This article is an Advance Online Publication of the authors’ corrected proof. Note that minor changes may be made before final version publication.

The Horticulture Journal Preview e Japanese Society for doi: 10.2503/hortj.UTD-100 JSHS Horticultural Science http://www.jshs.jp/

Production of Novel Red-purple Flowers Containing Cyanidin- based Anthocyanin Using Hybridization Breeding

Kimitoshi Sakaguchi1, Chisato Isobe2, Kazuyoshi Fujita2, Yoshihiro Ozeki3 and Taira Miyahara4*

1Miyoshi Agritech Co., Ltd sales department, Hokuto 408-0041, Japan 2Miyoshi & Co., Ltd., R & D Center, Hokuto 408-0041, Japan 3Department of Biotechnology, Tokyo University of Agriculture and Technology, Koganei 184-8588, Japan 4Graduate School of Horticulture, Chiba University, Chiba 263-8522, Japan

Modern molecular biology techniques have enabled the generation of novel flower colors. Standard cultivated varieties of delphinium have blue flowers as a result of the biosynthesis and accumulation of delphinidin-based anthocyanins. Some cultivars have pink flowers due to the biosynthesis and accumulation of pelargonidin- based anthocyanins. The biosynthetic pathway of the latter becomes active due to the inactivation of flavonoid 3',5'-hydroxylase. Cyanidin-based red-purple flowers have not been identified to date in because these species do not express the flavonoid 3'-hydroxylase gene. However, in our previous work, we identified expression of the flavonoid 3'-hydroxylase gene in a wild delphinium (Delphinium zalil) that accumulates quercetin 3-glycoside. D. zalil lacks the anthocyanidin synthase, the key enzyme to produce anthocyanins, so the flowers do not contain any anthocyanins. Here, we report the use of conventional breeding to introduce cyanidin biosynthesis into delphiniums. We introduced the flavonoid 3'-hydroxylase gene of D. zalil into D. cardinale by hybridization breeding, causing accumulation of cyanidin-based anthocyanin. In the hybrid , flavonoid 3'-hydroxylase was transcribed and a cyanidin-based anthocyanin was biosynthesized, generating novel purple-red flowers. Greater understanding of the anthocyanin biosynthetic genes expressed in wild species will benefit the development of breeding strategies to generate novel flower colors in cultivars of high horticultural value.

Key Words: Delphinium cardinale, Delphinium zalil, flavonoid 3'-hydroxylase, flower color, pelargonidin.

erate the molecular breeding process to produce diverse Introduction flower colors and shapes, disease resistance, extended Recent advances in molecular biology have acceler‐ vase life, and other traits (Matsubara et al., 2006; ated the hybridization process by reducing the Nakatsuka and Koishi, 2018; Nakatsuka et al., 2011; time and space required for breeding. Next-generation Nishihara et al., 2018; Yagi, 2013, 2018). sequencing (NGS) platforms now provide one of the Conventional cross-breeding programs make use of most potent tools in molecular biology and whole- variants that have been developed spontaneously or in‐ genome sequencing of many horticultural crops has duced artificially after treatment with mutagens such as been undertaken (Hirakawa et al., 2014; Hoshino et al., gamma rays and heavy-ion beams. However, this ap‐ 2016; Yagi et al., 2014). In particular, NGS is capable proach requires the cultivation of a large number of of providing genetic markers for ornamental plant seedlings and the availability of large fields for plant‐ breeding, and marker-assisted selection can then accel‐ ing, growth, and improvement of novel cultivars (Okamura et al., 2013; Yamaguchi et al., 2008, 2009, 2010). Breeders must then wait until the plants bloom Received; April 17, 2019. Accepted; June 4, 2019. to identify novel flower-color traits. Therefore, the First Published Online in J-STAGE on July 17, 2019. process of defining a novel mutant and establishing a This study was supported by JSPS KAKENHI Grant Number 16K18564 and 19K15829 to TM and 18K06279 to YO. new cultivar for commercial production has many spa‐ * Corresponding author (E-mail: [email protected]). tial and temporal requirements. Further, the success of

