International Journal of Molecular Sciences

Article Cloning and Functional Characterization of Dihydroflavonol 4-Reductase Gene Involved in Anthocyanin Biosynthesis of Chrysanthemum

Sun-Hyung Lim 1,*, Bora Park 2,3, Da-Hye Kim 2, Sangkyu Park 2, Ju-Hee Yang 2, Jae-A Jung 4 , JeMin Lee 3 and Jong-Yeol Lee 2,*

1 Division of Horticultural Biotechnology, School of Biotechnology, Hankyong National University, Anseong 17579, Korea 2 National Institute of Agricultural Sciences, Rural Development Administration, Jeonju 54874, Korea; [email protected] (B.P.); [email protected] (D.-H.K.); [email protected] (S.P.); [email protected] (J.-H.Y.) 3 Department of Horticultural Science, Kyungpook National University, Daegu 41566, Korea; [email protected] 4 Floriculture Research Division, National Institute of Horticultural & Herbal Science, Rural Development Administration, Wanju 55365, Korea; [email protected] * Correspondence: [email protected] (S.-H.L.); [email protected] (J.-Y.L.); Tel.: +82-31-670-5105 (S.-H.L.); +82-63-238-4616 (J.-Y.L.)  Received: 30 September 2020; Accepted: 25 October 2020; Published: 27 October 2020 

Abstract: Dihydroflavonol 4-reductase (DFR) catalyzes a committed step in anthocyanin and biosynthesis by reducing dihydroflavonols to . However, the role of this in determining flower color in the economically important crop chrysanthemum (Chrysanthemum morifolium Ramat.) is unknown. Here, we isolated cDNAs encoding DFR from two chrysanthemum cultivars, the white-flowered chrysanthemum “OhBlang” (CmDFR-OB) and the red-flowered chrysanthemum “RedMarble” (CmDFR-RM) and identified variations in the C-terminus between the two sequences. An enzyme assay using recombinant proteins revealed that both catalyzed the reduction of dihydroflavonol substrates, but CmDFR-OB showed significantly reduced DFR activity for dihydrokaempferol (DHK) as compared with CmDFR-RM. Transcript levels of anthocyanin biosynthetic genes were consistent with the anthocyanin contents at different flower developmental stages of both cultivars. The in planta complementation assay, using Arabidopsis thaliana dfr mutant (tt3-1), revealed that CmDFR-RM, but not CmDFR-OB, transgenes restored defective anthocyanin biosynthesis of this mutant at the seedling stage, as well as proanthocyanidin biosynthesis in the seed. The difference in the flower color of two chrysanthemums can be explained by the C-terminal variation of CmDFR combined with the loss of CmF3H expression during flower development.

Keywords: anthocyanin; chrysanthemum; dihydroflavonol 4-reductase; flavonoid; flower color

1. Introduction Flower color influences the commercial value of ornamental . In general, floral coloration involves the accumulation of pigments such as flavonoids (including anthocyanins), carotenoids, and . The flavonoid class anthocyanins are responsible for flower colors ranging from pale pink to blue in many ornamental plants including chrysanthemum (Chrysanthemum morifolium), carnation (Dianthus caryophyllus), lily (Lilium longiflorum), gerbera (Gerbera hybrida), petunia (Petunia hybrida), and rose (Rosa hybrida). The mechanism of anthocyanin biosynthesis has been extensively investigated, and most structural and regulatory genes involved in this process have been identified in several

Int. J. Mol. Sci. 2020, 21, 7960; doi:10.3390/ijms21217960 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 2 of 16

Int. J. Mol. Sci. 2020, 21, 7960 2 of 17 investigated, and most structural and regulatory genes involved in this process have been identified in several species [1–3]. Dihydroflavonol 4-reductase (DFR) plays an important role in the plantformation species of anthocyanins [1–3]. Dihydroflavonol and , 4-reductase (DFR) as shown plays anin the important Figure role1. DFR in thecatalyzes formation the ofconversion anthocyanins of andthe proanthocyanidins, colorless dihydroflavonols, as shown in theincluding Figure1. DFRdihydrokaempferol catalyzes the conversion (DHK), ofdihydroquercetin the colorless dihydroflavonols, (DHQ), and dihydromyricetin including dihydrokaempferol (DHM), into leucoanthocyanidins. (DHK), dihydroquercetin These (DHQ), three andcolorless dihydromyricetin dihydroflavonols (DHM), are intosubs leucoanthocyanidins.trates of flavonol biosynthesis These three that colorless are catalyzed dihydroflavonols by flavonol are substratessynthase (FLS). of flavonol Competition biosynthesis between that areFLS catalyzed and DFR by can flavonol modify synthase the metabolic (FLS). Competitionflux and alter between flower FLScolor and [4–6]. DFR can modify the metabolic flux and alter flower color [4–6].

Figure 1.1. Schematic representation of flavonols,flavonols, proanthocyanidins,proanthocyanidins, and anthocyaninsanthocyanins biosyntheticbiosynthetic pathway in in plants plants The The abbreviations abbreviations of ofenzyme enzyme names names are areas follows: as follows: CHS, CHS, chalcone chalcone synthase; synthase; CHI, CHI,chalcone chalcone ; isomerase; F3H, F3H,flavanone flavanone 3-hydroxylase; 3-hydroxylase; FLS, FLS,flavonol flavonol synthase; synthase; F3′H, F3 0flavonoidH, flavonoid 3′- 3hydroxylase;0-hydroxylase; F3 F3′5′H,050 H, flavonoid 3′ 3,50′,5-hydroxylase;0-hydroxylase; DFR, DFR, dihydroflavonol dihydroflavonol 4-reductase; 4-reductase; LAR, reductase; reductase; ANR, ANR, anthocyanidi anthocyanidinn reductase; reductase; ANS, ANS, anthocyanidin anthocyanidin synthase; synthase; and andUFGT, UFGT, UDP-glucose: UDP-glucose: flavonoid flavonoid 3-O-glucosyltransferase. 3-O-glucosyltransferase.

Several studies have revealed that the substrate specificityspecificity of DFR plays an important role in the determination ofof anthocyaninanthocyanin types types [ 7[7–10].–10]. According According to to di differencesfferences at at the the 134th 134th amino amino acid acid residues residues in theirin their substrate substrate binding binding domains, domains, DFRs DFRs are divided are divi intoded three into types,three thetypes, asparagine the asparagine (Asn)-type, (Asn)-type, aspartic acidaspartic (Asp)-type, acid (Asp)-type, and non-Asn and non-Asn/Asp-type./Asp-type. Previous Prev studiesious have studies reported have reported that Asn-type that Asn-type DFRs catalyze DFRs thecatalyze conversion the conversion of all three of dihydroflavonolsall three dihydrofla (DHK,vonols DHQ, (DHK, and DHQ, DHM), and whereas DHM), Asp-typewhereas DFRsAsp-type are highlyDFRs are specific highly for specific DHQ and for DHMDHQ ratherand DHM than rather for DHK than as for a substrate DHK as [a8– substrate10]. However, [8–10]. not However, all Asn-type not DFRsall Asn-type can catalyze DFRs all can three catalyze dihydroflavonols all three dihydroflavonols as substrates. DFRs as substrates. from sweet DFRs potato from (IbDFR) sweet and potato from freesia(IbDFR) (FeDFR2) and from are freesia classified (FeDFR2) into the are Asn-type classified DFR, into duethe toAsn-type their 134th DFR, amino due to acid their residues 134th [amino10,11]. IbDFRacid residues catalyzes [10,11]. only IbDFR the DHK catalyzes but not only DHQ the and DHK DHM, but not while DHQ FeDFR2 and DHM, utilizes while both FeDFR2 the DHQ utilizes and DHMboth the but DHQ not the and DHK DHM as a substrate.but not the Other DHK studies as a havesubstrate. shown Other that thestudies neighboring have shown residues that of the substrateneighboring binding residues sites of and the NADPsubstrate binding binding regions sites and of DFRs NADP play binding pivotal regions roles inof proteinDFRs play activity pivotal in buckwheatroles in protein (Fagopyrum activity esculentumin buckwheat), gerbera, (Fagopyrum and grape esculentum hyacinth), gerbera, (Muscari and armeniacum grape hyacinth)[11–13 (].Muscari These resultsarmeniacum imply) [11–13]. that other These factors results can imply influence that the other substrate factors specificity can influence of DFR the enzyme. substrate specificity of DFR Chrysanthemum,enzyme. a perennial herbaceous flowering plant, is an important floricultural crop worldwide.Chrysanthemum, The ray florets a perennial of chrysanthemum herbaceous flowers flowering can beplant, pink, is red,an important yellow, white, floricultural purple, green, crop orworldwide. various bicolored The ray florets forms of due chry tosanthemum the accumulation flowersof can di befferent pink, proportions red, yellow,of white, two pigments,purple, green, i.e., anthocyaninsor various bicolored and carotenoids forms due [14 ].to Therethe accumulation are three basic of types different of anthocyanidins, proportions of namely, two pigments, pelargonidin, i.e., ,anthocyanins and delphinidin.and carotenoids Although [14]. the There ray florets are three of chrysanthemum basic types onlyof anthocyanidins, accumulate the cyanidinnamely, derivatives,pelargonidin, the cyanidin, substrate and feeding delphinidin. assay revealed Although that CmDFRthe ray couldflorets utilize of chrysanthemum DHK and DHM only as a substrateaccumulate for the pelargonidin cyanidin derivatives, and delphinidin the substrat biosynthesis,e feeding respectively assay revealed [14–16 that]. CmDFR could utilize DHKIn and the DHM current as study,a substrate we identified for pelargonidin two DFR and genes, delphinidinCmDFR-OB biosynthesis,and CmDFR-RM, respectivelywith [14–16]. different C-terminus, from chrysanthemum flowers with white and red colors, respectively, and demonstrated that these genes function in anthocyanin biosynthesis. We performed an in vitro enzyme assay with

