Introduction

Preharvest sprouting (PHS) is the precocious germination of grains in the spike before harvesting. PHS in wheat induces deterioration of flour quality because starch breakdown occurs in germinated grains. Red-grained wheat varieties are usually more tolerant to PHS than white-grained wheat varieties (Flintham 2000; Warner et al. 2000; Himi et al. 2002). This association between PHS tolerance and red pigmentation of the grain is likely due to a pleiotropic effect of the genes controlling grain color (Flintham 2000; Warner et al. 2000; Himi et al. 2002). The grain color of wheat is controlled by the R-1 (red) genes located on the distal region of the long arm of homoeologous chromosome 3 (3AL, 3BL, and 3DL). White-grained wheat varieties are homozygous for the recessive alleles R-A1a, R-B1a, and R-D1a (r2, r3, and r1, respectively, in a former notation). One or more dominant alleles, R-A1b, R-B1b, and/or R-D1b (R2, R3, and R1, respectively, in a former notation) confer red pigmentation onto grains (McIntosh et al. 1998). We previously demonstrated that the dominant R-1 genes enhance grain dormancy using near isogenic lines carrying red grains (ANK-1A to 1D) with the genetic background of white-grained wheat (Novosibirskaya 67) and a white-grained mutant (EMS-AUS) induced by EMS treatment of a red-grained AUS1490 line (Himi et al. 2002). In spite of the important trait of grain color in wheat breeding, genetic analysis of grain color remains inconclusive because grain color phenotypes were determined by the analysis of F3 grains harvested from individual F2 plants. Since grain color is determined by the maternal genotype, the phenotype of F3 grains from F2 plants is controlled by the F2 genotype. Therefore, the development of genetic markers for R-1 genotypes is required for not only genetic interests but also wheat breeding.

The pigment of red wheat grain is identified to be catechin and PAs (Miyamoto and Everson 1958; McCallum and Walker 1990). In addition, flavonols and stilbenes have been reported to be included in red grain (Matus-Cadiz et al. 2008). In barley grains, PAs consist of the oligomeric and polymeric structures of the flavan-3-ol monomers, (+)-catechin and (+)-gallocatechin, and the most abundant PAs in barley are dimeric PAs (Quinde-Axtell and Baik 2006). This suggests that PAs in wheat might also consist of oligo- and/or polymeric structures. In maize kernels, anthocyanins for purple pigment and phlobaphenes for red pigment have been characterized (Styles and Ceska 1975). Arabidopsis seeds accumulate flavonols and PAs (Routaboul et al. 2006). Anthocyanins, phlobaphenes, flavonols, and PAs are all synthesized through the same early flavonoid biosynthetic pathway and branched out into the individual pathway (Fig. 1). We cloned chalcone synthase (CHS), chalcone flavanone isomerase (CHI), flavonone 3-hydroxylase (F3H), and dihydroflavonol 4-reductase (DFR) genes of wheat and found that these genes were expressed predominantly in the immature grain coat of red grain but were almost completely suppressed in white-grained lines (Himi and Noda 2004; Himi et al. 2005). These results suggest that the R-1 genes of wheat are transcription factors to regulate the flavonoid biosynthetic pathway.

Fig. 1
figure 1

The flavonoid biosynthetic pathway. Enzymes are abbreviated as follows: CHS chalcone synthase, CHI chalcone isomerase, F3H flavanone 3-hydroxylase, DFR dihydroflavonol 4-reductase, ANS anthocyanidin synthase, UFGT UDPG-flavonoid glucosyl transferase, BAN banyuls, LAR leucoanthocyanidin reductase, GST glutathione S-transferase. The pathway is demonstrated according to Winkel-Shirley (2001) and Xie et al. (2004). Wheat genes for the enzymes shown in gray have not yet been cloned

Full size image

Several regulatory proteins involved in flavonoid biosynthesis have been reported in various species, such as maize, petunia, snapdragon, and Arabidopsis (Winkel-Shirley 2001; Mol et al. 1998). Most of these transcription factors belong to two large families, MYB and bHLH, and a small family of WD40 repeats (WDR) (Stracke et al. 2001; Buck and Atchley 2003; Ramsay and Glover 2005). Anthocyanin synthesis requires the synergy of a MYB protein (such as C1/Pl in maize) and a bHLH protein (such as R/B/Lc/Sn in maize). PA synthesis in Arabidopsis seeds requires not only both MYB (TT2) and bHLH (TT8), but also WDR (TTG1). However, TTG1 is expressed in both seeds and vegetative tissues, and TTG protein has other functions, such as the development of trichomes and root hairs and the accumulation of seed mucilage (Walker et al. 1999). Phlobaphene synthesis in maize kernel is regulated by P, a MYB protein, alone (Grotewold et al. 1994). Since each wheat R-1 gene appears to regulate flavonoid biosynthesis and the accumulation of flavonoid pigments in grain coat, we presumed that R-1 genes may encode MYB-type proteins, such as P of maize, and cloned 14 MYB genes from wheat. One of them, Tamyb10s, is located on the distal region of the long arm of chromosomes 3A, 3B, and 3D (Himi and Noda 2005).

In this paper, we demonstrate the variation of sequences of Tamyb10 genes from red- and white-grained wheat varieties, and Tamyb10 genotypes are consistent with the R-1 gene genotypes of the varieties, which were previously clarified from genetic analysis by crossing. The deduced Tamyb10 amino acid sequences have the same motif with a TT2 of Arabidopsis and OsMYB3 of rice. When the Tamyb10-A1 gene of red-grained AUS1490 and the Tamyb10-D1 genes of red-grained Chinese Spring (CS) with the CaMV35S promoter were delivered to a colorless coleoptile of CS, anthocyanin accumulation was observed. Thus, Tamyb10 is likely to be a strong candidate for the R-1 gene of wheat, which regulates wheat grain color. Furthermore, a novel transposon belonging to the hAT family, Genome Surfing Trader (GeST), was found in the Tamyb10-A1 gene of some R-A1a lines. From all of this information, we developed PCR-based markers to detect an R-1 genotype easily at an early stage.