© 2019 The Japanese Society for Horticultural Science (JSHS), All rights reserved. 2 K. Sakaguchi, C. Isobe, K. Fujita, Y. Ozeki and T. Miyahara such an approach depends on the experience of the three hydroxyl residues at the 3',4', and 5' positions breeder (Anderson, 2006; Onozaki et al., 2018; Shibata, (Davies, 2009; Tanaka and Brugliera, 2013). Studies of 2008). However, the application of molecular biological anthocyanin structure in delphinium have shown that methods can overcome these difficulties. For example, blue, blue-violet, and pink are generated by the obtaining gene expression information by RT-PCR in aglycones delphinidin and pelargonidin (Hashimoto potential parental lines before hybridization can provide et al., 2000, 2002; Honda et al., 1999; Kondo et al., vital information regarding the genetic backgrounds of 1990, 1991; Miyagawa et al., 2015). Based on the re‐ the prospective parents and enable breeding plans to be sults of anthocyanin structural analysis, molecular and developed that will achieve the desired objective. As a biochemical analyses of the flavonoid and anthocyanin result, the use of RT-PCR can eliminate laborious and biosynthesis pathways in delphinium have confirmed ineffectual crossing processes and reduce the time and these biosynthetic enzymes produce the anthocyanin space required for breeding. structures. These analyses also elucidated characteris‐ The genus Delphinium () comprises tics of the anthocyanin molecules associated with par‐ over 400 species (http://www.theplantlist.org/browse/A/ ticular flower color traits (Fig. 1; Ishii et al., 2017; Ranunculaceae/Delphinium/). Only a few species of Matsuba et al., 2010; Miyagawa et al., 2014, 2015; Delphinium, such as D. elatum L., D. grandiflorum L., Miyahara et al., 2016; Nishizaki et al., 2013, 2014). and Delphinium × belladonna (a hybrid between However, delphinium cultivars do not bear cyanidin- D. elatum × D. grandiflorum ex Bergmans) are grown based red-purple flowers due to the absence of flavo‐ worldwide. In Japan, approximately 30 million flowers noid 3'-hydroxylase (F3'H), which is required to are produced annually and they are used as cut flowers. generate cyanidin derivatives; this enzyme catalyzes Considerable effort has been devoted to expanding the hydroxylation at the 3' position of the B ring (Fig. 1). variety of flower colors available, resulting in the gen‐ Our previous study showed that D. zalil Aitch. & eration of blue, light blue, violet, purple, lavender, Hemsi. has F3'H activity, but defective anthocyanidin white, and pink flowers (Hashimoto et al., 2002; Katoh synthase (ANS) expression. Therefore, D. zalil does not et al., 2004; Legro, 1961; Miyagawa et al., 2014). produce anthocyanin, but rather accumulates flavonol The anthocyanin pigments that produce flower colors glycosides in its sepals. The recombinant D. zalil F3'H vary in terms of the hydroxylation pattern of the of an‐ enzyme protein expressed in yeast showed hydroxyla‐ thocyanidin B ring (anthocyanidin is an aglycone of an‐ tion activity to convert naringenin, apigenin, dihydro‐ thocyanin). For example, pelargonidin has a hydroxyl kaempferol, and kaempferol to eriodictyol, luteolin, residue at the 4' position, cyanidin has two hydroxyl dihydroquercetin, and quercetin, respectively. Although residues at the 3' and 4' positions, and delphinidin has kaempferol and quercetin glycosides are accumulated in

Phenylalanine

Quercetin Kaempferol F3'H FLS

Dihydroquercetin Dihydrokaempferol Dihydromyricetin F3'H F3'5'H DFR DFR DFR

D. zalil ANS ANS ANS

Cyanidin Pelargonidin Delphinidin

No delphiniums

D. cardinale D. grandiflorum Fig. 1. Schematic pathways of anthocyanin and flavonol biosynthesis in delphiniums. Hort. J. Preview 3