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 3 of 16

In the current study, we identified two DFR genes, CmDFR-OB and CmDFR-RM, with different C-terminus, from chrysanthemum flowers with white and red colors, respectively, and demonstrated that these genes function in anthocyanin biosynthesis. We performed an in vitro enzyme assay with recombinant CmDFRs and verified their catalytic activities for different dihydroflavonol substrates. Int. J. Mol. Sci. 2020, 21, 7960 3 of 17 In a complementation assay with the Arabidopsis thaliana mutant transparent testa3-1 (tt3-1), CmDFR- RM restored anthocyanin accumulation in rosette leaves and proanthocyanidin accumulation in the recombinantseeds, but CmDFR-OB CmDFRs and did verified not. Our their study catalytic indicates activities that forthe di CmDFRsfferent dihydroflavonol play a functional substrates. role in Inanthocyanin a complementation biosynthesis assay and with pigmentation the Arabidopsis in thalianathe ray mutantflorets oftransparent chrysanthemum. testa3-1 ( tt3-1), CmDFR-RM restored anthocyanin accumulation in rosette leaves and proanthocyanidin accumulation in the seeds, but2. ResultsCmDFR-OB did not. Our study indicates that the CmDFRs play a functional role in anthocyanin biosynthesis and pigmentation in the ray florets of chrysanthemum. 2.1. Nucleotide Polymorphisms of the CmDFR Gene in Different Colored Ray Florets Chrysanthemum 2.Cultivars Results

2.1. NucleotideThe NCBI Polymorphisms database lists of thefour CmDFR full-length Gene inDFR Di ffcloneserent Colored from chrysanthe Ray Florets Chrysanthemummum. All of these Cultivars genes share at least 95% sequence similarity with the first DFR isolated from chrysanthemum CmDFR (accessionThe NCBI ADC96612.1), database lists a 1125 four bp full-length sequenceDFR withclones an ORF from encoding chrysanthemum. a deduced All protein of these of genes 374 amino share atacids. least To 95% investigate sequence the similarity role of withDFR thein anthocyanin first DFR isolated biosynthesis, from chrysanthemum we isolated DFRCmDFR sequences(accession from ADC96612.1),OhBlang (OB) a and 1125 RedMarble bp sequence (RM) with using an ORF primers encoding designed a deduced based proteinon the ofCmDFR 374 amino sequences. acids. ToUnexpectedly, investigate the the role cDNA of DFR sequences in anthocyanin from OB biosynthesis, and RM showed we isolated a variationDFR insequences the C-terminal from OhBlangregion as (OB)compared and RedMarble with ADC96612.1. (RM) using Thus, primers we designeddesignated based CmDFR on the fromCmDFR OB assequences. CmDFR-OB Unexpectedly,, a 1059 bp thesequence cDNA with sequences an ORF from encoding OB and a RM deduced showed protei a variationn of 352 in amino the C-terminal acids (GenBank region asaccession compared number with ADC96612.1.MT513763). CmDFR Thus, we from designated RM (CmDFR-RMCmDFR from) is 1128 OB asbp CmDFR-OBlong, with an, a ORF 1059 encoding bp sequence a 375 with amino an ORF acid encodingprotein (GenBank a deduced accession protein of number 352 amino MT513764). acids (GenBank The high accession level numberof amino MT513763). acid sequenceCmDFR identityfrom RMamong (CmDFR-RM CmDFR,) CmDFR-OB, is 1128 bp long, and with CmDFR-RM an ORF encoding suggests a 375that amino the genes acid encoding protein (GenBank these proteins accession are numberallelic. Multiple MT513764). sequence The high alignment level of of amino DFRs acid from sequence chrysanthemum identity among revealed CmDFR, a single CmDFR-OB, nucleotide andsubstitution CmDFR-RM (T→G) suggests that resulted that the in a genes premature encoding stop thesecodon, proteins and also are a 7 allelic.bp deletion Multiple within sequence the sixth alignment of DFRs from chrysanthemum revealed a single nucleotide substitution (T G) that resulted exon of CmDFR-OB (Figure 2). → in a premature stop codon, and also a 7 bp deletion within the sixth exon of CmDFR-OB (Figure2).

Figure 2. Schematic representation and comparison of the nucleotide sequences of DFR genes in Figure 2. Schematic representation and comparison of the nucleotide sequences of DFR genes in chrysanthemum cultivars OhBlang (OB) and RedMarble (RM). Exons are indicated by open boxes and chrysanthemum cultivars OhBlang (OB) and RedMarble (RM). Exons are indicated by open boxes introns by a thin line. Different sequence regions are marked with a dashed red line, and the asterisk and introns by a thin line. Different sequence regions are marked with a dashed red line, and the indicates a stop codon. asterisk indicates a stop codon. According to the nucleotide polymorphism of CmDFR, we screened 22 chrysanthemum cultivars (14 white,According 4 pink, to and the 4nucleotide red) using polymorphism an insertion and of CmDFR deletion, (InDel)we screened maker 22 (Figure chrysanthemum3). To validate cultivars the accuracy(14 white, of 4 the pink, InDel and marker, 4 red) using we used an insertion plasmids and containing deletionCmDFR-OB (InDel) makerand (FigureCmDFR-RM 3). To validateas positive the controlsaccuracy for of analysis.the InDel Consistent marker, we with used the plasmids results ofcontaining DNA sequencing CmDFR-OB analysis, and CmDFR-RM the CmDFR asfragment positive amplifiedcontrols for from analysis. RM was Consistent 127 bp in with size the and results that fromof DNA OB sequencing was 120 bp analysis, in size due the to CmDFR a 7 bp fragment deletion (Figureamplified3B). from PCR usingRM was InDel 127 marker bp in generatedsize and that 120 from bp amplicons OB was from120 bp all in 11 size chrysanthemum due to a 7 bp cultivars deletion with(Figure white 3B). ray PCR florets using except InDel for AG,marker MY, generate and SD. Byd 120 contrast, bp amplicons 127 bp amplicons from all were 11 chrysanthemum produced from allcultivars eight chrysanthemum with white ray cultivarsflorets except with for pink AG, or redMY, colored and SD. ray By florets contrast, and 127 in thebp positiveamplicons control were CmDFR-RMproduced from. These all resultseight chrysanthemum suggest that that cultivars the DFR withgene pink alone or cannot red colored be used ray to determineflorets and the in raythe floret color.

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 4 of 16 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 4 of 16 positive control CmDFR-RM. These results suggest that that the DFR gene alone cannot be used to positive control CmDFR-RM. These results suggest that that the DFR gene alone cannot be used to Int.determine J. Mol. Sci. the2020 ray, 21 ,floret 7960 color. 4 of 17 determine the ray floret color.

Figure 3. Insertion and deletion (InDel) analysis of the CmDFR gene in chrysanthemum. (A) FigureFigure 3. InsertionInsertion and and deletion deletion (InDel) (InDel) analysis analysis of ofthe theCmDFRCmDFR genegene in chrysanthemum. in chrysanthemum. (A) PhotographsA of 22 chrysanthemum flowers showing different ray floret colors, scale bar indicates 1 (Photographs) Photographs of 22 of chrysanthemum 22 chrysanthemum flowers flowers showing showing different different ray ray floret floret colors, colors, scale scale bar bar indicates indicates 1 1cm; cm; (B (B) )PCR PCR products products amplifiedamplified usingusing InDel-DFRInDel-DFR primers primers and and separated separated on on an an 6% 6% polyacrylamide polyacrylamide gel. cm; (B) PCR products amplified using InDel-DFR primers and separated on an 6% polyacrylamide M,gel. DNA M, DNA molecular molecular weight weight markers. markers. PR and PR PO and used PO as used InDel as marker InDel controlmarker represent control represent the amplified the gel. M, DNA molecular weight markers. PR and PO used as InDel marker control represent the fragmentamplified fromfragment the positive from the plasmids positive containingplasmids containingCmDFR-RM CmDFR-RMand CmDFR-RM and CmDFR-RM, respectively., respectively. amplified fragment from the positive plasmids containing CmDFR-RM and CmDFR-RM, respectively. Sequence alignment showed that CmDFRs containcontain a highly conservedconserved NADP bindingbinding region Sequence alignment showed that CmDFRs contain a highly conserved NADP binding region and a substrate binding region at their N-termini, whereas their C-termini are diverse (Figure 44A).A). and a substrate binding region at their N-termini, whereas their C-termini are diverse (Figure 4A). These conserved NADP binding domains and substratesubstrate binding domains are present in dicot plants These conserved NADP binding domains and substrate binding domains are present in dicot plants such asas Arabidopsis,Arabidopsis, cupflowercupflower ((NierembergiaNierembergia spathulataspathulata), gentiangentian ((Gentiana trifloratriflora), gerbera,gerbera, grapegrape such as Arabidopsis, cupflower ( spathulata), gentian (Gentiana triflora), gerbera, grape ((Vitis vinifera), and petunia,petunia, and also in monocotmonocot species such as grape hyacinth, rice (Oryza sativa ), (Vitis vinifera), and petunia, and also in monocot species such as grape hyacinth, rice (Oryza sativa), and (Zea mays),), suggestingsuggesting thatthat thesethese proteinsproteins areare highlyhighly conservedconserved amongamong angiosperms.angiosperms. and maize (Zea mays), suggesting that these proteins are highly conserved among angiosperms. Phylogenetic analysis of DFRs fromfrom various plant species clearly classifiedclassified DFRs from monocots Phylogenetic analysis of DFRs from various plant species clearly classified DFRs from monocots and into didifferentfferent branchesbranches (Figure(Figure4 4B).B). DFR DFR proteins proteins including including CmDFR,CmDFR, CmDFR-OB,CmDFR-OB, andand and eudicots into different branches (Figure 4B). DFR proteins including CmDFR, CmDFR-OB, and CmDFR-RM were grouped grouped into into the the same same subclade subclade wi withth gerbera gerbera in in the the Asteraceae Asteraceae family. family. Similar Similar to CmDFR-RM were grouped into the same subclade with gerbera in the Asteraceae family. Similar to tothe the DFRs DFRs of ofthe the Asteraceae Asteraceae family, family, petunia petunia PhDF PhDFRR was was classified classified close close to to potato StDFR, pepperpepper the DFRs of the Asteraceae family, petunia PhDFR was classified close to potato StDFR, pepper CsDFR, and cupflowercupflower NsDFR in thethe SolanaceaeSolanaceae family.family. The resultsresults ofof phylogeneticphylogenetic analysisanalysis alignalign CsDFR, and cupflower NsDFR in the family. The results of phylogenetic analysis align wellwell withwith thethe geneticgenetic relationshipsrelationships amongamong thethe species.species. Collectively, our data suggests that thesethese well with the genetic relationships among the species. Collectively, our data suggests that these chrysanthemum DFRs possessing the conservative characteristics can have a functional DFR role in chrysanthemum DFRs possessing the conservative characteristics can have a functional DFR role in flavonoidflavonoid biosynthesis. .