Materials and methods

Plant materials

Wheat (Triticum aestivum) varieties or lines used in this study are listed in Table 1. DH populations used in this study were derived from F1 plants of the cross between red-grained Zenkoji Komugi (R-A1a/R-B1b/R-D1b) and white-grained Tamaizumi (R-A1a/R-B1a/R-D1a) consisting of 142 individuals. They were grown under a semi-transparent plastic roof house at Institute of Plant Science and Resources, Kurashiki, Japan. The spikes were tagged at anthesis and harvested 5 days post-anthesis. Grains were collected from the primary and secondary florets of the central spikelets of the spikes. Leaves and roots were collected from 3-day-old seedlings germinated on plastic plates with 2 pieces of filter paper. For DNA and RNA extraction, developing grains and seedlings grown at 20°C under 12 h of UV light (about 100 mol m−2 s−1, UV lamp) or dark conditions were used.

Table 1 Genotypes of R-1 gene in red-or white-grained lines or varieties
Full size table

Genomic DNA, RNA, and cDNA preparation

DNA was isolated from 1 g of 10-day-old seedlings according to Murray and Thompson (1980). Total RNA was extracted from about 0.5 g of grains and 1 g of leaves and roots of 3-day-old seedlings by the SDS-phenol method (Himi and Noda 2004). Poly (A)+ RNA was isolated from 10 mg of total RNA with an mRNA isolation kit (Roche Diagnostics) according to the instructions from the supplier. cDNA was obtained by the reverse-transcription reaction using SuperScript II (Invitrogen).

Isolation of Tamyb10-A1, Tamyb10-B1, and Tamyb10-D1

All primers used in this study are listed in Table 2. The positions of primers are shown in Suppl Fig. S1. Rapid amplification of cDNA 3′ ends (3′ RACE) was performed to isolate a 3′ region of Tamyb10 using mRNA of CS grains of 5 DPA. The primers of P1LP and P2LP were designed on the basis of the MYB consensus region of the ZmP gene (accession number Z11879). The 3′ region of Tamyb10 cDNA was amplified with P1LP and 3′ adapter primers, and nested PCR was carried out with P2LP and 3′ adapter primers. Cloning of PCR fragments was done with pGEM-T vector systems (Promega). DNA sequences were determined with the PRISM 310 ABI DNA sequencing system (Applied Biosystems).

Table 2 List of primers used in this study
Full size table

Three primers, Tamyb10-LP1, Tamyb10-RP1, and Tamyb10-RP2, were designed for the 3′ RACE product, except for the MYB consensus region. Tamyb10-LP1 and Tamyb10-RP1 primers could recognize and amplify a partial sequence of Tamyb10-A1 from the PCR with genomic DNA of N3A3TD, N3BT3D, and N3DT3A (Suppl Fig. S2a). To obtain Tamyb10-B1 and Tamyb10-D1, PCR was performed using an N3AT3D line with the sets of P1LP and Tamyb10-RP1 and P1LP and Tamyb10-RP2 in low stringency. Two different sequences were obtained, and primers that recognize each sequence were designed (Tamyb10-LP2 and Tamyb10-LP3). These primers were used for 3′ RACE, and each product was sequenced. Tamyb10-RP3 was designed in the latter 3′ RACE products. The primer set of Tamyb10-LP2 and Tamyb10-RP1 could recognize Tamyb10-B1, and the primer set of Tamyb10-LP3 and Tamyb10-RP3 could recognize Tamyb10-D1 using N3AT3D, N3BT3D, and N3DT3A lines (Suppl Fig. S2b, c). The upstream regions of each Tamyb10 gene were isolated by the inverse PCR method. Genomic DNA of CS was treated with PvuII and self-ligated. Then, PCR using Tamyb10-LP4 and Tamyb10-RP4 was performed, and the products were used for nested PCR with 3 sets of primers (Tamyb10-LP1 and Tamyb10-RP4, Tamyb10-LP2 and Tamyb10-RP4, and Tamyb10-LP3 and Tamyb10-RP4). The Tamyb10-LP5 and Tamyb10-RP5 primers, which were designed in the region including the start codon and the stop codon, respectively, were used to obtain full sequences of the Tamyb10 gene.

Expression analysis of the Tamyb10 gene

RT-PCR for examining the level of Tamyb10 gene expression was conducted using Tamyb10-LP4 and Tamyb10-RP1 primers. The primers for CHS, CHI, F3H, DFR, and ANS genes are listed in Table 2. Primers ubi-LP and ubi-RP for the ubiquitin gene and TaActin LP and TaActin RP for the actin gene were used as internal controls.

PCR genotyping for Tamyb10-A1, B1, and D1

The primers and PCR conditions are described in Table 2 and Suppl Fig. S3a–e. Tamyb10-A1 genes are classified into three groups (functional Tamyb10-A1 of Norin 61 type (R-A1b), non-functional Tamyb10-A1 of CS type (R-A1a), and non-functional Tamyb10-A1 of Norin 17 type (R-A1a)). Therefore, three PCR should be required to distinguish each Tamyb10-A gene.

Transient assay

A transient vector, pBI221 (AF502128), which has the beta-glucuronidase gene driven by the CaMV 35S promoter, was used for the transient assay. We inserted Tamyb10-D1 of CS, Tamyb10-A1 of AUS1490, or Tamyb10-A1 of EMS-AUS genes instead of the beta-glucuronidase gene of pBI221 and the designated pBI-myb10D (CS), pBI-myb10A (AUS), or pBI-myb10A (EMS), respectively. These genes were amplified with the Tamyb10-LP5 and Tamyb10-RP5 primers using cDNA of 5 DPA of CS, AUS1490, and EMS-AUS grains, respectively. Grains of CS were imbibed at 25°C for 48 h under dark conditions. pBI-myb10D (CS), pBI-myb10A (AUS), pBI-myb10A (EMS), and pBI221 (control vector) were delivered into CS coleoptiles by particle bombardment according to our previous report (Ahmed et al. 2003). Pigmentation of coleoptiles was observed after 24 h at 25°C under dark conditions.