D. zalil sepals, the recombinant enzyme activity showed Analysis of anthocyanin aglycones in the sepals of F1 a preference for dihydrokaempferol over other flavo‐ hybrids noids (Fig. 1; Miyahara et al., 2016). In addition, we The extracts in 80% methanol/0.1% TFA were have also shown that D. cardinale Hook. accumulates allowed to dry, the residue was dissolved in 100 μL of large amounts of a single pelargonidin-based anthocya‐ water. Then, 100 μL of 12 N HCl was added, and hy‐ nin, which leads to the production of vivid red flowers drolysis was performed at 80°C for 1 h. The aglycones (Miyagawa et al., 2015). However, neither flavonoid 3', in the crude hydrolysis solution were extracted by the 5'-hydroxylase (F3'5'H) activity nor accumulation of addition of 200 μL of ethyl acetate. The organic layer F3'5'H reaction products has been detected in wild-type was recovered and dried, then dissolved in 50 μL of D. zalil or D. cardinale. 0.1% TFA, and a 10 μL of an aliquot of this solution, In this study, we introduced the F3'H gene from containing the hydrolyzed aglycones, was analyzed D. zalil into D. cardinale with the expectation that this using HPLC. The equipment used was the same as de‐ would enable generation of a new cultivar that would scribed above, and the elution conditions were identical bear red-purple flowers as a result of cyanidin biosyn‐ to those described by Miyagawa et al. (2014). The agly‐ thesis. cones delphinidin, cyanidin, and pelargonidin were pur‐ chased from Extrasynthese Co., Genay, France, for use Materials and Methods as standards. Plant materials and hybridization of D. zalil and D. cardinale Genomic PCR and RT-PCR analysis Seedlings of D. zalil and D. cardinale were obtained Total RNA was extracted from fully-opened sepals of from Miyoshi & Co., Ltd. and cultivated in a green‐ D. zalil, D. cardinale, and F1 hybrids, and first strand house at the Research and Development Center of cDNAs were synthesized from 1 μg of total RNA, the Miyoshi & Co., Ltd. Immature anthers were removed same as described previously (Ishii et al., 2017). The from the maternal parent (D. zalil) before the flower first strand cDNA reaction mixture was diluted 3-fold, opened. After ripening, the pistils were hand-pollinated then 1 μL of the cDNA mixture was used in PCR as a with freshly collected mature pollen from D. cardinale. template. Full-length cDNA corresponding to the F3'H Then, ~40 days after pollination, the fruit capsules were gene was amplified by PCR using the primers F3'H-F harvested and the seeds inside were collected to devel‐ (5'-ATGCCTTCTCTATACTTTCTA-3') and F3'H-R (5'- op breeding lines of D. zalil × D. cardinale. We used CTATACTTGATAGACACTTGG-3'). The reference D. zalil and D. grandiflorum to confirm the genomic gene Actin was amplified using the primer sets de‐ flavonoid 3'-hydroxylase (F3'H) nucleotide sequences. scribed previously (Miyahara et al., 2016). PCR was performed using PrimeStar GXL DNA polymerase Flower color assessment (Takara Bio Inc., Shiga, Japan) under the following