(A) (A) Figure 4. Cont. Int.Int. J.J. Mol.Mol. Sci. 2020,, 21,, 7960x FOR PEER REVIEW 55 ofof 1716

(B)

FigureFigure 4.4.Amino Amino acid acid sequence sequence alignment alignment and phylogenetic and phyl analysisogenetic of analysis DFR proteins of DFR in Chrysanthemum proteins in morifoliumChrysanthemumand thosemorifolium from and other those species. from other (A) Multiple species. ( sequenceA) Multiple alignment sequence ofalignment CmDFRs of and CmDFRs DFR proteinsand DFR from proteins other from species. other Numbers species. Numbers indicate theindica positionte the ofposition the last of aminothe last acid amino in each acid proteinin each sequence.protein sequence. The putative The NADPputative binding NADP sitebinding and presumed site and presumed substratebinding substrate region binding are markedregion are by bluemarked and by red blue boxes, and respectively. red boxes, respectively. Substrate specificity Substrate is specificity associated is with associated aa134 and with aa145 aa134 (highlighted and aa145 with(highlighted an inverted with red an arrowheadinverted red and arrowhead an inverted and black an inverted arrowhead, black respectively); arrowhead, (respectively);B) Phylogenetic (B) relationshipsPhylogenetic amongrelationships DFR proteins among from DFR chrysanthemum proteins from and chrysanthemum other plants. The and phylogenetic other plants. tree wasThe constructedphylogenetic using tree the was neighbor-joining constructed using method the with neighbor-joining MEGA6 software. method Numbers with next MEGA6 to the nodessoftware. are bootstrapNumbers valuesnext to fromthe nodes 1000 are replications. bootstrap Thevalues tree from is drawn 1000 replications. to scale and The has tree branch is drawn lengths to scale with and the samehas branch units aslengths those with of the the evolutionary same units distancesas those of that the were evolutionary used to inferdistances the phylogenetic that were used tree to (scale infer bar,the 0.05phylogenetic amino acid tree substitutions (scale bar,per 0.05 site). amino The acid GenBank substitutions accession per numbers site). ofThe the GenBank proteinsequences accession usednumbers are as of follows: the proteinActinidia sequences chinensis usedAcDFR are (PSS36490.1); as follows: AgapanthusActinidia chinensis praecox ApDFRAcDFR (AB099529.1);(PSS36490.1); AntirrhinumAgapanthus praecox majus AmDFR ApDFR (X15536);(AB099529.1);Arabidopsis Antirrhinum thaliana majusAtDFR AmDFR (AB033294); (X15536);Callistephus Arabidopsis chinensis thaliana CcDFRAtDFR (P51103.1);(AB033294);Camellia Callistephus sinensis chinensisCsDFR CcDFR (AAT84073.1); (P51103.1);Capsicum Camellia annuum sinensisCaDFR CsDFR (NP_001311706.1); (AAT84073.1); ChrysanthemumCapsicum annuum morifolium CaDFRCmDFR (NP_001311706.1); (ADC96612), CmDFR-OBChrysanthemum (MT513763), morifolium CmDFR-RM CmDFR (ADC96612), (MT513764); DaucusCmDFR-OB carota (MT513763),DcDFR (XP_017254990.1); CmDFR-RM (MT513764);Freesia hybrida DaucusFhDFR1 carota (KU132393); DcDFR (XP_017254990.1);Gentiana triflora GtDFR Freesia (BAA12736.1);hybrida FhDFR1Gerbera (KU132393); hybrida GhDFR Gentiana (AKN56969.1); triflora GtDFRGynura (BAA12736.1); bicolor GbDFR Gerbera (BAJ17657.1); hybridaHelianthus GhDFR annuus(AKN56969.1);HaDFR Gynura (XP_022001438.1); bicolor GbDFRIpomoea (BAJ17657.1); batatas IbDFR Helianthus (BAD05164.1); annuus HaDFRIpomoea (XP_022001438.1); purpurea IpDFR (BAA36406.1); Iris hollandica IhDFR (BAF93856.1); Lilium hybrida LhDFR (BAB40789.1); Lonicera Ipomoea batatas IbDFR× (BAD05164.1); Ipomoea purpurea IpDFR× (BAA36406.1); Iris × hollandica IhDFR caerulea(BAF93856.1);LcDFR Lilium (ALU09329.1); × hybridaMuscari LhDFR armeniacum (BAB40789.1);MaDFR Lonicera (AIC33028.1); caerulea LcDFRNierembergia (ALU09329.1);sp. NB17 Muscari NsDFR (BAC10993.1);armeniacum MaDFROryza (AIC33028.1); sativa, OsDFR Nierembergia (AB003495); sp.Perilla NB17 frutescens NsDFR PfDFR(BAC10993.1); (AB002817); OryzaPetunia sativa, hybridaOsDFR PhDFR(AB003495); (CAA56160); Perilla frutescensSolanum tuberosum PfDFR (AB002817);StDFR (AEN83503.1); Petunia hybridaVitis vinifera PhDFRVvDFR (CAA56160); (X75964); Solanumand Zea maystuberosumZmDFR StDFR (Y16040). (AEN83503.1); Vitis vinifera VvDFR (X75964); and Zea mays ZmDFR (Y16040). 2.2. Recombinant CmDFRs Exhibit Different Enzymatic Properties 2.2. Recombinant CmDFRs Exhibit Different Enzymatic Properties We assayed the enzyme activity of recombinant CmDFR proteins in vitro. We successfully We assayed the enzyme activity of recombinant CmDFR proteins in vitro. We successfully produced two types of recombinant CmDFRs (CmDFR-OB, and CmDFR-RM) in IPTG-treated produced two types of recombinant CmDFRs (CmDFR-OB, and CmDFR-RM) in IPTG-treated

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 6 of 16 Int. J. Mol. Sci. 2020, 21, 7960 6 of 17 transformed bacterial cells. The majority of glutathione S- (GST)-CmDFR-RM was transformeddetected in the bacterial soluble cells. fraction, The majority whereas of GST-CmDFR-OB glutathione S-transferase was mostly (GST)-CmDFR-RM partitioned into wasthe insoluble detected ininclusion the soluble body fraction, (Figure whereas 5). Although GST-CmDFR-OB a relatively was small mostly amount partitioned of CmDFR-OB into the insoluble was obtained inclusion as bodycompared (Figure with5). AlthoughCmDFR-RM a relatively due to small their amountreduced of solubility, CmDFR-OB both was GST-cleaved obtained as compared proteins withwere CmDFR-RMsuccessfully purified. due to their reduced solubility, both GST-cleaved proteins were successfully purified.

FigureFigure 5.5.Expression Expression and and purification purification of recombinant of recombinan CmDFRt CmDFR proteins. proteins. The coding The regionscoding ofregionsCmDFRs of fromCmDFRs RM andfrom OB RM without and OB the without stop codon the stop were codon cloned we intore cloned pGEX-6P-1 into in-framepGEX-6P-1 with in-frame a glutathione with a Sglutathione-transferase S (GST)-transferase tag and (GST) expressed tag and in Escherichia expressedcoli in EscherichiaBL21 (DE3) coli cells. BL21 Bacterial (DE3) lysatescells. Bacterial were prepared lysates forwere gel prepared loading before for gel (–) loading and after befo (+re) IPTG (–) and induction. after (+) The IPTG IPTG-induced induction. The cells IPTG-induced were lysed by sonicationcells were andlysed separated by sonication into the and aqueous separate fractiond into (soluble) the aqueous and non-aqueous fraction (s pelletoluble) (insoluble) and non-aqueous by centrifugation. pellet Recombinant(insoluble) by proteins centrifugation. in the aqueous Recombinant fraction proteins were collected in the usingaqueous Glutathione fraction were Sepharose collected 4B beads,using and their GST tag was removed by PreScission protease (cleaved). The GST moiety that remained on Glutathione Sepharose 4B beads, and their GST tag was removed by PreScission protease (cleaved). the beads was eluted (GST-eluted). The samples were subjected to SDS-PAGE to confirm the expression The GST moiety that remained on the beads was eluted (GST-eluted). The samples were subjected to and purification of the proteins. Asterisks indicate induced recombinant CmDFR proteins displayed at SDS-PAGE to confirm the expression and purification of the proteins. Asterisks indicate induced corresponding size with or without GST moiety. recombinant CmDFR proteins displayed at corresponding size with or without GST moiety. We performed and in vitro enzyme activity assay using dihydroflavonol substrates and equal We performed and in vitro enzyme activity assay using dihydroflavonol substrates and equal amounts of the two recombinant CmDFR proteins. Since the products generated from these amounts of the two recombinant CmDFR proteins. Since the products generated from these reactions, reactions, leucoanthocyanidins, were unstable, we chemically converted these compounds to stable leucoanthocyanidins, were unstable, we chemically converted these compounds to stable anthocyanidins with acidic alcohol and identified the corresponding anthocyanidins based on anthocyanidins with acidic alcohol and identified the corresponding anthocyanidins based on specific color development and HPLC analysis. Both CmDFRs converted DHK to , specific color development and HPLC analysis. Both CmDFRs converted DHK to leucopelargonidin, as identified by color development (Figure6A) and peaks corresponding to pelargonidin in HPLC as identified by color development (Figure 6A) and peaks corresponding to pelargonidin in HPLC (Figure6B). In this reaction, CmDFR-RM exhibited high enzymatic activity, whereas CmDFR-OB (Figure 6B). In this reaction, CmDFR-RM exhibited high enzymatic activity, whereas CmDFR-OB showed only slight activity. When DHQ was used as the substrate, both CmDFRs showed comparable showed only slight activity. When DHQ was used as the substrate, both CmDFRs showed levels of activity, although CmDFR-OB activity was somewhat lower than CmDFR-RM. The catalytic comparable levels of activity, although CmDFR-OB activity was somewhat lower than CmDFR-RM. activity for the substrate dihydromyricetin (DHM) was detected only in CmDFR-RM. When extracting The catalytic activity for the substrate dihydromyricetin (DHM) was detected only in CmDFR-RM. the reactant with ethyl acetate, the recovery rate of each leucoanthocyanidin varies depending When extracting the reactant with ethyl acetate, the recovery rate of each leucoanthocyanidin product on their B-ring hydroxylation [7]. Therefore, the difference in substrate preference between each varies depending on their B-ring hydroxylation [7]. Therefore, the difference in substrate preference dihydroflavonol substrate could not be determined from the results of this experiment, but these between each dihydroflavonol substrate could not be determined from the results of this experiment, results clearly showed that the catalytic activity of CmDFR-OB was lower than that of CmDFR-RM, but these results clearly showed that the catalytic activity of CmDFR-OB was lower than that of and in particular, its activity for DHK substrate was drastically decreased. CmDFR-RM, and in particular, its activity for DHK substrate was drastically decreased.