A transient assay using pBI-myb10D (CS) was also performed with coleoptiles of white-grained varieties (listed in Table 4).

Results

Isolation of Tamyb10 genes

To complete the full sequence of Tamyb10 genes expressed predominantly in immature red grains (Himi and Noda 2005), we then cloned and sequenced the 3′ region of Tamyb10 in cDNA derived from the immature grains of wheat cv. Chinese Spring (CS; R-A1a/R-B1a/R-D1b) by the 3′RACE method. We also amplified the partial sequences of Tamyb10 from the genomic DNA of normal CS and three chromosome addition-deletion lines of CS, Nulli3ATetra3D (N3AT3D), Nulli3BTetra3D (N3BT3D), and Nulli3DTetra3A (N3DT3A), which lacked chromosomes 3A, 3B, and 3D, respectively. Thus, we identified three Tamyb10 genes located on chromosomes 3A, 3B, and 3D (Fig. 2; Suppl Fig. S2a–c). Each fragment included a MYB region and characteristic intron sequences. The 5′ regions of the Tamyb10 genes on chromosomes 3A, 3B, and 3D of CS were obtained by the inverse PCR method. The full genomic sequences of the Tamyb10 genes on chromosomes 3A, 3B, and 3D of CS were then cloned using primers specific to each Tamyb10 gene. We designated the Tamyb10 genes on chromosomes 3A, 3B, and 3D as Tamyb10-A1 (Genbank no. AB191458), Tamyb10-B1 (AB191459), and Tamyb10-D1 (AB191460), respectively.

Fig. 2
figure 2

Structure of Tamyb10 genes and putative Tamyb10 proteins. a Genomic organization of the Tamyb10-1 genes. Boxes indicate exons, and lines indicate introns and untranslated regions. Functional Tamyb10-A1, Tamyb10-B1, and Tamyb10-D1 from the lines of R-A1b, R-B1b, and R-D1b alleles, respectively, have R2 (blue box) and R3 (green box) repeats of the MYB consensus region and conserved sequences (IRTKAL/IRC, red box) among Tamyb10, Arabidopsis TT2, and rice OsMYB3. Tamyb10-A1 gene (CS type) of recessive R-A1a alleles (such as CS) was rearranged in the upper region to an unknown sequence (grey lines and boxes). The hAT family-like 2.2-kb sequence (GeST) was inserted into a second intron of the Tamyb10-A1 gene (Norin 17 type) of the recessive R-A1a alleles (such as Norin 17). Tamyb10-B1 gene of the recessive R-B1a alleles had a 19-bp deletion in the third exon (black arrow). Tamyb10-D1 gene is unidentified. Bar 1 kb. b Functional Tamyb10 encoded by Tamyb10-A1, Tamyb10-B1, and Tamyb10-D1 from the lines of R-A1b, R-B1b, and R-D1b alleles, respectively, has R2 (blue box) and R3 (green box) repeats of the MYB consensus region. The IRTKAL/IRC motif is represented by a red box. Tamyb10-A1 of CS (R-A1a) altered the first half of the R2 repeat (grey box). Tamyb10-A1 of Norin 17 (R-A1a) was truncated to lack the last half of the R3 repeat to C-terminus region. Tamyb10-B1 of recessive R-B1a alleles had a 19-bp deletion in the middle part (black arrow) and was frame-shifted to different amino acid sequences after the deletion (grey box). The Tamyb10-D1 sequence of the R-D1a allele is still unknown. Bar 50 amino acids. c Region of conserved sequences (IRTKAL/IRC) of Tamyb10 (wheat), OsMYB3 (rice), AtTT2 (Arabidopsis), DkMYB2 (persimmon), PtMYB134 (quaking aspen), LjTT2a-c (Lotus japonicas), VvMYBPA2 (grape), SbMYB1 (sorghum), and ZmC1 (maize)

Full size image

Furthermore, to elucidate the relationship between Tamyb10 and dominant R-1 genotype, Tamyb10-A1 of AUS1490 carrying R-A1b and Tamyb10-B1 of Norin 61 carrying R-B1b were isolated (Table 1). These sequences were isolated from both genomic DNA and cDNA (derived from the immature grains).

Thus, Tamyb10-A1, B1, and D1 derived from each dominant allele of R-1 gene have 3 exons and 2 introns (Fig. 2a) and encode 259, 268, and 265 amino acid residues, respectively (Suppl Fig. S4). The deduced amino acid sequences of Tamyb10-A1 showed high identity to those of Tamyb10-B1 and D1 (90 and 91%, respectively), and the deduced amino acid sequence of Tamyb10-B1 also showed high identity to that of Tamyb10-D1 (91%) (Suppl Fig. S4).

A phylogenetic tree was constructed using the predicted amino acid sequences of the R2R3 MYB domains of Tamyb10 genes and flavonoid regulatory R2R3 MYB proteins from other species (Fig. 3). Tamyb10s, OsMYB3, and SbMYB1 are most closely related to the PA-clade 2 rather than to other flavonoid regulatory MYB subgroups, such as PA-clade 1 and other anthocyanin- and phlobaphene regulators. While the functions of OsMYB3 and SbMYB1 have not been identified, these MYB proteins and Tamyb10s seem likely to act as PA regulators.