The sepal colors of D. zalil, D. cardinale, and the F1 conditions: denaturation for 2 min at 94°C, followed by hybrid lines of D. zalil × D. cardinale were evaluated 30 cycles of 5 s at 98°C, 15 s at 55°C, and 20 s at 68°C. using the Royal Horticultural Society Colour Chart The amplified PCR products were then analyzed by (RHS CC; 5th edition, 2007). agarose gel electrophoresis. Genomic sequence analysis was also performed as described previously (Ishii et al., Anthocyanin profiles of D. zalil × D. cardinale hybrid 2017). F3'H genomic sequences in D. zalil and lines D. grandiflorum were confirmed using the same primer Approximately 200 mg of sepals from fully-opened set used for RT-PCR. The genomic sequence of F3'H flowers were frozen in liquid nitrogen and then ground from D. grandiflorum was registered under accession into powder using a mortar and pestle. Subsequently, number LC441150 in the GenBank/EMBL/DDBJ data‐ the powder was mixed in 100 μL of 80% methanol con‐ bases. taining 0.1% trifluoroacetic acid (TFA). Cell debris was Results and Discussion removed by centrifugation at 15,000 × g for 10 min and the supernatant was analyzed by high performance liq‐ Production of D. zalil and D. cardinale F1 hybrids uid chromatography (HPLC) using an LaChrom Elite Initially, we attempted to use D. cardinale as the seed HPLC System (pump L-2130, column oven L-2300 at parent and D. zalil as the pollen parent; however, pollen 30°C, diode array detector L-2450 [Hitachi High- development in D. zalil flowers failed because of defi‐ Technologies, Tokyo, Japan] with a Handy octadecylsil‐ cient stamen development under the prevailing cultiva‐ yl column [internal diameter 4.6 mm × length 250 mm, tion conditions. Therefore, we chose to use D. zalil as Wako Pure Chemical Industries, Osaka, Japan]). The the seed parent and D. cardinale as the pollen parent. injected sample was separated by linear gradient elution Legro (1961) reported that seeds could be obtained by at a flow rate of 1 mL·min−1 in 20%–80% methanol/ this type of cross and that over 80% of the seeds are 1.5% aqueous phosphoric acid for 20 min and detected capable of germination. We collected more than 1,000 at 520 nm. seeds from cross-pollinated D. zalil flowers, sowed 765 4 K. Sakaguchi, C. Isobe, K. Fujita, Y. Ozeki and T. Miyahara in a 406-hole tray, and achieved a germination rate of Anthocyanin composition of D. zalil × D. cardinale F1 43.5%. All the resulting seedlings were planted, of hybrids which 126 developed to the flowering stage. All the Next, we performed HPLC analysis to elucidate the hybrid plants had a single-flower structure, as in the composition of the anthocyanin molecules in the sepals parental plants, but the hybrid flowers displayed of F1 hybrids (Fig. 3). All the anthocyanin retention characteristics that were intermediate in comparison to times were consistent in each F1 hybrid. Three of the the parental types. The growth period from sowing to four breeding lines had one major anthocyanin mole‐ flowering was shorter in the hybrid plants than in the cule and four or more minor anthocyanin molecules. parental varieties, and all the hybrid plants had red- The exception was line No. 3, which was assessed as purple sepals (Fig. 2). 72B using the RHS CC assessment, but had a different We selected four representative hybrids (numbered major anthocyanin molecule. This major anthocyanin 1–4), assessed their sepal colors against the RHS CC, molecule was the second largest peak in other lines. Al‐ and compared the colors with those of the parental though the detailed chemical structures of the anthocya‐ D. zalil and D. cardinale (Table 1). Miyahara et al. nins in the hybrids’ sepals were not identified, the (2016) reported that D. zalil could not biosynthesize backbone structure of these anthocyanins, aglycone, anthocyanin, and its RHS CC value was in the green- was analyzed after hydrolysis of the anthocyanins yellow group (1C). Delphinium cardinale accumulates (Fig. 4). The major anthocyanidin molecule was cyani‐ pelargonidin glycosides (Miyagawa et al., 2015), and its din in these hybrids, while pelargonidin, the major agly‐ RHS CC value is in the red group (42B). The sepal col‐ cone in the parental line D. cardinale (Miyagawa et al., ors of the four hybrids fell into the red-purple group 2015), was only present in small quantities. Miyagawa (72A–72D), and the purple group (N78A; Table 1). The et al. (2015) reported that D. cardinale accumulates sepal colors in the hybrids showed little individual var‐ pelargonidin 3-(6-malonyl)-glucoside-7-(6-(4-(6-(p- iation, but no plant demonstrated a red value (42B) sec‐ hydroxybenzoyl)-glucosyl)-oxybenzoyl)-glucoside) (Sup‐ ondary to the accumulation of pelargonidin derivatives. plemental Fig. 1). We speculate that F1 hybrids may accumulate cyanidin-type anthocyanin molecules with the same modification as in D. cardinale [cyanidin 3- (6‑malonyl)-glucoside-7-(6-(4-(6-(p-hydroxybenzoyl)- glucosyl)-oxybenzoyl)-glucoside)], and that the other D. zalil D. cardinale minor anthocyanin molecules are perhaps precursors of the major anthocyanin molecules. A previous study showed that the flowers resulting from hybridization breeding of D. cardinale × D. grandiflorum (which ac‐ cumulates delphinidin-based anthocyanin) and D. nudicaule Torrey & A. Gray (which accumulates pelargonidin-based anthocyanin) × D. grandiflorum ac‐ F1 hybrids

400 No.1 ) U 200

No.1 No.2 No.3 No.4 mA ( Fig. 2. Flower phenotypes of the parental lines (D. zalil and 0

nm 160 No.2 D. cardinale) and representative F1 hybrids. Table 1. Flower colors of the parental lines D. zalil and D. cardinale, and the F1 hybrid lines 1–4.