Int. J. Mol. Sci. 2020, 21, 7960 7 of 17 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 7 of 16

Figure 6. In vitro assays of recombinant CmDFR proteins. (A) DFR activities of recombinant CmDFR-OB,Figure 6. In andvitro CmDFR-RM assays of recombinant using the dihydroflavonol CmDFR proteins. substrates (A) DFR DHK, activities DHQ, of and recombinant DHM. The unstableCmDFR- leucoanthocyanidinsOB, and CmDFR-RM produced using the by dihydroflavonol DFR activity were su chemicallybstrates DHK, converted DHQ, toand stable DHM. anthocyanidins, The unstable andleucoanthocyanidins the corresponding produced color development by DFR activity was were observed; chemically (B) converted HPLC analysis to stable of anthocyanidins, anthocyanidin metabolitesand the corresponding generated by color the enzymedevelopment activity was of CmDFR-OB,observed; (B)and HPLC CmDFR-RM. analysis of Black anthocyanidin and grey chromatogramsmetabolites generated and arrows by the indicate enzyme anthocyanidins activity of CmDFR-OB, production and and CmDFR-RM. dihydroflavonols Black substrates and grey detectedchromatograms at 520 and and 288 arrows nm, respectively. indicate anthocyanidins PG, pelargonidin; production CY, cyanidin; and dihydroflavonols DEL, delphinidin. substrates detected at 520 and 288 nm, respectively. PG, pelargonidin; CY, cyanidin; DEL, delphinidin. 2.3. Anthocyanin Content and Expression Analysis of CmDFR in OB and RM 2.3. AnthocyaninTo investigate Content the relationship and Expression between Analysis pigmentation of CmDFR and in OB anthocyanin and RM content, we measured the anthocyaninTo investigate contents the in relationship ray florets at between three developmental pigmentation stagesand anthocyanin of the white content, cultivar we OB measured and the redthe cultivaranthocyanin RM (Figure contents7A). in The ray anthocyanin florets at three contents developmental were consistent stages with of the the white visible cultivar appearance OB and of thethe ray red florets cultivar of RM RM (Figure (Figure7B). 7A). In these The plants,anthocyani the anthocyaninn contents contentwere consis wastent highest with at floweringthe visible stageappearance 1 (FS1) andof the gradually ray florets decreased of RM (Figure during flower7B). In development. these plants, the These anthoc resultsyanin indicate content that was variations highest inat anthocyaninflowering stage content 1 (FS1) are and responsible gradually for decreased the color du diringfferences flower in development. ray florets of OB These vs. results RM. indicate that Tovariations investigate in anthocyanin the relationship content between are responsibl color phenotypee for the color and differences transcript levelsin ray offlorets anthocyanin of OB vs. biosyntheticRM. genes, we analyzed the expression of seven structural genes including four early biosyntheticTo investigate genes (EBG),the relationship i.e., chalcone between synthase color phenotype (CHS), chalcone and transcript isomerase levels (CHI of), anthocyanin flavanone 3-hydroxylasebiosynthetic genes, (F3H), we and analyzed flavonid 3the0-hydroxylase expression (ofF3 0sevenH), and structural three late genes biosynthetic including genes four (LBG), early i.e.,biosynthetic dihydroflavonol genes (EBG), 4-reductase i.e., (chalconeDFR), anthocyanidin synthase (CHS synthase), chalcone (ANS ),isomerase and UDP-glucose: (CHI), flavanone flavonoid 3- 3-O-glucosyltransferasehydroxylase (F3H), and (UFGT flavonid) by 3 quantitative′-hydroxylase reverse-transcription (F3′H), and three late PCR biosynthetic (qPCR) (Figure genes7C). (LBG), For both i.e., OBdihydroflavonol and RM, transcript 4-reductase level of(DFR EBG), wasanthocyanidin highly detected synthase at FS1 (ANS and), dramaticallyand UDP-glucose: decreased flavonoid at FS2. 3- ExceptO-glucosyltransferase for CmCHS, all of ( EBGUFGT was) by highly quantitative detected reverse-transcription in RM as compared with PCR OB. (qPCR) Interestingly, (Figure transcript 7C). For levelboth ofOBCmF3H and RM,was transcript barely detected level of EBG in all was flower highly developmental detected at FS1 stage and of dramatically OB, while it decreased was highly at expressedFS2. Except in for those CmCHS in RM., all In of the EBG expression was highly profile detected of LBG, in RM OB as showed compared a high with transcript OB. Interestingly, level at FS1,transcript but a sharplylevel of lowCmF3H transcript was barely level detected according in to all the flower flower developmental development. stage However, of OB, RM while showed it was highhighly expression expressed pattern in those at in FS1 RM. and In gradually the expression reduced profile transcript of LBG, level OB showed until FS3. a high Overall, transcript it showed level thatat FS1, coordinated but a sharply expression low transcript of anthocyanin level according biosynthetic to the genes,flower except development. of CmCHS However,coincided RM with showed the patternshigh expression of anthocyanin pattern accumulation at FS1 and gradually at different reduce stagesd transcript of flower development.level until FS3. Overall, it showed that coordinated expression of anthocyanin biosynthetic genes, except of CmCHS coincided with the patterns of anthocyanin accumulation at different stages of flower development.

Int. J. Mol. Sci. 2020, 21, 7960 8 of 17 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 8 of 16

FigureFigure 7. 7. Phenotypes,Phenotypes, anthocyanin anthocyanin contents, contents, and and anthocyani anthocyaninn biosynthetic biosynthetic gene gene expression expression levels levels in in chrysanthemumchrysanthemum cultivars cultivars OB andand RM.RM. ( A(A)) Photographs Photographs of of flowers flowers at diatff dierentfferent developmental developmental stages stages (FS1, (FS1,FS2, andFS2, FS3)and fromFS3) from the two the chrysanthemum two chrysanthe cultivarsmum cultivars examined examined in this study;in this (study;B) Anthocyanin (B) Anthocyanin contents contentsin ray florets in ray of bothflorets cultivars; of both (C )cultivars; qPCR analysis (C) qPCR of anthocyanin analysis biosyntheticof anthocyanin genes biosynthetic including four genes early includingbiosynthetic four genes early (chalcone biosynthetic synthase genes (CHS), (chalconechalcone isomerasesynthase (CHI),(CHS),flavanone chalcone 3-hydroxylase isomerase (CHI), (F3H) , flavanoneand flavonoid 3-hydroxylase 30-hydroxylase (F3H) (F3, and0H) ),flavonoid and three 3 late′-hydroxylase biosynthetic (F3 genes′H)), (dihydroflavonoland three late biosynthetic 4-reductase genes(DFR) ,(dihydroflavonol anthocyanidin synthase 4-reductase (ANS), (andDFR) UDP-glucose:, anthocyanidin flavonoid synthase 3-O -glucosyltransferase(ANS), and UDP-glucose: (UFGT)). Results represent mean values SD from three biological replicates. CmEF1α was used as the reference flavonoid 3-O-glucosyltransferase± (UFGT)). Results represent mean values ± SD from three biological replicates.gene. Diff erentCmEF1 lettersα was above used the as bars the indicate reference significantly gene. Diffe diffrenterent letters values above (p < 0.05) the calculated bars indicate using significantlyone-way ANOVA different followed values by (p a< Duncan’s 0.05) calculated multiple using range one-wa tests. y ANOVA followed by a Duncan’s 2.4.multiple CmDFR-RM range Complements tests. the Arabidopsis tt3-1 Mutant