Fig. 3
figure 3

Phylogenetic analysis showing plant MYB transcription factors. The tree was constructed from the ClustalW alignment using the neighbor-joining method. The scale bar represents 0.1 substitutions per site. The GenBank accession numbers of the MYB proteins are as follows: Tamyb10-A1, B1, and D1 (wheat, AB599721, AB599722, and AB191460), OsMYB3 (rice, BAA23339), SbMYB1 (sorghum, ADD18214), AtPAP1 (Arabidopsis, AAG42001), AtPAP2 (Arabidopsis, AAG42002), DkMYB4 (persimmon, BAI49721), VvMYBPA1 (grape, CAJ90831), ZmP (maize, AAC49394), ZmC1 (maize, P10290), OsC1 (rice, BAD04030), LjTT2a-c (Lotus japonicas, BAG12893-BAG12895), AtTT2 (Arabidopsis, NP_198405), DkMYB2 (persimmon, BAI49719), VvMYBPA2 (grape, ACK56131), and PtMYB134 (quaking aspen, ACR83705)

Full size image

Furthermore, the amino acid sequences encoded by Tamyb10s share a conserved sequence (IRTKAL/IRC) between Arabidopsis TT2 (Nesi et al. 2001) and rice OsMYB3 (Suzuki et al. 1997) in the middle of the sequences (Fig. 2a–c; Suppl Fig. S4). This motif is also found in DkMYB2 (Akagi et al. 2010), PtMYB134 (Mellway et al. 2009), LjTT2s (Yoshida et al. 2008), VvMYBPA2 (Terrier et al. 2009), and SbMYB1 (accession number: GU479928) (Fig. 2c). TT2 is necessary for the expression of late flavonoid biosynthetic genes (PA pathway genes), such as DFR, anthocyanidin synthase (ANS), and banylus (BAN), in immature seeds and for PA accumulation in seed testa (Fig. 1). DkMYB2, PtMYB134, LjTT2s, and VvMYBPA2 were also reported as regulators to accumulate PA. While more than 100 R2R3 MYB genes have been identified in Arabidopsis, no significant similarity was found within the C-terminal regions between TT2 and any other MYB protein (Nesi et al. 2001). On the other hand, DkMYB4 (Akagi et al. 2009) and VvMYBPA1 (Bogs et al. 2007) were also reported as PA regulators, but these proteins do not have this motif and can be classified into another clade (PA-clade 1) rather than the Arabidopsis TT2 group (PA-clade 2) (Akagi et al. 2010).

The N-terminal amino acid sequence encoded by Tamyb10-A1 of CS (R-A1a) was quite different from that of the Tamyb10-A1 genes of AUS1490 (Fig. 2a, b; Suppl Fig. S4). The Tamyb10-A1 protein of CS appears to lack the function to bind DNA because of the loss of the first half of R2 repeat of MYB domain. On the other hand, another R-A1 recessive line, Norin 17 (R-A1a), encodes a similar N-terminal amino acid sequence to Tamyb10-A1 of AUS1490 (R-A1b) but has a 2.2-kb insertion into the second intron (Fig. 2a). This insertion may cause incomplete transcription, since the expression of Tamyb10-A1 in Norin 17 was not detected by RT-PCR (Himi and Noda 2005). While the Tamyb10-B1 of CS has conserved R2R3 repeats, a 19-bp deletion of the CCG repeat region on the downstream of R2R3 repeats caused a frame shift, resulting in different amino acid sequences from that of Tamyb10-B1 of Norin 61 (Fig. 2a, b; Suppl Fig. S4). On the other hand, no Tamyb10-D1 of other R-D1a varieties was isolated using specific primers for Tamyb10-D1 or common primers for Tamyb10s, suggesting that Tamyb10-D1 genes of R-D1a varieties might be deleted.

Linkage between R-1 genotype and Tamyb10 genotype

We investigated Tamyb10 genotypes with 33 wheat varieties with known R-1 genotypes using Tamyb10-A1-, B1-, and D1-specific primers. To detect 3 types (the functional type, non-functional CS type, and non-functional Norin 17 type) of Tamyb10-A1 (Fig. 2a, b), we designed 3 sets of primers for: (1) an upstream region of the functional type, (2) an upstream region of the non-functional-CS type, and (3) a second intron with/without a 2.2-kb insertion (Suppl Fig. S3). All R-A1b varieties (Norin 61, RL4137, Fukuho Komugi, Norin 10, Asakaze Komugi, Norin 66, Fukuwase Komugi, and AUS1490) showed a 665-bp fragment with primer set (1) and a 565-bp fragment with primer set (3) but not with primer set (2) (Fig. 4a, b). These results showed that these lines have the same type of sequences that encode the R2R3 repeat and the IRTKAL/IRC motif. Although three R-A1a varieties (Norin 17, Kitakami Komugi, and AUS1408) also showed the same amplified pattern as R-A1b varieties with primer sets (1) and (2), 2,750-bp fragments were amplified with the primer set (3) (Fig. 4a, b). This result suggested that these three lines lack the lower region of Tamyb10-A1, where the last half of the R3 repeat and the IRTKAL/IRC motif exist because of the 2.2-kb insertion in the second intron (Fig. 2b). Other R-A1a varieties (CS, Zenkoji Komugi, Tamaizumi, Ackarma, SZF, Cornell 595, 8019R1, 8021V2, BL1496, Cadoux, Clark’s cream, Gaines, Hakei 91-64, Prina, and Ryubaku 7) showed amplified fragments with primer sets (2) and (3) but not with primer set (1) (Fig. 4a, b). Therefore, Tamyb10-A1 of these lines lost the function, since the first half of R2 repeat is deleted though IRTKAL/IRC motif exists. Furthermore, Tamyb10-A1 was examined in white-grained EMS-AUS with R-A1a induced by EMS treatment of red-grained AUS1490 with R-A1b (Mares et al. 2005). The PCR amplified patterns were exactly the same as those AUS1490 using primer sets (1), (2), and (3) for Tamyb10-A1. However, we found a single nucleotide substitution of guanine to adenine at position 52, resulting in a single amino acid substitution of glycine (G) to glutamic acid (E) at position 15 (Suppl Fig. S4). These results indicated that the PCR-fragment patterns of Tamyb10-A1 using 3 specific sets of primers and a single nucleotide change of EMS-AUS are completely correlated to the R-A1 genotypes in 33 varieties/lines. Therefore, these sets of primers for Tamyb10-A1 are useful to distinguishing the R-A1 genotype.