520 80

at Table 1. Flower colors of the parental lines D. zalil and 0 D. cardinale, and the F1 hybrid lines 1–4. 240 No.3 RHS CC Line 120 Color group Value

absorbance 0

D. zalil green-yellow 1C e 400 No.4 D. cardinale red 42B iv

D. z × D. c 200 Relat F1 hybrids 0 0 10 20 No. 1 red-purple 72A Retention time (min) No. 2 red-purple 72D No. 3 red-purple 72B Fig. 3. HPLC profile demonstrating the anthocyanins in the sepals of individual F hybrids. Traces for four hybrids are shown, No. 4 purple N78A 1 numbered 1–4. Hort. J. Preview 5

A 1 2 B 1 2 3 4 F3'H F3'H Actin Actin

Fig. 5. Qualitative RT-PCR analysis of F3'H expression. A: F3'H gene (upper) and Actin gene (reference, lower) expression in the sepals of the parental lines D. zalil (1) and D. cardinale (2).

B: F3'H and Actin expression in F1 hybrids, numbered 1–4.

D. cardinale, the main anthocyanin has been structural‐ ly identified as a pelargonidin-based anthocyanin. This fact indicates that both DFR and ANS must be ex‐ pressed in D. cardinale. These findings suggest that the F3'H gene introduced into D. cardinale by hybridiza‐ tion with D. zalil is expressed in the hybrids and can catalyze the conversion of dihydrokaempferol to dihy‐ droquercetin as part of the cyanidin biosynthesis path‐ way (Fig. 1). The fact that the ANS gene was defective in D. zalil implies that all the hybrids had obtained ac‐ tive ANS genes from D. cardinale, while the DFR gene could have come from either parental line and become

active in the F1 hybrids. This indicates that the substrate specificities of DFR and ANS have little effect on an‐ thocyanidin structures such as pelargonidin and cyani‐ din in delphiniums. A previous report described the expression of F3'H Fig. 4. Aglycone analysis of the anthocyanins in the sepals of rep‐ in D. zalil, but not in D. grandiflorum. The genomic resentative F1 D. zalil and D. cardinale hybrids. Standard num‐ structure of the F3'H coding region in D. zalil is the bers are 1, delphinidin; 2, cyanidin; 3, pelargonidin. same as that of the cDNA structure; there are no intron‐ ic sequences and it is directly connected to exons (Sup‐ cumulate delphinidin-based anthocyanins, but not plemental Fig. 2). The genomic structure is similar to pelargonidin-based anthocyanins. This led to the con‐ that of F3'5'H in D. nudicaule (Miyagawa et al., 2014). clusion that delphinidin is dominant over pelargonidin However, F3'5'H is not expressed in D. nudicaule. In in delphiniums (Honda et al., 1999). Our data show that D. grandiflorum, the F3'H genomic sequence comprises cyanidin is also dominant over pelargonidin in delphini‐ three exons and two introns. Moreover, each exonic se‐ ums. quence in D. grandiflorum has > 90% identity with that of D. zalil (Supplemental Fig. 2). Therefore, although Expression of anthocyanidin biosynthetic genes in the the F3'H coding region sequence of D. grandiflorum

F1 hybrids has been confirmed, the reason for its lack of expres‐ Delphinium zalil has an active F3'H that catalyzes the sion has not been clarified. We speculate that there is a conversion of the aglycone flavonoids naringenin, api‐ similar reason for the lack of expression of F3'H in both genin, dihydrokaempferol, and kaempferol to eriodic‐ D. cardinale and D. grandiflorum. tyol, luteolin, dihydroquercetin, and quercetin, Legro (1961) reported that hybridization of D. zalil respectively (Fig. 1; Miyahara et al., 2016). Anthocya‐ and D. cardinale produced hybrids with wide color var‐ nidin profiling of the F1 hybrids showed that they all ac‐ iations in their main sepals, including apple blossom- cumulate cyanidin compounds, implying the inheritance white, creamy-green, soft-pink, raspberry-red, and of an active F3'H gene from D. zalil. Qualitative RT- magenta. In contrast, the F1 hybrids produced here dem‐ PCR analysis demonstrated the expression of F3'H in onstrated a comparatively narrow range of red colors.

