Nishihara,  K.  Mishiba and  S.  Yamamura. 2005b. Two different mutations are involved in the formation of white-flowered gentian plants. Plant Sci. 169: 949–958.
•  Nakatsuka,  T.,  M.  Saito,  Y.  Sato-Ushiku,  E.  Yamada,  T.  Nakasato,  N.  Hoshi,  K.  Fujiwara,  T.  Hikage and  M.  Nishihara. 2012. Development of DNA markers that discriminate between white- and blue-flowers in Japanese gentian plants. Euphytica 184: 335–344.
•  Neff,  M. M.,  J. D.  Neff,  J.  Chory and  A. E.  Pepper. 1998. dCAPS, a simple technique for the genetic analysis of single nucleotide polymorphisms: experimental applications in Arabidopsis thaliana genetics. Plant J. 14: 387–392.
• Nishihara, M., K. Mishiba, T. Imamura, H. Takahashi and T. Nakatsuka. 2015. Molecular breeding of Japanese gentians—Applications of genetic transformation, metabolome analyses, and genetic markers—. p. 239–265. In: J. Rybczyński, M. Davey andA. Mikuła (eds.). The Gentianaceae Volume 2: Biotechnology and Applications. Springer. Berlin, Heidelberg.
•  Nishihara,  M.,  E.  Yamada,  M.  Saito,  K.  Fujita,  H.  Takahashi and  T.  Nakatsuka. 2014. Molecular characterization of mutations in white-flowered torenia plants. BMC Plant Biol. 14: 86.
•  Nuraini,  L.,  Y.  Ando,  K.  Kawai,  F.  Tatsuzawa,  K.  Tanaka,  M.  Ochiai,  K.  Suzuki,  V.  Aragones,  J. A.  Daros and  T.  Nakatsuka. 2020. Anthocyanin regulatory and structural genes associated with violet flower color of Matthiola incana. Planta 251: 61.
•  Ohno,  S.,  M.  Hosokawa,  A.  Hoshino,  Y.  Kitamura,  Y.  Morita,  K. I.  Park,  A.  Nakashima,  A.  Deguchi,  F.  Tatsuzawa,  M.  Doi,  S.  Iida and  S.  Yazawa. 2011. A bHLH transcription factor, DvIVS, is involved in regulation of anthocyanin synthesis in dahlia (Dahlia variabilis). J. Exp. Bot. 62: 5105–5116.
•  Ramsay,  N. A. and  B. J.  Glover. 2005. MYB-bHLH-WD40 protein complex and the evolution of cellular diversity. Trends Plant Sci. 10: 63–70.
•  Ramsay,  N. A.,  A. R.  Walker,  M.  Mooney and  J. C.  Gray. 2003. Two basic-helix-loop-helix genes (MYC-146 and GL3) from Arabidopsis can activate anthocyanin biosynthesis in a white-flowered Matthiola incana mutant. Plant Mol. Biol. 52: 679–688.
•  Saunders,  E. R. 1928. Matthiola. Bibl. Genet. 4: 141–170.
•  Seitz,  C.,  C.  Eder,  B.  Deiml,  S.  Kellner,  S.  Martens and  G.  Forkmann. 2006. Cloning, functional identification and sequence analysis of flavonoid 3'-hydroxylase and flavonoid 3',5'-hydroxylase cDNAs reveals independent evolution of flavonoid 3',5'-hydroxylase in the Asteraceae family. Plant Mol. Biol. 61: 365–381.
•  Seyffert,  W. 1971. Simulation of quantitative characters by genes with biochemically definable action. Theor. Appl. Genet. 41: 285–291.
•  Shimizu,  K.,  N.  Ohnishi,  N.  Morikawa,  A.  Ishigami,  S.  Otake,  I. O.  Rabah,  Y.  Sakata and  F.  Hashimoto. 2011. A 94-bp deletion of anthocyanidin synthase gene in acyanic flower lines of lisianthus [Eustoma grandiflorum (Raf.) Shinn.]. J. Japan. Soc. Hort. Sci. 80: 434–442.
•  Spribille,  R. and  G.  Forkmann. 1981. Genetic control of chalcone synthase activity in flowers of Matthiola incana R. Br. Z. Naturforsch. 36c: 619–624.
•  Suzuki,  K.,  K.  Tasaki and  M.  Yamagishi. 2015. Two distinct spontaneous mutations involved in white flower development in Lilium speciosum. Mol. Breed. 35: 193.
•  Tatsuzawa,  F.,  N.  Okuyama,  K.  Kato,  H.  Shono,  J.  Takeda and  H.  Kofujita. 2014. Sinapoylglucoside and kaempferol glycosides in flowers of Matthiola incana cultivars (Brassicaceae). Hort. Res. (Japan) 13: 85–89 (In Japanese with English abstract).
•  Tatsuzawa,  F.,  N.  Saito,  K.  Toki,  K.  Shinoda and  T.  Honda. 2012. Flower colors and their anthocyanins in Matthiola incana cultivars (Brassicaceae). J. Japan. Soc. Hort. Sci. 81: 91–100.