Fig. 4
figure 4

Amplification profile of Tamyb10-A1, B1, and D1 genes with sets of Tamyb10-A (1), (2), and (3), respectively, and Tamyb10-B and Tamyb10-D-specific primers with wheat varieties/lines of the known R-1 genotype. a, c PCR patterns of each Tamyb gene. In tables under gel image, the + indicates PCR amplification with primer sets for Tamyb10-A1 (1) and Tamyb10-D1 and PCR amplification of a larger fragment with primer sets for Tamyb10-B1. CS indicates PCR amplification with primer sets for Tamyb10-A1 (2) and N17 indicates PCR amplification of a larger fragment with primer sets for Tamyb10-A1 (3). The wheat varieties/lines used in this study are listed in Table 1, and the varieties/lines written in red letters indicate red-grained varieties/lines. The sequences and positions of the primers and PCR conditions are shown in Table 2 and Suppl Fig. S3, respectively. The sizes (bp) of amplified fragments are shown on the right-hand side. b, e R-1 genotype of each wheat line. Dominant alleles (R-A1b, R-B1b, and R-D1b) are listed as b in red, and recessive alleles (R-A1a, R-B1a, and R-D1a) are listed as a. d grain color of Novosibirskaya 67 (white-grained variety) and its near isogenic lines, ANK-1A to 1E (red-grained lines)

Full size image

The linkage between the Tamyb10-B1 and R-B1 genotypes was also examined using presence/absence detection primers for 19-bp deletion in exon 2 of Tamyb10-B1 (Suppl Fig. S3). 282-bp fragments were amplified in all R-B1b varieties (Norin 61, RL4137, Fukuho Komugi, Kitakami Komugi, and Zenkoji Komugi). On the other hand, a 263-bp fragment (19-bp deleted fragments) was found in all R-B1a varieties (Norin 10, Asakaze Komugi, Norin 66, Fukuwase Komugi, AUS1490, EMS-AUS, Norin 17, AUS1408, CS, Tamaizumi, Ackarma SZF, Cornell 595, 8019R1, 8021V2, BL1496, Cadoux, Clark’s cream, Gaines, Hakei 91-64, Prina, and Ryubaku 7) (Fig. 4a, b).

Furthermore, primers for Tamyb10-D1 derived from CS amplified 1,353-bp fragment in all R-D1b varieties (Norin 61, RL4137, Norin 10, Asakaze Komigi, Kitakami Komugi, CS, and Zenkoji Komugi) but no amplification was found in any of the R-D1a varieties (Fukuho Komugi, Norin 66, Fukuwase Komugi, AUS1490, EMS-AUS, Norin 17, AUS1408, Tamaizumi, Ackarma SZF, Cornell 595, 8019R1, 8021V2, BL1496, Cadoux, Clark’s cream, Gaines, Hakei 91-64, Prina, and Ryubaku 7) (Fig. 4a, b).

These correlations between the R-1 genotypes and PCR-fragment patterns of Tamyb10 were also confirmed in white-grained Novosibirskaya 67 and its near isogenic lines, ANK-1A to 1E, carrying only one allele of R-1b each. Whereas Novosibirskaya 67 has the non-functional CS type of Tamyb10-A1, 19-bp deleted Tamyb10-B1, and non-amplified Tamyb10-D1, introduction of one of each dominant allele of R-1b into Novosibirskaya 67 was coincident with the introduction of each functional type of Tamyb10, as shown in Fig. 4c–e.

PCR genotyping of 142 DH lines produced from the cross between Zenkoji Komugi (R-A1a/R-B1b/R-D1b) and Tamaizumi (R-A1a/R-B1a/R-D1a) was conducted. All 36 white-grained lines showed 19-bp deleted amplification (263 bp) with a set of primers for Tamyb10-B1 and no amplification with a set of primers for Tamyb10-D1 (Table 3). On the other hand, red-grained 106 lines were grouped into 3 types: (1) 282-bp amplification with Tamyb10-B1 primers and no amplification with Tamyb10-D1 primers (31 lines), (2) 263-bp amplification with Tamyb10-B1 primers and 1,353-bp amplification with Tamyb10-D1 primers (36 lines), and (3) 282-bp amplification with Tamyb10-B1 primers and 1,353-bp amplification with Tamyb10-D1 primers (39 lines). Although none of the R-1 genotypes of any of the red-grained lines could be determined, it was clear that at least one functional type of Tamyb10 produced red color in the grain. These results showed that the genome-specific primer sets could be useful to detect each allele of the R-1 gene in wheat breeding programs.

Table 3 Cosegregation of genotype of Tamyb10-B1 and D1, and grain color in doubled haploid (DH) lines derived from the cross between Zenkoji Komugi and Tamaizumi
Full size table

An inserted 2.2-kb sequence, Genome Surfing Trader (GeST), into the second intron in Tamyb10-A1 is a transposon-like sequence belonging to the hAT family

The inserted 2.2-kb sequences and the insertion site into the second intron of Tamyb10-A1 in Norin 17, AUS1408, and Kitakami Komugi are identical. The inserted sequence was revealed to have 17-bp terminal inverted repeats (TIRs) on both ends and to be flanked by 8-bp target site duplications (TSDs) (Fig. 5a, b). Among the transposases of hAT, BED zinc finger motif near N-terminus and hAT dimerization motif at C-terminus are highly conserved (Huang et al. 2009) (Fig. 5c). Furthermore, a BLASTX search revealed that the 2.2-kb sequence possesses a putative BED zinc finger motif at N-terminus that the hAT family transposons commonly carry (Fig. 5b). These characteristics suggested that the inserted sequence likely belongs to the hAT-family transposon (Kempken and Windhofer 2001). However, the putative amino acid sequence of the 2.2-kb sequence is short, and hAT dimerization motif, which is highly conserved in the transposase of the hAT family transposons, was not detected in the sequence (Fig. 5b). From these results, the 2.2-kb sequence, named Genome Surfing Trader (GeST), may be changed from the original active autonomous form to the 2.2-kb sequence as a non-autonomous form.