D. zalil and the F1 hybrids, but it was undetectable in Legro (1961) used D. zalil plants raised from seeds ob‐ D. cardinale sepals (Fig. 5). Our previous studies tained from the Botanical Garden in Geneva, while the showed that ANS gene expression cannot be detected in D. cardinale plants were raised from seeds collected D. zalil, although DFR can be detected and has been se‐ from the Santa Monica Mountain, at the junction of quenced. These results indicate that a defect in ANS Tuna Canyon and Saddle Park, California. At present, gene expression is responsible for the lack of anthocya‐ the genus Delphinium is cultivated worldwide, and nin biosynthesis in D. zalil (Miyahara et al., 2016). In D. zalil and D. cardinale are important species that are 6 K. Sakaguchi, C. Isobe, K. Fujita, Y. Ozeki and T. Miyahara grown and distributed as the commercial brands “Zalil” K. Nanri, A. Komaki, T. Yanagi, Q. Guoxin, F. Maeda, M. and “Cardinal larkspur (Scarlet larkspur)”. It is possible Ishikawa, S. Kuhara, S. Sato, S. Tabata and S. N. Isobe. that Legro (1961) used heterogeneous populations as 2014. Dissection of the octoploid strawberry genome by deep sequencing of the genomes of Fragaria species. DNA hybridization parents in his study, whereas the parental Res. 21: 169–181. plants used here were thought to be relatively homoge‐ Honda, K., K. Tsutsui and K. Hosokawa. 1999. Analysis of the neous because they had been derived by self- flower pigments of some Delphinium species and their inter‐ hybridization from selected lines to provide stable specific hybrids produced via ovule culture. Sci. Hort. 82: phenotypes. 125–134. In this study, we used wild-type delphiniums as a Hoshino, A., V. Jayakumar, E. Nitasaka, A. Toyoda, H. Noguchi, gene resource to produce novel delphinium flower col‐ T. Itoh, T. Shin-I, Y. Minakuchi, Y. Koda, A. J. Nagano, M. Yasugi, M. N. Honjo, H. Kudoh, M. Seki, A. Kamiya, T. ors associated with the accumulation of cyanidin. These Shiraki, P. Carninci, E. Asamizu, H. Nishide, S. Tanaka, K. hybrid lines can be used to generate novel varieties col‐ Park, Y. Morita, K. Yokoyama, I. Uchiyama, Y. Tanaka, S. ored by cyanidin, e.g., black (Dahlia variabilis: Tabata, K. Shinozaki, Y. Hayashizaki, Y. Kohara, Y. Suzuki, Deguchi et al., 2016), sky blue (Meconopsis grandis: S. Sugano, A. Fujiyama, S. Iida and Y. Sakakibara. 2016. Tanaka et al., 2001), blue (Centaurea cyanus: Takeda Genome sequence and analysis of the Japanese morning et al., 2005), and a rich variety of reds. Consequently, glory Ipomoea nil. Nat. Commun. 7: 13295. DOI: 10.1038/ our progress in the generation of delphinium hybrids ncomms13295. Ishii, I., K. Sakaguchi, K. Fujita, Y. Ozeki and T. Miyahara. 2017. that accumulate cyanidin will undoubtedly lead to an A double knockout mutant of acyl-glucose-dependent antho‐ expansion of the available flower color varieties of del‐ cyanin glucosyltransferase genes in Delphinium phiniums. grandiflorum. J. Plant Physiol. 216: 74–78. Katoh, N., K. Tokuhiro, T. Nakabayashi, H. Yoshida and M. Conclusions Hagimori. 