2.4. CmDFR-RMWe evaluated Complements the effects the ofArabidopsisCmDFR-OB tt3-1and MutantCmDFR-RM on anthocyanin biosynthesis by introducing these genes into the Arabidopsis tt3-1 mutant, which failed to accumulate brown We evaluated the effects of CmDFR-OB and CmDFR-RM on anthocyanin biosynthesis by proanthocyanidins in seeds and anthocyanins in the junction between the stem and rosette leaves. introducing these genes into the Arabidopsis tt3-1 mutant, which failed to accumulate brown We obtained transgenic CmDFR-OB and CmDFR-RM plants via the floral dip method and selected proanthocyanidins in seeds and anthocyanins in the junction between the stem and rosette leaves. transformants by spraying the plants with 0.3% basta. To investigate the role of the CmDFRs in We obtained transgenic CmDFR-OB and CmDFR-RM plants via the floral dip method and selected anthocyanin accumulation, wild-type seeds, mutant, and T transgenic lines were germinated and transformants by spraying the plants with 0.3% basta. To investigate2 the role of the CmDFRs in grown on 1/2 Murashige and Skoog (MS) medium containing 1% sucrose (Figure8A). Transformation anthocyanin accumulation, wild-type seeds, mutant, and T2 transgenic lines were germinated and with CmDFR-OB failed to restore the mutant phenotypes of tt3-1, including a lack of proanthocyanidins grown on 1/2 Murashige and Skoog (MS) medium containing 1% sucrose (Figure 8A). in seeds and anthocyanins at the junction of the cotyledon and hypocotyl. However, transformation with Transformation with CmDFR-OB failed to restore the mutant phenotypes of tt3-1, including a lack of CmDFR-RM restored these mutant phenotypes, as the transgenic plants accumulated proanthocyanidins proanthocyanidins in seeds and anthocyanins at the junction of the cotyledon and hypocotyl. in seeds and anthocyanins at the junction of the cotyledon and hypocotyl. However, transformation with CmDFR-RM restored these mutant phenotypes, as the transgenic plants accumulated proanthocyanidins in seeds and anthocyanins at the junction of the cotyledon and hypocotyl.

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 9 of 16

At the mature stages, we performed the RT-PCR to check the presence and expression of the CmDFR gene (Figure 8B). It revealed that the CmDFR gene was successfully expressed in all the transgenic Arabidopsis plants. Furthermore, we extracted pigments from whole leaves of wild-type, tt3-1, and tt3-1 plants complemented with CmDFRs (Figure 8C). The anthocyanin content of tt3-1 was half that of wild-type plants. The anthocyanin contents of complemented tt3-1 plants harboring CmDFR-RM were similar to that of the wild-type plants. By contrast, the anthocyanin content of tt3- 1 plants complemented with CmDFR-OB was not restored. These results suggest that CmDFR-RM functions in proanthocyanidins biosynthesis in seeds and anthocyanins biosynthesis in leaves and Int. J. Mol. Sci. 2020, 21, 7960 9 of 17 stems.

FigureFigure 8. 8.Overexpression Overexpression of ofCmDFR-OB CmDFR-OBand andCmDFR-RM CmDFR-RMin in Arabidopsis Arabidopsistt3-1 tt3-1mutants. mutants. ( A(A)) Seeds Seeds tt3-1 (top)(top) and and 3-day-old 3-day-old seedlings seedlings (bottom) (bottom) of of wild-type wild-type Col-0, Col-0, tt3-1, and, and representative representative T 2T2progeny progeny of of homozygoushomozygous Arabidopsis Arabidopsistt3-1 tt3-1lines lines transformed transformed with withCmDFR-OB CmDFR-OBand andCmDFR-RM CmDFR-RM,, respectively.respectively. WhiteWhite bar bar indicates indicates 0.5 0.5 mm mm in in size; size; (B ()B Expression) Expression analysis analysis of ofCmDFR CmDFRin in the the leaves leaves of wild-type,of wild-type,tt3-1 tt3-, and1, and transgenic transgenic plants plants by RT-PCR.by RT-PCR.AtEF1 AtEF1α wasα was used used as the as referencethe reference gene; gene; (C) Anthocyanin(C) Anthocyanin contents contents of four-week-old Arabidopsis plants. Results represent mean values SD from three biological replicates. of four-week-old Arabidopsis plants. Results represent mean± values ± SD from three biological Direplicates.fferent letters Different above letters the bars above indicate the bars significantly indicate significantly different valuesdifferent (p values< 0.0001) (p < calculated0.0001) calculated using one-wayusing one-way ANOVA ANOVA followed followed by a Duncan’s by a Duncan’s multiple multiple range tests. range tests. At the mature stages, we performed the RT-PCR to check the presence and expression of the CmDFR 3. Discussion gene (Figure8B). It revealed that the CmDFR gene was successfully expressed in all the transgenic ArabidopsisDFR is a plants. NADPH-dependent Furthermore, reductase we extracted that conver pigmentsts dihydroflavonols from whole leaves such of as wild-type, DHK, DHQ,tt3-1 and, andDHMtt3-1 intoplants leucoanthocyanidins complemented with suchCmDFRs as leucopelar(Figure8gonidin,C). The anthocyaninleucocyanidin, content and ofleucodelphinidin,tt3-1 was half thatrespectively. of wild-type DFR plants. proteins The anthocyanin contain contents the hi ofghly complemented conservedtt3-1 NADP(H)plants harboring bindingCmDFR-RM domain were“VTGAAGFIGSWLIMRLLERGY” similar to that of the wild-type and plants. a substrate By contrast, binding domain. the anthocyanin The latter content domain of istt3-1 dividedplants into complementedthree types including with CmDFR-OB the Asn type,was notAsp restored. type, and These non-Asn/Asp results suggest type, which that CmDFR-RM differ in the functions amino acid in proanthocyanidinsresidue at the 134th biosynthesis position. Due in seeds to the and differen anthocyaninst preferences biosynthesis of DFRs in for leaves dihydroflavonols, and stems. these enzymes affect the biosynthesis of anthocyanin metabolites. Asn-type DFRs can use all three 3.dihydroflavonols Discussion (DHK, DHQ, and DKM) as substrates, whereas Asp-type DFRs cannot use DHK DFR is a NADPH-dependent reductase that converts dihydroflavonols such as DHK, DHQ, and DHM into leucoanthocyanidins such as leucopelargonidin, , and , respectively. DFR proteins contain the highly conserved NADP(H) binding domain “VTGAAGFIGSWL IMRLLERGY” and a substrate binding domain. The latter domain is divided into three types including the Asn type, Asp type, and non-Asn/Asp type, which differ in the amino acid residue at the 134th position. Due to the different preferences of DFRs for dihydroflavonols, these enzymes affect the biosynthesis of anthocyanin metabolites. Asn-type DFRs can use all three dihydroflavonols (DHK, DHQ, and DKM) as substrates, whereas Asp-type DFRs cannot use DHK efficiently [8–10]. For example, Int. J. Mol. Sci. 2020, 21, 7960 10 of 17 as Asp-type DFRs from petunia and cymbidium cannot catalyze the conversion of DHK efficiently, they fail to produce brick red flowers in nature [8,9]. DFR is the key flavonoid biosynthetic enzyme contributing to the accumulation of pigments including anthocyanin and proanthocyanidin. In this study, we purified the two recombinant CmDFRs from bacterial expression (Figure5) and demonstrated their DFR activities (Figure6). During the process of protein purification, unlike the recombinant CmDFR-OB, the majority of recombinant CmDFR-RM protein was detected in the soluble fraction, which appeared to be associated with the C-terminal region of CmDFR-RM containing multiple hydrophilic residues (-SSSSKERT-). Although it cannot be ascertained whether these variations at the C-terminus of the two DFRs directly or indirectly affect the enzyme activity, at least the C-terminal hydrophilic region appears to affect the solubility of DFR protein. CmDFR-OB and CmDFR-RM both contain the 134th Asn residue, which enable the of all three dihydroflavonol substrates including DHK. The result of in vitro assay showed that both enzymes produced leucopelargonidin and leucocyanidin from DHK and DHQ, respectively. Leucodelphinidin production was very scare or absent both of CmDFR-OB and CmDFR-RM, but it needs to expand its detection range to confirm. The enzymes both showed similar activities for leucocyanidin production. Interestingly,there was a striking difference in the production of leucopelargonidin. CmDFR-OB showed a dramatically reduced activity as compared with CmDFR-RM on leucopelargonidin production. This remarkable decrease in the activity of CmDFR-OB can be assumed to be due to the C-terminal truncation, suggesting that the variable C-terminus of DFR may be an important region conferring structural stability of binding pocket, especially near the B-ring . A recent study showed that MaDFR harboring a point mutation at the 134th or 145th residue was still able to catalyze reactions using all three dihydroflavonol substrates [17], which indicated that the 134th or 145th residue was not an absolute factor for substrate specificity of DFR. Therefore, it can be implied that the C-terminal variable region of DFR may be one of the important factors that determine substrate specificity of DFR. Differences in the C-terminal region of DFR proteins have been associated with different substrate preferences and enzyme activities [10,18]. In poplar, two DFR proteins, PtrDFR1 and PtrDFR2, belonging to the Asp type DFRs, exhibited the variability at the C-terminus. The tobacco flower color change, due to anthocyanin accumulation, was observed in the ectopic expression of PtrDFR1, but not in that of PtrDFR2. In freesia, two DFR proteins (FeDFR1 and FeDFR2) classified into the same Asn type with high sequence similarity, but they exhibited the C-terminal variation each other. In vitro enzyme assay showed that FeDFR2 was able to convert DHQ into leucocyanidin, while FeDFR1 had no catalytic activity on DHQ. Taken together these results suggest that the variable C-terminus of DFR plays an important role in substrate preference and enzyme activity. Several studies have indicated that anthocyanin biosynthesis takes place within a metabolon complex consisting of CHS, F3H, F30H, DFR, and ANS, in which DFR might interact with ER-bounded cytochrome P450 to direct the metabolic flux towards anthocyanin biosynthesis [19,20]. In the metabolon complex, the F30H and flavonoid 30,50-hydroxylase (F3050H) enzymes catalyze B ring hydroxylation. The sequential interaction between F30H and DFR or between F3050H and DFR results in the accumulation of cyanidin- and delphinidin-derived anthocyanins, respectively. Additionally, substrate channeling between DFR and FLS for anthocyanin and flavonol biosynthesis contributes to color pattern formation in flowers [4,21]. Several reports have indicated that ray florets of chrysanthemum with bronze, pink, or purplish-red coloration accumulate only cyanidin derivatives, whereas white ray florets do not accumulate any anthocyanins [14,15,22]. In the current study, we detected anthocyanins in the red-flowered cultivar RM but not in the white-flowered cultivar OB, indicating that anthocyanins were responsible for the coloration of these flowers. Additionally, we analyzed the expression of anthocyanin biosynthetic genes in ray floret of OB and RM to better understand anthocyanin biosynthetic mechanism in chrysanthemum (Figure7). Our study showed that most flavonoid biosynthetic genes, except for CmCHS, were expressed at dramatically low levels in all floral developmental stage of OB as compared Int. J. Mol. Sci. 2020, 21, 7960 11 of 17 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 11 of 16 detectedwith that in in OB RM. during Interestingly, all flower the transcriptdevelopmental level ofstages.CmF3H Althoughamong EBGall of was LBG hardly at FS1 detected in OB inwas OB expressedduring all at flower a low developmentallevel, it was not stages. able to Althoughaccumulate all anthocyanin of LBG at FS1 in the in OB ray was florets expressed of OB. As at awell low as,level, we itverified was not that able CmDFR-OB to accumulate can anthocyanincatalyze the inDHK the rayand florets DHQ ofas OB.substrates As well with as, we in verifiedvitro assay that (FigureCmDFR-OB 6). Taken can catalyze together the these DHK re andsults DHQ suggest as substrates that the lack with ofin CmF3H vitro assay transcripts (Figure6 ).may Taken lead together to the deficientthese results of DHK suggest and that block the lackoff oftheCmF3H metabolictranscripts flow toward may lead anthocyanin to the deficient biosynthesis of DHK and in blockOB. Additionally,off the metabolic it is thought flow toward that the anthocyanin high activities biosynthesis of F3′H can in fully OB. Additionally,convert DHK itinto is thoughtDHQ, resulting that the inhigh the activitiespredominant of F3 0accumulationH can fully convert of B-ring DHK dihydr into DHQ,oxylated resulting metabolites, in the predominantwhich lead to accumulation accumulate cyanidin-basedof B-ring dihydroxylated anthocyanins metabolites, and quercetin which derivati lead toves accumulate [14,15]. On cyanidin-based the basis of anthocyaninsthese results, andwe proposequercetin the derivatives anthocyanin [14 biosynthetic,15]. On the basispathway of these in ra results,y florets we of proposechrysanthemum, the anthocyanin as shown biosynthetic in Figure 9.pathway in ray florets of chrysanthemum, as shown in Figure9.