Fig. 5
figure 5

Structures of GeST. a Sequence of the GeST insertion site. The sequences of target site duplication (TSD) and terminal inverted repeat (TIR) are underlined. b A structure of GeST. A region with similarity to the zinc finger-BED motif (zf-BED) is marked with a black box. Putative initiation and termination sites of coding region are shown with single and double asterisks, respectively. The black triangles at both ends represent TIRs. Bar 0.5 kb. c Structures of putative transposase proteins of hAT family in plant. The black boxes and grey boxes represent zf-BED and hAT dimerization motifs, respectively. The GenBank accession numbers of the proteins are as follows: Ac (maize, P08770), Tip100 (common morning glory, Q9ZWT4), Tam (snapdragon, CAA38906), DaiZ (rice, ACN38703), and Dart (rice, BAI39457)

Full size image

Since GeSTs of Norin 17, AUS1408, and Kitakami Komugi are the same and are inserted into the same position (Fig. 5a), a progenitor of these varieties is expected to carry the same GeST in the same insertion site. AUS1408 was originated from the Transvaal region of South Africa (Mares et al. 2009). Norin 17 and Kitakami Komugi were bred in Japan and registered to the Ministry of Agriculture, Forestry, and Fisheries (MAFF) in 1936 and 1959, respectively (The NIAS Genebank, http://www.gene.affrc.go.jp/index_en.php).

Since the highly strong linkage between the R-1 genotype and Tamyb10 genotype is revealed, the R-1 genotypes of progenitors were estimated through PCR amplification patterns using Tamyb10-detection-specific primer sets (Fig. 6a, b). Although some lines were unavailable from the NIAS Genebank, we were able to estimate the R-1 genotype(s) from the parents and descendants. For example, the R-1 genotype of F5-31 line was estimated to be R-A1b/R-B1b/R-D1b (R1R2R3 in a former style) because both Shiro Daruma and Velvet as its parents carry all dominant alleles of R-1. The R-1 genotype of Kitakanto 5 was also estimated to be R-A1a (Norin 17 type) /R-B1b (r2R3) since both parents have R-B1b (R3), and its progeny, Tohoku 71, has R-A1a (Norin 17 type) (r2) and Norin 10 as the parent of Tohoku 71 has R-A1b/R-B1a/R-D1b (R1R2r3). From these results, Goshu 13, as the common progenitor of Kitakami Komugi and Norin 17, was presumably the origin of GeST. The R-1 genotype of Goshu 13 was not surveyed because the seeds are not preserved in the NIAS Genebank. Interestingly, the R-1 genotype of Goshu Komugi, which is available from NIAS, was R-A1a (Norin 17 type)/R-B1a/R-D1a, and “Goshu” is synonymous with “Australia”. Although there is no evidence to demonstrate the relation between the progenitor of Goshu 13/Goshu Komugi and AUS1408, these varieties possibly have the same progenitor, which has the GeST insertion.

Fig. 6
figure 6

Pedigree charts of Kitakami Komugi and Norin 17. R1/r1, R2/r2, and R3/r3 are synonymous with R-D1b/R-D1a, R-A1b/R-A1a, and R-B1b/R-B1a, respectively. Tamyb10-A1 genes of CS type are written as “r2 (CS)”, and those of the Norin 17 type are written as “r2 (No17)” in bold font. Estimated R-1 gene genotypes are shown within brackets. The progenitors of Kitakami Komugi (a) and Norin 17 (b) were cited from Yamada (1990). Wheat varieties with asterisks were distributed from a genetic resource of the NIAS (National Institute of Agrobiological Sciences)

Full size image

Activation of flavonoid biosynthetic genes by Tamyb10

Since the red pigment of wheat grain was identified as catechin and PAs (Miyamoto and Everson 1958; McCallum and Walker 1990), biosynthesis of these pigments is expected to be regulated by the transcription factor of flavonoid biosynthetic pathway. Thus, it is important to examine whether Tamyb10 gene carrying R2R3 MYB domain can regulate the biosynthesis. To investigate the ability of Tamyb10 to activate flavonoid biosynthetic genes, a transient assay system using white coleoptiles previously developed (Ahmed et al. 2006) was applied. We constructed a plasmid, pBI-myb10D, with the Tamyb10-D1 gene of CS driven by the CaMV35S promoter and introduced it into white coleoptiles of CS by particle bombardment. Figure 7a shows an obvious red pigmentation in the coleoptiles 24 h after bombardment of Tamyb10-D1 of CS. The red pigment appeared to be anthocyanin from the absorbance spectra with acidic methanol extract of the pigmented coleoptiles (data not shown). On the other hand, the delivered empty vector (pBI221) did not induce red pigmentation (Fig. 7b). We also examined the expression of flavonoid biosynthetic genes, including CHS, CHI, F3H, DFR, and ANS, in Tamyb10-D1-induced red-pigmented coleoptiles. The expressions of CHI, F3H, DFR, and ANS were higher in pigmented coleoptiles than in white coleoptiles, which harbor a control vector (Fig. 7c). It is noteworthy that a particularly high expression of the DFR gene was found in pigmented coleoptiles but not in white ones. These results suggest that the product of Tamyb10-D1 of CS can induce the expression of flavonoid biosynthetic genes and regulate anthocyanin synthesis.