2004. Breeding of red-flowering delphiniums The F3'H gene of D. zalil was introduced into using interspecific hybrids produced by in vitro germination. D. cardinale by conventional hybridization breeding. Breed. Sci. 54: 99–103. Kondo, T., K. Oki, K. Yoshida and T. Goto. 1990. Structure of The F hybrids produced by this cross showed novel 1 violdelphin, an anthocyanin from violet flower of red-colored sepals containing cyanidin-based anthocya‐ Delphinium hybridum. Chem. Lett. 19: 137–138. nin, which has not been found to date in delphiniums. Kondo, T., K. Suzuki, K. Yoshida, K. Oki, M. Ueda, M. Isobe and Our results demonstrate the feasibility of a strategy T. Goto. 1991. Structure of cyanodelphin, a tetra-p- based on the identification of the chemical structures of hydroxybenzoated anthocyanin from blue flower of flower pigments and the screening of the genes for en‐ Delphinium hybridum. Tetrahedron Lett. 32: 6375–6378. zymes that can generate these structures. This approach Legro, R. A. H. 1961. Species hybrids in delphinium. Euphytica 10: 1–21. can be applied to both horticultural and wild cultivars Matsuba, Y., N. Sasaki, M. Tera, M. Okamura, Y. Abe, E. and has the potential to identify novel traits and under- Okamoto, H. Funabashi, M. Takatsu, M. Saito, H. Matsuoka, utilized gene resources that could be used with conven‐ K. Nagasawa and Y. Ozeki. 2010. A novel glucosylation re‐ tional hybridization techniques to generate novel action on anthocyanins catalyzed by acyl-glucose-dependent phenotypes. glucosyltransferase in the petals of carnation and delphinium. Plant Cell 22: 3374–3389. Literature Cited Matsubara, K., S. Chen, J. Lee, H. Kodama, H. Kokubun, H. Watanabe and T. Ando. 2006. PCR-based markers for the Anderson, N. O. 2006. Flower breeding and genetics: Issues, genotype identification of flavonoid-3',5'-hydroxylase genes challenges and opportunities for the 21st century. Springer, governing floral anthocyanin biosynthesis in commercial The Netherlands. petunias. Breed. Sci. 5: 389–397. Davies, K. M. 2009. Modifying anthocyanin production in flow‐ Miyagawa, N., T. Miyahara, M. Okamoto, Y. Hirose, K. ers. p. 49–80. In: K. Gould, K. M. Davies and C. Winefield Sakaguchi, S. Hatano and Y. Ozeki. 2015. Dihydroflavonol (eds.). Anthocyanins: Biosynthesis, Functions, and Applica‐ 4-reductase activity is associated with the intensity of flower tions. Springer Science + Business Media, LLC, New York. colors in delphinium. Plant Biotechnol. 32: 249–255. Deguchi, A., F. Tatsuzawa, M. Hosokawa, M. Doi and S. Ohno. Miyagawa, N., Y. Nishizaki, T. Miyahara, M. Okamoto, Y. 2016. Quantitative evaluation of the contribution of four Hirose, Y. Ozeki and N. Sasaki. 2014. Sequence variations major anthocyanins to black flower coloring of dahlia petals. in the flavonoid 3',5'-hydroxylase gene associated with red‐ Hort. J. 85: 340–350. dish flower phenotypes in three delphinium varieties. Plant Hashimoto, F., M. Tanaka, H. Maeda, S. Fukuda, K. Shimizu and Biotechnol. 31: 83–87. Y. Sakata. 2002. Changes in flower coloration and sepal an‐ Miyahara, T., A. Hamada, M. Okamoto, Y. Hirose, K. Sakaguchi, thocyanins of cyanic Delphinium cultivars during flowering. S. Hatano and Y. Ozeki. 2016. Identification of flavonoid 3'- Biosci. Biotech. Biochem. 66: 1652–1659. hydroxylase in the yellow flower of Delphinium zalil. J. Hashimoto, F., M. Tanaka, H. Maeda, K. Shimizu and Y. Sakata. Plant Physiol. 202: 92–96. 2000. Characterization of cyanic flower color of Delphinium Nakatsuka, T. and K. Koishi. 2018. Molecular characterization of cultivars. J. Japan. Soc. Hort. Sci. 69: 428–434. double-flower mutation in Matthiola incana. Plant Sci. 268: Hirakawa, H., K. Shirasawa, S. Kosugi, K. Tashiro, S. Nakayama, 39–46. M. Yamada, M. Kohara, A. Watanabe, Y. Kishida, T. Nakatsuka, T., M. Saito, Y. Sato-Ushiku, E. Yamada, T. Nakasato, Fujishiro, H. Tsuruoka, C. Minami, S. Sasamoto, M. Kato, N. Hoshi, K. Fujiwara, T. Hikage and M. Nishihara. 2011. Hort. J. Preview 7