FigureFigure 9. 9. TheThe proposed proposed anthocyanin anthocyanin biosynthetic biosynthetic pathway pathway in the in ray the florets ray florets of chrysanthemum. of chrysanthemum. The boldThe boldblack black lines linesindicate indicate the thebiosynthetic biosynthetic pathwa pathwayy of ofcyanidin-based cyanidin-based anthocyanins anthocyanins that that are are only only accumulatedaccumulated in in the the ray ray florets florets of ofChrysanthemum Chrysanthemum flowers. flowers. The Theabbreviations abbreviations of enzyme of enzyme names names are asare follows: as follows: CHS, CHS,chalcone chalcone synthase; synthase; CHI, chalcone CHI, chalcone isomerase; isomerase; F3H, F3H, flavanoneone 3-hydroxylase; 3-hydroxylase; FLS, flavonolFLS, flavonol synthase; synthase; F3′H, F3 0flavonoidH, flavonoid 3′-hydroxylase; 30-hydroxylase; F3′ F35′H,050 H,flavonoid flavonoid 3′ 3,50,5′-hydroxylase,0-hydroxylase, DFR, DFR, dihydroflavonoldihydroflavonol 4-reductase; 4-reductase; ANS, ANS, anthocyanidi anthocyanidinn synthase; synthase; and andUFGT, UFGT, UDP-glucose: UDP-glucose: flavonoid flavonoid 3-O- glucosyltransferase.3-O-glucosyltransferase.

TheThe inin planta planta assayassay confirmed confirmed that that the the ectopic ectopic expression expression of of CmDFR-RMCmDFR-RM restoredrestored the the phenotype phenotype ofof Arabidopsis Arabidopsis tt3-1tt3-1 toto wild wild type, type, and and therefore therefore this this muta mutantnt accumulated accumulated proanthocyanidin proanthocyanidin in in its its seedsseeds and and anthocyanin anthocyanin in in its its hypocotyls hypocotyls and and matu maturere leaves. leaves. However, However, the the ectopic ectopic expression expression of of CmDFR-OBCmDFR-OB did notnot restorerestore the the accumulation accumulation of proanthocyanidinof proanthocyanidin and and anthocyanin anthocyanin to wild-type to wild-type levels. levels.In wild-type In wild-type Arabidopsis Arabidopsis seedling, seedling, cyanidin andcyanid pelargonidinin and pelargonidin accumulate accumulate at comparable at levelscomparable and the levelsmajor and flavonol the major is kaempferol, flavonol[ 23is ]kaempferol, indicating that [23] Arabidopsis indicating F3that0H Arabidopsis capacity only F3 accepts′H capacity a fraction only of acceptsthe total a flavonoidfraction of metabolites. the total flavonoid Therefore, metabolites. the failure ofTherefore,tt3-1 restoration the failure by CmDFR-OB of tt3-1 restoration suggests thatby CmDFR-OBpelargonidin, suggests which accountsthat pelargonidin, for about halfwhich of accounts anthocyanins, for about was barelyhalf of producedanthocyanins, in the was transgenic barely producedlines due in to the dramatically transgenic decreased lines due DHKto dramatically acceptance decreased of CmDFR-OB. DHK acceptance of CmDFR-OB. TakenTaken together, ourour resultsresults indicate indicate that that the the diff dierencefference of DFR of DFR enzyme enzyme activity activity and lack and of lackCmF3H of CmF3Htranscripts transcripts may contribute may contribute to the anthocyanin to the anthocya accumulationnin accumulation of chrysanthemum. of chrysanthemum. Further studies Further on studiesthe interactions on the interactions between metabolon between metabolon enzyme complex enzyme in complex chrysanthemum in chrysanthemum would provide would additional provide additionalinsights into insights anthocyanin into anthocyanin biosynthesis biosyn in thethesis ray floretsin the ray of this florets crop. of this crop.

Int. J. Mol. Sci. 2020, 21, 7960 12 of 17

4. Materials and Methods

4.1. Plant Materials Chrysanthemum plants were grown in a greenhouse under short day conditions at the National Institute of Horticultural and Herbal Science (Wanju, Korea). Two chrysanthemum cultivars named as OhBlang (OB) with white flower and RedMarble (RM) with red flower, respectively were used to analyze anthocyanin levels and gene expression. For the InDel marker analysis, we used twenty-two chrysanthemum cultivars. These cultivars are categorized based on their ray floret color as follows: white (BaekGang (BG), JimMa (JM), WonGyo 184 (WG184), BaekMa (BM), UnBaek (UB), BaekSeol (BS), JimMa2 (JM2), OB, SnowDream (SD), 13-62, AnGel (AG), 13-61, MoYa (MY), and PureAngel (PA)); pink (FreeMadona (FM); DonaPink (DP), CherryBlossom (CB), and PinkPride (PP)); and red (13-86, 12-71, BlackMarble (BM), and RM). Transformation experiments were conducted using Arabidopsis thaliana transparent testa (tt) dfr mutant line tt3-1, which was obtained from the Arabidopsis Biological Resource Center (ABRC). All Arabidopsis plants were grown on 1/2 MS medium containing 1% sucrose or in soil under long-day conditions (LD, 16 h light/8 h dark) at 22 ◦C. All samples were frozen rapidly in liquid nitrogen and kept at 80 C. A portion of the samples − ◦ was used for RNA extractions and anthocyanin measurements.

4.2. CmDFR Gene Cloning and Sequence Analysis The full-length ORF (open reading frame) of CmDFR was identified from the chrysanthemum genome database based on Chrysanthemum boreale and was predicted using FGENESH (http://www. softberry.com/berry.phtml?topic=fgenesh&group=programs&subgroup=gfind). Genomic DNA was obtained from chrysanthemum leaves using a DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA), according to the manufacturer’s instructions. Total RNA was extracted from ray florets of flowers at different developmental stages from two different chrysanthemum cultivars OB and RM using Fruit-mate for RNA Purification solution (Takara, Otsu, Japan) and Plant RNA Purification Reagent (Invitrogen, Carlsbad, CA, USA), as described previously [24]. The total RNA was purified using a FavorPrep™ Plant Total RNA Mini Kit (Favorgen, Changzhi, Taiwan), according to the manufacturer’s instructions. cDNA was synthesized from 2 µg of total RNA using amfiRivert cDNA Synthesis Platinum Master Mix (GenDEPOT, Barker, TX, USA). The CmDFR gene was amplified from cDNA and genomic DNA by PCR with PrimeSTAR® HS DNA Polymerase (Takara) and the primer set CmDFR-F/R. All PCR fragments were subcloned into the pENTR/D-TOPO vector (Invitrogen) to validate the DNA sequences. All primer sequences are listed in Supplementary Table S1. Nucleotide sequence, deduced amino acid sequence, and ORF of CmDFR were determined online (http://www.ncbi.nlm.nih.gov). Structural analysis of the deduced protein was carried out using the ExPASy Molecular Biology Server (http://cn.expasy.org/tools/). Multiple sequence alignments were performed using the CLUSTALW program (https://www.genome.jp/tools-bin/clustalw). A phylogenetic tree was constructed using the neighbor-joining method [25] with MEGA version 6 software.