Fig. 7
figure 7

Pigmentation of the coleoptile by delivering the plasmid, pBI-myb10, that introduced the Tamyb10-D1 gene of CS under the control of the CaMV 35 promoter. a Coleoptile 24 h after bombardment with pBI-myb10. b Coleoptile 24 h after bombardment with a plasmid pBI221. Bar 1 mm. c Expression of flavonoid biosynthetic genes in coleoptiles with Tamyb10-D1 of CS (+) or pBI-221 (−) bombardment. RNA samples of coleoptiles were isolated 24 h after bombardment and used for RT-PCR

Full size image

The function of Tamyb10-A1 genes of red-grained AUS1490 and its mutant, white-grained EMS-AUS, was also analyzed in our transient system (Fig. 8a). Both Tamyb10-A1 genes were expressed in immature grain (DPA 5). Whereas Tamyb10-A1 from AUS1490 made the CS coleoptiles pigmented through a transient assay, that from EMS-AUS did not induce anthocyanin pigmentation (Fig. 8b, c). As reported before, guanine at position 52 of AUS1490 was replaced into adenine in EMS-AUS, resulting in a single amino acid substitution, glycine to glutamic acid, at position 15 in EMS-AUS (Fig. 8d). Glycine at position 15 was extremely conserved among functional MYB proteins, including TT2 (PA regulator) (Nesi et al. 2001), PAP1, PAP2 (anthocyanin regulator) (Borevitz et al. 2000), AtMYB4 (UV-protecting sunscreens regulator) (Jin et al. 2000), AtMYB34 (ATR1, tryptophan pathway regulator) (Bender and Fink 1998), GL1 (AtMYB0, regulator of trichome development) (Oppenheimer et al. 1991), AtMYB66 (WER, regulator of epidermal-cell patterning) (Lee and Schiefelbein 1999), and AtMYB2 (regulator of salt- and dehydration-responsive genes) (Urao et al. 1993) (Fig. 8d). In addition, the glycine site was conserved among 96% of 125 R2R3-MYB proteins of Arabidopsis (Fig. 8d) (Stracke et al. 2001). Furthermore, the predicted alpha-helix structure differs between Tamyb10-A1 of AUS1490 and EMS-AUS, as determined using the New Joint method (Akiyama et al. 1998) (data not shown). These findings suggest that the glycine of red-grained AUS1490 is essential as a functional regulator.

Fig. 8
figure 8

Functional analysis of Tamyb10-A1 of AUS1490 and EMS-AUS. a Grain color of AUS1490 and its mutant, EMS-AUS. b Expression of Tamyb10-A1 in AUS1490 and EMS-AUS. The actin gene is used as the control, showing that the same amount of cDNA was applied in the PCR. c Pigmentation of the 24-hour coleoptile by delivering the plasmid with the Tamyb10-A1 gene of AUS1490 (upper) or EMS-AUS (lower) under control of the CaMV 35 promoter. d R2 repeat of MYB protein sequences. The alteration site from conserved glycine (G), colored in red, to glutamic acid (E) in EMS-AUS is indicated by a black arrow. Repeated W (tryptophan) residues in the R2R3 motif are shown in blue. Other conservative residues are shown in green. The numbers at the bottom are the conservation ratio among 125 R2R3-MYB proteins of Arabidopsis (Stracke et al. 2001)

Full size image

Since Tamyb10-B1 derived from red-grained Norin 61 showed high identity to Tamyb10-A1 and Tamyb10-D1, the product of Tamyb10-B1 of dominant R-B1b alleles presumably has the same active function as that of Tamyb10-D1 of CS or Tamyb10-A1 of AUS1490. Consequently, Tamyb10-A1 and Tamyb10-D1 derived from red-grained AUS1490 and CS, respectively, were demonstrated to activate anthocyanin biosynthetic enzymes, suggesting that Tamyb10 is a strong candidate for the R-1 gene.

Detection of non-functional genes in anthocyanin biosynthetic pathway through a transient assay of Tamyb10 in white coleoptile of white-grained varieties

In Arabidopsis, mutants impaired in flavonoid accumulation in seeds have been identified as transparent testa (tt) mutants (Lepiniec et al. 2006). Characterized loci involved in PA biosynthesis in Arabidopsis seed have been classified into structural genes (such as CHS, CHI, F3H, F3′H, DFR, and ANS) and regulatory genes (encoding regulatory proteins such as MYB, bHLH, MADS, and WD40) (Lepiniec et al. 2006). Mutants of anthocyanidin-less grain (ant mutants) are also known in barley, and one of the ant loci, ant 18, was reported as DFR gene (Kristiansen and Rohde 1991). However, no mutants of structural genes involved in flavonoid biosynthesis in wheat have been reported. It is still unknown whether mutation(s) occurred in structural gene(s) in white-grained wheat.

We performed a transient assay with pBI-myb10D (CS) into white coleoptiles of 11 white-grained lines to survey malformed structural genes of these lines. As shown in Table 4, red pigmentation was observed in coleoptiles of CS and other 10 white-grained lines, but no pigmentation was found in the coleoptile of Clark’s cream by introducing the Tamyb10 gene alone. On the other hand, delivering both Tamyb10 and B-peru, the bHLH protein of maize induced red pigmentation in the coleoptile (Table 4). These results suggest that all structural genes involved in anthocyanin accumulation in these white-grained lines act normally. However, it was demonstrated that anthocyanin accumulation in coleoptile requires not only a MYB gene, such as Tamyb10 or C1, but also unknown bHLH proteins, such as B-peru, which may be inactive in Clark’s cream coleoptiles.

Table 4 Anthocyanin pigmentation after bombardment of Tamyb10-D1 of CS, or Tamyb10-D1 of CS together with B- peru of maize
Full size table

Discussion

The first report of the inheritance of red grain color of hexaploid wheat was published in 1905 (Biffen 1905). Subsequent reports verified the existence of the R-1 gene for the redness of grain on the chromosome of the homoeologous group 3 and suggested that the R-1 gene alone controls the grain color, as maize P (ZmP) alone regulates red pigment accumulation in the pericarp (Grotewold et al. 1994). ZmP is a member of the R2R3-type MYB transcription factors that control a set of flavonoid biosynthesis genes (CHS, CHI, and DFR) that are independent of bHLH (Grotewold et al. 1994). While we isolated several sequences that have the bHLH motif in wheat, these bHLH genes were expressed in leaves, and no expression of these bHLH genes was found in immature grain (Himi and Noda 2005). Similar results were obtained from the Wheat Genetic Resource Database (http://www.shigen.nig.ac.jp/wheat/komugi/top/top.jsp), suggesting that bHLH proteins are less active than MYB proteins in developing grains.