Development of DNA markers that discriminate between Centaurea cyanus. Phytochemistry 66: 1607–1613. white-and blue-flowers in Japanese gentian plants. Tanaka, M., T. Fujimori, I. Uchida, S. Yamaguchi and K. Takeda. Euphytica 184: 335–344. 2001. A malonylated anthocyanin and flavonols in blue Nishihara, M., K. Tasaki, N. Sasaki and H. Takahashi. 2018. De‐ Meconopsis flowers. Phytochemistry 56: 373–376. velopment of basic technologies for improvement of breed‐ Tanaka, Y. and F. Brugliera. 2013. Flower colour and cyto‐ ing and cultivation of Japanese gentian. Breed. Sci. 68: 14– chromes P450. Phil. Trans. R. Soc. B. 368: 20120432. DOI: 24. 10.1098/rstb.2012.0432. Nishizaki, Y., N. Sasaki, M. Yasunaga, T. Miyahara, E. Okamoto, Yagi, M. 2013. Application of DNA markers for breeding carna‐ M. Okamoto, Y. Hirose and Y. Ozeki. 2014. Identification of tions resistant to bacterial wilt. Jpn. Agr. Res. Q. 47: 29–35. the glucosyltransferase genes that supplies the p- Yagi, M. 2018. Recent progress in whole genome sequencing, hydroxybenzoyl-glucose for 7-polyacylation of anthocyanin high-density linkage maps, and genomic databases of orna‐ in delphinium. J. Exp. Bot. 65: 2495–2506. mental plants. Breed. Sci. 68: 62–70. Nishizaki, Y., M. Yasunaga, E. Okamoto, M. Okamoto, Y. Hirose, Yagi, M., S. Kosugi, H. Hirakawa, A. Ohmiya, K. Tanase, T. M. Yamaguchi, Y. Ozeki and N. Sasaki. 2013. p- Harada, K. Kishimoto, M. Nakayama, K. Ichimura, T. Hydoroxybenzoyl-glucose is a Zwitter donor for the biosyn‐ Onozaki, H. Yamaguchi, N. Sasaki, T. Miyahara, Y. thesis of 7-polyacylated anthocyanin in delphinium. Plant Nishizaki, Y. Ozeki, N. Nakamura, T. Suzuki, Y. Tanaka, S. Cell 25: 4150–4165. Sato, K. Shirasawa, S. Isobe, Y. Miyamura, A. Watanabe, S. Okamura, M., M. Nakayama, N. Umemoto, E. A. Cano, Y. Hase, Nakayama, Y. Kishida, M. Kohara and S. Tabata. 2014. Y. Nishizaki, N. Sasaki and Y. Ozeki. 2013. Crossbreeding Sequence analysis of the genome of carnation (Dianthus of a metallic color carnation and diversification of the pecu‐ caryophyllus L.). DNA Res. 21: 231–241. liar coloration by ion-beam irradiation. Euphytica 191: 45– Yamaguchi, H., A. Shimizu, K. Degi and T. Morishita. 2008. 56. Effects of dose and dose rate of gamma ray irradiation on Onozaki, T., M. Yamada, M. Yagi, K. Tanase and M. Shibata. mutation induction and nuclear DNA content in 2018. Effects of crossing and selection for seven generations chrysanthemum. Breed. Sci. 58: 331–335. based on flower vase life in carnation (Dianthus Yamaguchi, H., A. Shimizu, Y. Hase, K. Degi, A. Tanaka and T. caryophyllus L.) and the relationship between ethylene pro‐ Morishita. 2009. Mutation induction with ion beam irradia‐ duction and flower vase life in the breeding lines. Hort. J. tion of lateral buds of chrysanthemum and analysis of 87: 106–114. chimeric structure of induced mutants. Euphytica 165: 97– Shibata, M. 2008. Importance of genetic transformation in orna‐ 103. mental plant breeding. Plant Biotechnol. 25: 3–8. Yamaguchi, H., A. Shimizu, Y. Hase, A. Tanaka, N. Shikazono, Takeda, K., A. Osakabe, S. Saito, D. Furuyama, A. Tomita, Y. K. Degi and T. Morishita. 2010. Effects of ion beam irradia‐ Kojima, M. Yamadera and M. Sakuta. 2005. Components of tion on mutation induction and nuclear DNA content in protocyanin, a blue pigment from the blue flowers of chrysanthemum. Breed. Sci. 60: 398–404.