4.3. InDel Analysis Genomic DNA was extracted from the leaves of the 22 chrysanthemum cultivars using a DNeasy Plant Mini Kit (Qiagen), following the manufacturer’s instructions. A pair of PCR primers was designed to detect the 7-bp deletion in the sixth exon of CmDFR (Supplementary Figure S1). The PCR mixture contained 100 ng of genomic DNA, 5 pmol of each primer, 10 pmol of dNTPs, and 1 unit of PrimeSTAR HS DNA Polymerase in 1 PrimeSTAR Buffer (Takara) in a total volume of 25 µL. The PCR conditions × were as follows: An initial denaturation at 98 ◦C for 2 min, followed by 30 cycles of denaturation at 98 ◦C for 10 s, annealing at 60 ◦C for 10 s, and an extension at 72 ◦C for 20 s, and a final incubation at 72 C for 5 min. The PCR products were separated in a 5% nondenaturing polyacrylamide gel in 1 ◦ × Int. J. Mol. Sci. 2020, 21, 7960 13 of 17

TBE buffer (90 mM Tris-borate, 2 mM EDTA, pH 8.0). Following electrophoresis, the DNA fragments were visualized by silver staining, as previously described [26].

4.4. Expression and Purification of Recombinant CmDFR The CmDFR ORFs were amplified using primer sets designed as shown in Supplementary Table S1. The PCR products were inserted into the pGEX-6P-1 vector linearized by BamHI digestion using an In-Fusion Advantage PCR Cloning Kit (Clontech, Mountain View, CA, USA) in-frame with the sequence encoding the N-terminal glutathione S-transferase (GST) tag. The resulting pGEX-6P-1:CmDFR vectors were verified by sequencing and transformed into Escherichia coli strain BL21 (DE3) cells (Novagen, Darmstadt, Germany). Transformed bacterial cells were cultured in 50 mL of LB broth containing 50 µg mL–1 ampicillin. Protein expression was induced by the addition · of 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 20 ◦C for 20 h. After centrifugation, the collected bacterial cells were lysed by sonication in 2 mL of sonication buffer containing 50 mM sodium phosphate (pH 8.0), 150 mM NaCl, 1% Triton X-100, 10% glycerol, 2 mM dithiothreitol (DTT), and 1 mM phenylmethylsulfonyl fluoride. Following centrifugation (13,000 g, 4 C, 10 min), 100 µL × ◦ of Glutathione Sepharose 4B beads (GE Healthcare, Pittsburgh, PA, USA) were added to the soluble bacterial lysate and incubated at 4 ◦C for 2 h, with gentle rotation. The GST-tagged protein-bound beads were collected and washed five times with 1 PBS (137 mM NaCl, 2.7 mM KCl, 100 mM Na HPO , × 2 4 and 2 mM K2HPO4 (pH 7.4)). For GST cleavage, a mixture of 4 µL of PreScission protease (GenScript, Piscataway, NJ) and 96 µL of cleavage buffer (50 mM Tris-HCl (pH 7.0), 150 mM NaCl, 1 mM EDTA, and 1 mM DTT) was added to the beads and incubated for 3 h, at 4 ◦C. Following incubation, the beads were pelleted by centrifugation at 500 g for 3 min and the eluate was transferred to a new × tube. The beads were washed with two bed volume of cleavage buffer, and the washing solution was combined with the eluate. The concentrations of the GST-cleaved CmDFR proteins were equalized to 1.1 µg/µL by adding cleavage buffer, and the same volume of glycerol was added to obtain a final concentration of 0.55 µg/µL. These were used for assay and the rest were stored in 20 C. The purified − ◦ untagged proteins in the eluate were quantified by the Bradford method and verified by SDS-PAGE.

4.5. In Vitro Assay for CmDFR Activity The enzymatic activities of the CmDFRs were assayed in a 200 µL reaction containing 3.3 µg of GST-cleaved CmDFR protein, 100 mM Tris-HCl (pH 7.5), 1 mM glucose 6-phosphate, 0.75 mM nicotinamide adenine dinucleotide phosphate reduced, 1 unit of glucose 6-phosphaste dehydrogenase, and 1 mM of dihydroflavonol substrates. The reactions were initiated by adding dihydroflavonol substrates and incubated at 30 ◦C for 50 min, and 10 µL acetic acid was added to stop the reaction. The reactants were extracted twice with 500 µL of ethyl acetate and dried using nitrogen gas. The products of the reactions were unstable leucoanthocyanidins; therefore, the residues were dissolved in 100 µL of acidic alcohol, n-butanol:HCl (95:5, v/v) to generate the corresponding stable anthocyanidins. Following incubation at 95 ◦C for 30 min, 10 µL of the solution was analyzed by high-performance liquid chromatography (HPLC).

4.6. HPLC Analysis HPLC analysis was conducted on an LC-20A HPLC system with a diode array detector (Shimadzu, Kyoto, Japan). The separation of dihydroflavonol substrates and anthocyanidin products was accomplished on an Inertsil-ODS3 C18 column (5 µm, 250 4.6 mm, GL Science). The mobile × phase consisted of 0.1% (v/v) formic acid (A) and acetonitrile containing 0.1% (v/v) formic acid (B). The gradient profile was optimized as follows: 0 min, 95% A/5% B; 30 min, 45% A/55% B; 45 min, 35% A/65% B; 50 min, 0% A/100% B; 52 min, 95% A/5% B; and 60 min, 95% A/5% B. The flow rate was 1 mL min 1, and the column temperature was maintained at 30 C. The detection wavelength was · − ◦ 288 nm for dihydroflavonols and 520 nm for anthocyanidins. Int. J. Mol. Sci. 2020, 21, 7960 14 of 17

4.7. Chemical Standards ( )-DHK, ( )-DHQ, and ( )-DHM were purchased from Sigma-Aldrich (St. Louis, MO, USA), and ± ± ± pelargonidin chloride, cyanidin chloride, and delphinidin chloride were purchased from Extrasynthese (Extrasynthese, Genay Cedex, France). The dihydroflavonols were prepared as 100 mM stock solutions in DMSO, and the anthocyanidins were prepared as a 100 mM solution in 50% methanol containing 1.2 N HCl.

4.8. Measurement of Total Anthocyanin Contents Total anthocyanin contents were measured from ray florets of flowers at the following different developmental stages: FS1, 6.0 weeks under short-day (SD) treatment, on which the first coloration of ray florets was first visible; FS2, 7.0 week under SD treatment that is the date of early flowering; and FS3, 7.5 weeks under SD treatment that is the date of full flowering. As described by [24], powdered ray floret samples were incubated in 600 µL extraction buffer (methanol containing 1% HCl) for 6 h, at 4 ◦C, with moderate shaking. Following the addition of 200 µL water and 200 µL chloroform, the samples were centrifuged at 14,000 g for 5 min at 4 C to sediment the plant material. The absorbance of the × ◦ supernatant was recorded at 530 nm (A530) and 657 nm (A657) using a microplate reader. Anthocyanin content was determined using the following equation: A 0.33 A . The anthocyanin content in 530 − × 657 each sample was measured in three independent experiments.

4.9. Quantitative Reverse-Transcription PCR (qPCR) Total RNA was prepared from ray florets, as described above. cDNA was synthesized from 2 µg of total RNA using amfiRivert cDNA Synthesis Platinum Master Mix (GenDEPOT). qPCR was performed using AccuPower 2x Greenstar qPCR Master Mix (Bioneer, Daejun, Korea) and a Bio-Rad CFX96 Detection System (Bio-Rad Laboratories, Hercules, CA, USA), according to the manufacturer’s instructions. Gene expression was normalized using the elongation factor 1α (EF1α) as the reference gene. The gene-specific primers used for qPCR analysis are listed in Supplementary Table S1. Three biological replicates were performed per sample.

4.10. In Planta Assay of CmDFR Function The plasmid used for stable transformation of Arabidopsis was constructed as follows: The ORFs of CmDFR isolated from cultivars OB (CmDFR-OB) and RM (CmDFR-RM) were subcloned into the pENTR/D-TOPO vector (Invitrogen) and incorporated into the Gateway destination vector pB7WG2D (VIB-Ghent University, Ghent, Belgium) via several Gateway cloning steps. The resulting vector was maintained in Agrobacterium tumefaciens strain GV3101 and transformed into the Arabidopsis tt3-1 mutant using the floral dip method. Transformed Arabidopsis seeds were grown in soil under LD conditions at 22 ◦C. Transgenic Arabidopsis plants were selected by spraying the plants with 0.3% basta solution. Homozygous T2 lines were subjected to evaluate expression level of exogenous CmDFR gene and phenotypic investigation. Arabidopsis elongation factor 1α (EF1α) gene was used as internal reference.

4.11. Statistical Analysis For qPCR analysis, results were represented as mean values SD from three biological replicates. ± For the analysis of anthocyanin contents, experimental data were presented as mean values SD of ± three biological replicates. Statistical significance was determined by one-way ANOVA followed by a Duncan’s multiple range tests. Int. J. Mol. Sci. 2020, 21, 7960 15 of 17

5. Conclusions In this study, we characterized the DFR derived from different ray floret colored OB and RM chrysanthemum cultivars. CmDFR2-OB contains a premature stop codon due to a single base substitution. Differences at the C-terminus of CmDFR proteins affect their catalytic efficiency, as well as protein solubility. Lack of CmF3H gene expression was observed in OB during all flower developmental stages. Complementation tests indicated that the C-terminal variation in CmDFR also affects anthocyanin levels in leaves and proanthocyanidin biosynthesis in seeds, as revealed in transgenic Arabidopsis. Taken together, these results provide a new insight to better understand the anthocyanin biosynthesis in chrysanthemum.