We isolated the Tamyb10 genes of wheat, which encode R2R3-type MYB transcription factors. These proteins have a unique motif (IRTKAL/IRC) in the C-terminus region, which was also found in TT2 of Arabidopsis, OsMYB3 of rice, and other PA regulators (Fig. 2c; Suppl Fig. S4). The KAxRC sequences in the motif were also seen in the C1 protein, which is also a regulator of the anthocyanin biosynthesis of maize, suggesting that this site might play an essential role as a flavonoid biosynthesis regulator. Although the Tamyb10-B1 gene was expressed in both R-B1a and R-B1b varieties (data not shown), the deduced amino acid sequences encoded by Tamyb10-B1 of CS and other varieties carrying R-B1a, a recessive allele, may not have the function because they lack this conserved motif. We demonstrated the critical difference of sequences between two types of Tamyb10-A1 genes of R-A1a varieties and R-A1b varieties and the Tamyb10-B1 genes of R-B1a varieties and R-B1b varieties. On the other hand, the Tamyb10-D1 sequence of the R-D1a varieties remains to be isolated (Fig. 2a, b). No amplified fragments from genomic DNA and cDNA of R-D1a varieties were obtained by PCR using Tamyb10-D1-specific primers, suggesting that the Tamyb10-D1 sequence of R-D1a might be deleted or that the sequence of the sites where the primers anneal might be varied. However, the presence/absence of Tamyb10-D1 is demonstrated through PCR amplification using Tamyb10-D1-specific primers, although heterozygotes of Tamyb10-D1 cannot be detected by PCR. Consequently, through strong linkage between the R-1 genotype and the Tamyb10-sequence type of 33 varieties carrying known R-1 genotypes and DH plants, Tamyb10-genome-specific primers were found to be efficient for R-1 genotyping.

Previous reports indicated that the dominant R-A1b, R-B1b, and R-D1b loci are functionally equivalent and have a quantitatively similar effect (Flintham 2000). The putative amino acid sequences of Tamyb10-A1, B1, and D1 of the dominant R-A1b, R-B1b, and R-D1b alleles are similar to each other, and each Tamyb10 appears to act as a transcription factor. On the other hand, Tamyb10-A1 of CS, which has a recessive R-A1a allele, was found to lack the first half of the R2 repeat of the MYB region as if it were bound to another gene through a large deletion, including first half of the R2 repeat (Fig. 2a, b). Interestingly, the promoter sequence of the Tamyb10-A1 gene isolated from diploid wheat (Triticum monococcum, 2n = 14; genome formula, AmAm) showed similarity to that of Tamyb10-A1 of Norin 61, which has a dominant R-A1b allele (data not shown). Furthermore, two mutation patterns of Tamyb10-A1 of CS type (lack of first half of the R2 repeat) and Norin 17 type (GeST insertion in second intron) are not found in one Tamyb10-A1 gene simultaneously. Presumably, each mutation could result from a different event in the wheat line of the dominant R-A1b in an evolutionary step. The 19-bp deletion in the Tamyb10-B1 gene of the recessive R-B1a allele could also result from a single event in the wheat line of the dominant R-B1b, as well as R-A1.

Recently, more than three thousand transposable elements (TEs) were identified from wheat chromosome 3B (Choulet et al. 2010). While characterized mutable alleles caused by TEs in the hAT family have been reported in many species (Kempken and Windhofer 2001), few mutable alleles have been reported in wheat even though the wheat genome contains TEs in the hAT family. We indentified a novel transposon, GeST, belonging to the hAT family, in Tamyb10-A1 of Norin 17. However, GeST has no hAT dimerization motif and is locked in the same positions in several varieties or from generation to generation, suggesting that GeST is inactive.

Zimmermann et al. (2004) reported a conserved amino acid signature (D/ELx2R/Kx3Lx6Lx3R) as the structural basis for interaction between MYB and bHLH proteins. Interestingly, this signature was found in Tamyb10 proteins (Suppl Fig. S4) but not in ZmP, an independent transcription factor for phlobaphene. In this report, we show that the introduced Tamyb10 gene induced red anthocyanin pigmentation in white coleoptiles of CS but not in white coleoptiles of Clark’s cream (Table 4). We previously showed that the introduction of C1 (a MYB-type regulatory gene of maize) and B-peru (a bHLH-type regulatory gene of maize) induced anthocyanin accumulation in coleoptiles of CS (Ahmed et al. 2003). From these results, not only a MYB gene such as Tamyb10 or C1 but also unknown bHLH proteins such as B-peru, which may be inactive in Clark’s cream coleoptiles, regulate to activate anthocyanin biosynthesis genes in coleoptiles. Khlestkina et al. (2008) reported that Rc (red coleoptiles) genes located on homoeologous group 7 are responsible for anthocyanin pigmentation, which regulates F3H expression in wheat coleoptiles. The introduction of Tamyb10 also upregulated F3H and other structural genes, suggesting that Rc might be a MYB-type regulatory gene, such as Tamyb10. Tamyb10 was demonstrated to activate anthocyanin biosynthesis genes with synergy of the bHLH-type protein through a transient assay. However, no bHLH-type proteins have been detected in immature grains of red-grained wheat by bHLH-specific PCR or registered in the wheat database. Furthermore, genetic analysis for red grain has indicated that the genome-specific single gene R-1 controlled the red pigmentation of grains. This evidence suggests that Tamyb10 solely activates catechin and PA synthesis genes. From these results, it is likely that Tamyb10 is a strong candidate for the R-1 gene.

An association between grain color and depth of dormancy has been noted in a viviparous 1 (vp1) mutant of maize and in transparent testa (tt) mutants of Arabidopsis (McCarty et al. 1991; Debeaujon et al. 2000). The R-1 genes affect the sensitivity of embryos to abscisic acid (ABA) and the development of grain dormancy, and it has been proposed that one of the Myb-type genes of Arabidopsis, AtMYB2, might be involved in ABA signal transduction (Abe et al. 2003). It remains to be determined whether Tamyb10 is involved in the development of grain dormancy via control of the sensitivity of the embryo to ABA. The gene structure characterization described in this paper may provide important information related to simple and easy genetic markers for wheat grain color and the association between grain color and grain dormancy.