Supplementary Materials: Supplementary Materials can be found at http://www.mdpi.com/1422-0067/21/21/ 7960/s1, Table S1: Primer sequences used in this study. Author Contributions: S.-H.L. and J.-Y.L. conceived and designed the research; B.P., D.-H.K., S.P., J.-H.Y., J.-A.J. and J.L. conducted experiments; S.-H.L. and S.P. drafted the manuscript. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by funds from the National Institute of Agricultural Science (PJ014838) and a grant from the Next-Generation BioGreen 21 Program (PJ013360/PJ013346), Rural Development Administration, Korea. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations

ANS Anthocyanidin synthase CHI Chalcone isomerase CHS Chalcone synthase DFR Dihydroflavonol 4-reductase F3H Flavanone 3-hydroxylase F30H Flavonoid 30-hydroxylase F3050H Flavonoid 3050-hydroxylase FLS Flavonol synthase InDel marker Insertion and deletion marker qPCR Quantitative reverse-transcription PCR RT-PCR Reverse-transcription PCR UFGT UDP-glucose: flavonoid-3-O-glucosyltransferase.

References

1. Grotewold, E. The genetics and biochemistry of floral pigments. Annu. Rev. Plant Biol. 2006, 57, 761–780. [CrossRef][PubMed] 2. Lin-Wang, K.; Bolitho, K.; Grafton, K.; Kortstee, A.; Karunairetnam, A.; McGhie, T.K.; Espley,R.V.;Hellens, R.P.; Allan, A.C. An R2R3 MYB transcription factor associated with regulation of the anthocyanin biosynthetic pathway in Rosaceae. BMC Plant Biol. 2010, 10, 50. [CrossRef][PubMed] 3. Lim, S.H.; Ha, S.H. Marker development for identification of rice seed coat color. Plant Biotechnol. 2013, 7, 391–398. [CrossRef] 4. Luo, P.; Ning, G.; Wang, Z.; Shen, Y.; Jin, H.; Li, P.; Huang, S.; Zhao, J.; Bao, M. Disequilibrium of flavonol synthase and dihydroflavonol-4-reductase expression associated tightly to white vs. red color flower formation in plants. Front. Plant Sci. 2016, 6, 1257. [CrossRef][PubMed] 5. Park, S.; Kim, D.H.; Park, B.R.; Lee, J.Y.; Lim, S.H. Molecular and functional characterization of Oryza sativa FLAVONOL SYNTHASE (OsFLS), a bifunctional dioxygenase. J. Agric. Food Chem. 2019, 67, 7399–7409. [CrossRef] 6. Park, S.; Kim, D.H.; Yang, J.H.; Lee, J.Y.; Lim, S.H. Increased flavonol levels in tobacco expressing AcFLS affect flower color and root growth. Int. J. Mol. Sci. 2020, 21, 1011. [CrossRef] Int. J. Mol. Sci. 2020, 21, 7960 16 of 17

7. Gosch, C.; Nagesh, K.M.; Thill, J.; Miosic, S.; Plaschil, S.; Milosevic, M.; Olbricht, K.; Ejaz, S.; Rompel, A.; Stich, K.; et al. Isolation of Dihydroflavonol 4-Reductase cDNA Clones from Angelonia x angustifolia and Het-erologous Expression as GST Fusion Protein in Escherichia coli. PLoS ONE 2014, 9, e107755. [CrossRef] 8. Haselmair-Gosch, C.; Miosic, S.; Nitarska, D.; Roth, B.L.; Walliser, B.; Paltram, R.; Lucaciu, R.C.; Eidenberger, L.; Rattei, T.; Olbricht, K.; et al. Great cause—Small effect: Undeclared genetically engineered Orange Petunias Harbor an inefficient Dihydroflavonol 4-Reductase. Front. Plant Sci. 2018, 9, 149. [CrossRef] 9. Johnson, E.T.; Yi, H.; Shin, B.; Oh, B.J.; Cheong, H.; Choi, G. Cymbidium hybrida dihydroflavonol 4-reductase does not efficiently reduce dihydrokaempferol to produce orange pelargonidin-type anthocyanins. Plant J. 1999, 19, 81–85. [CrossRef] 10. Li, Y.; Liu, X.; Cai, X.; Shan, X.; Gao, R.; Yang, S.; Han, T.; Wang, S.; Wang, L.; Gao, X. Dihydroflavonol 4-reductase genes from Freesia hybrida play important and partially overlapping roles in the biosynthesis of flavonoids. Front. Plant Sci. 2017, 8, 428. [CrossRef] 11. Liu, X.; Xiang, M.; Fan, Y.; Yang, C.; Zeng, L.; Zhang, Q.; Chen, M.; Liao, Z. A root-preferential DFR-Like gene encoding dihydrokaempferol reductase involved in anthocyanin biosynthesis of purple-fleshed sweet potato. Front. Plant Sci. 2017, 8, 279. [CrossRef][PubMed] 12. Bashandy,H.; Pietiäinen, M.; Carvalho, E.; Lim, K.J.; Elomaa, P.; Martens, S.; Teeri, T. Anthocyanin biosynthesis in gerbera cultivar ‘Estelle’ and its acyanic sport ‘Ivory’. Planta 2015, 242, 601–611. [CrossRef][PubMed] 13. Katsu, K.; Suzuki, R.; Tsuchiya, W.; Inagaki, N.; Yamazaki, T.; Hisano, T.; Yasui, Y.; Komori, T.; Koshio, M.; Kubota, S.; et al. A new buckwheat dihydroflavonol 4-reductase (DFR), with a unique substrate binding structure, has altered substrate specificity. BMC Plant Biol. 2017, 17, 1–14. [CrossRef] 14. Park, C.H.; Chae, S.C.; Park, S.Y.; Kim, J.K.; Kim, Y.J.; Chung, S.O.; Arasu, M.V.; Al-Dhabi, N.A.; Park, S.U. Anthocyanin and carotenoid contents in different cultivars of chrysanthemum (Dendranthema grandiflorum Ramat.) flower. Molecules 2015, 20, 11090–11102. [CrossRef][PubMed] 15. Brugliera, F.; Tao, G.Q.; Tems, U.; Kalc, G.; Mouradova, E.; Price, K.; Stevenson, K.; Nakamura, N.; Stacey, I.; Katsumoto, Y.; et al. Violet/blue chrysanthemums—Metabolic engineering of the anthocyanin biosynthetic pathway results in novel petal colors. Plant. Cell Physiol. 2013, 54, 1696–1710. [CrossRef] 16. Schwinn, K.E.; Markham, K.R.; Giveno, N.K. Floral flavonoids and the potential for pelargonidin biosynthesis in commercial chrysanthemum cultivars. Phytochemistry 1993, 35, 145–150. [CrossRef] 17. Liu, H.; Lou, Q.; Ma, J.; Su, B.; Gao, Z.; Liu, Y. Cloning and functional characterization of dihydroflavonol 4-reductase gene involved in anthocyanidin biosynthesis of grape hyacinth. Int. J. Mol. Sci. 2019, 20, 4743. [CrossRef] 18. Huang, Y.; Gou, J.; Jia, Z.; Yang, L.; Sun, Y.; Xiao, X.; Song, F.; Lou, K. Molecular cloning and characterization of two genes encoding dihydroflavonol-4-reductase from Populus trichocarpa. PLoS ONE 2012, 7, e30364. [CrossRef] 19. Fujino, N.; Tenma, N.; Waki, T.; Ito, K.; Komatsuzaki, Y.; Sugiyama, K.; Yamazaki, T.; Yoshida, S.; Hatayama, M.; Yamashita, S.; et al. Physical interactions among flavonoid enzymes in snapdragon and torenia reveal the diversity in the flavonoid metabolon organization of different plant species. Plant. J. 2018, 94, 372–392. [CrossRef] 20. Nakayama, T.; Takahashi, S.; Waki, T. Formation of flavonoid metabolons: Functional significance of protein-protein interactions and impact on flavonoid chemodiversity. Front. Plant Sci. 2019, 10, 821. [CrossRef] 21. Yuan, Y.W.; Rebocho, A.B.; Sagawa, J.M.; Stanley, L.E.; Bradshaw, H.D. Competition between anthocyanin and flavonol biosynthesis produces spatial pattern variation of floral pigments between Mimulus species. Proc. Natl. Acad. Sci. USA 2016, 113, 2448–2453. [CrossRef][PubMed] 22. Chen, S.M.; Li, C.H.; Zhu, X.R.; Deng, Y.M.; Sun, W.; Wang, L.S.; Chen, F.D.; Zhang, Z. The identification of flavonoids and the expression of genes of anthocyanin biosynthesis in the chrysanthemum flowers. Biol. Plantarum 2012, 56, 458–464. [CrossRef] 23. Sun, W.; Liang, L.; Meng, X.; Li, Y.; Gao, F.; Liu, X.; Wang, S.; Gao, X.; Wang, L. Biochemical and molecular characterization of a flavonoid 3-O-glycosyltransferase responsible for anthocyanins and flavonols biosynthesis in Freesia hybrida. Front. Plant. Sci. 2016, 7, 410. [CrossRef][PubMed] 24. Kim, D.H.; Park, S.; Lee, J.Y.; Ha, S.H.; Lim, S.H. Enhancing flower color through simultaneous expression of the B-peru and mPAP1 transcription factors under control of a flower-specific promoter. Int. J. Mol. Sci. 2018, 19, 309. [CrossRef] Int. J. Mol. Sci. 2020, 21, 7960 17 of 17

25. Saitou, N.; Nei, M. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 1987, 4, 406–425. 26. Lim, S.H.; Cho, H.; Lee, S.; Cho, Y.H.; Kim, B.D. Identification and classification of S haplotypes in Raphanus sativus by PCR-RFLP of the S locus glycoprotein (SLG) gene and the S locus receptor kinase (SRK) gene. Theor. Appl. Genet. 2002, 104, 1253–1262. [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).