Complete coding sequence of NtMYB12 contain 819 bases contains the conserved R2R3 domain at the N-terminal end, which is also the DNA binding site of the R2R3 MYB proteins. Three tryptophan residue (W) in R2 domain and one phenylalanine (F) and two tryptophan residue (W) in R3 domain, which is important for formation of the core of helix-turn-helix (HTH) were presented (Yuan et al. 2016). Alignment of NtMYB12 with different R2R3 MYB proteins of Sg7 clade indicated presence of another conserved motif with a consensus sequence SG7 (KRRGGRTSRCTMK) and SG7-2 (WLLE), but there is likely no binding motif of bHLH. MEME software analysis further indicated presence of characteristic signature sequences of the Sg7-2 group of the R2R3 MYB proteins. Apart from C1 and C2 motifs, recently an additional motif known as C3-motif/ZnF-like (zinc finger like) has been identified in the flexible c-terminal of certain activator belonging to R2R3 MYB transcription factors. Similar C3 motif was also located in the C-terminal of NtMYB12 (Supplementary Fig. S1), suggesting its role as a regulator of general flavonol and phenylpropanoid pathway, similar to MdMYB12 related to high content flavonol in red-fleshed apple (Wang et al. 2017). The full length of coding sequence and globally analyzed its evolutionary position in the Fig. 1A.

We further identified NtMYB12 subcellular localization in Arabidopsis and Narcisssus cells by protoplast transit expression assay, and showed that NtMYB12 fusion GFP was localized in the nucleus and had stronger transcription activity (Fig. 1B). By using RT-qPCR, the transcript abundance profile of NtMYB12 in different organs was exhibited that NtMYB12 transcript was lowly expressed in vegetable organ (e.g., root, bulb, leaf, scarp), while it highly expressed in the reproductive organ tepals and corona (Fig. 1C-D); During the period of narcissus plant development from 7 to 21 days, the NtMYB12 expression level was declined especially 10 folds from stage 1 to stage 4 of flower development (Fig. 1E-F), NtMYB12 expression level declined significantly in tepal, corona, tube, and ovary, but increased at stage 4 in anther (Fig. 1G-I), which accompanied with these tissues were colorless, fait yellow, green and white colorization. It seems that this gene is closed related to the color formation in Narcissus tazetta “Jinzhanyintai” tepals.

In order to illustrate NtMYB12 function, loss- of NtMYB12 (NtMYB12-RNAi, Ntmyb12) and gain- of NtMYB12 (oeNtMYB12) constructs were generated. Callus cells of narcissus transformed with the mentioned construct (Fig. 2A) resulted in differentiation of callus firstly in differentiation medium supplemented with kanamycin (25 mg/l), which was detected positive transforms by GUS staining, then emergence of putative embryos on embryo development medium. The embryos appeared globular and lead to generation of white translucent secondary embryos. Emerging embryos converted to small bulb shoots on shoot development medium. Individual bulb shoots were numbered and then multiplied on shoot multiplication medium. Rooting of putatively transformed shoots was carried out on rooting medium and later transplanted in the green house (Patent CN202210162166.8) (Fig. 2B). Reverse transcript semi-quantitative PCR and RT-qPCR were used to proof Ntmyb12 knockdown and overexpressing NtMYB12 transforms. The transcript level of Ntmyb12 could not detect in the gel by semi-RT-qPCR, while stronger signal appeared in oeNtMYB12 and weaker signal appeared in wildtype. NtACTIN was used as an inner control (Fig. 2C-D).

To investigate the function of NtMYB12 on downstream nuclear genetic reprogramming, gain- of function or loss- of function of NtMYB12 tissues (Fig. 3A) were collected to perform genome-wide transcriptome sequencing. Total 26,480 unigenes (> 1 kb) were collected. Three replicates were got low variant (r2 = 0.97) (Supplementary Fig. S2). Comparison of oeNtMYB12 or Ntmyb12 reprograming transcriptome relative to wild-type plants enriched 497 DEGs (ICFI > = 2, P value < 0.01) with 340 transcripts upregulated and 157 transcripts downregulated (Supplemental Dataset S1), or 876 DEGs (ICFI > = 2, P value < 0.01) with 320 transcripts upregulated and 357 transcripts downregulated, respectively (Supplementary Dataset S2). The DEGs (ICFI > = 2, P value < 0.01) of Ntmyb12 relative to WT were subjected to MapMan analysis (Thimm et al., 2004). The biological processes with the most significant alteration were identified that protein metabolic process and cellular process particularly, protein catalytic activity and protein binding were the most significant terms, closely followed by biological regulation and response to stimulus, particularly, protein transporter activity, then was followed by localization and cellular component organization or biogenesis (Supplementary Fig. S2A), which did not surprisingly consider that metabolite biogenesis was affected by NtMYB12, similar to the closet homologous of MdMYB12 related to high content flavonol in red-fleshed apple (Wang et al. 2017).

To address the DEGs enrichment pathway of biological process and molecular function, these DEGs were further analyzed using KEGG pathway. Expectedly, the most significant downregulated pathway was photosynthesis, phenylpropanoid biosynthesis, flavonoid biosynthesis pathways; the second significant down-enrichment pathway was plant hormone signal transduction pathway and porphyrin and chlorophyll metabolism pathways in the Ntmyb12 line relative to WT (Fig. 3B; Figure S2B). In contrast, the most significant upregulated pathway was photosynthesis, phenylpropanoid biosynthesis, flavonoid biosynthesis pathways, the second significant enrichment pathway was plant hormone signal transduction pathway and porphyrin and chlorophyll metabolism pathways in the oeMYB12 line relative to WT (Fig. 3C; Figure S2C).

To characterize up- and downregulated genes of phenylpropanoid biosynthesis and flavonoid biosynthesis pathways in the Ntmyb12 line (Fig. 3A), these two subsets of genes were selected and analyzed using heatmap and MapMan metabolite pathway matching (Fig. 3B). Intriguingly, 4CL (c107941; c113409), C4H (c113570), CHS (c108532), CHI (c92629), FLS (c108471), LAR (c93779), and CCoMAOT (c91328) locus encoding enzymes of flavonoid biosynthesis pathway were significantly upregulated in oeNtMYB12 line, but downregulated in Ntmyb12 line. However, 4CL (c113374), AOC3 (c102005), BPBT (c92663), HCT (c95028), CCoMAOT (c96475) locus encoding enzymes of phenylpropanoid biosynthesis pathway were downregulated in Ntmyb12 line, and upregulated in oeNtMYB12 line (Fig. 3D, Supplementary Fig. S3). In order to focus on the pigment formation related flavonol and anthocyanin pathway, the RT-qPCR further confirmed the alteration of transcript level of the selected genes in oeNtMYB12 and Ntmyb12 mutant lines relative to the wild-type (Fig. 3E). Gain- or loss- of NtMYB12 up- or down-regulated NtF3’H, NtFLS, NtLAR gene expression, overexpressing NtMYB12 also upregulated NtANS gene expression in contract, knockdown Ntmyb12 downregulated their expression, however, knockdown Ntmyb12 did not detect NtANS transcript level, the transcript level of NtDFR was extremely low and could not detect its alteration in overexpressing NtMYB12 or Ntmyb12 callus (Fig. 3E). We further detected the transcript level of NtANS and NtDFR from transiently transformed narcissus petal protoplasts. The transcript level of NtANS was significantly upregulated in transformed oeNtMYB12 petal protoplasts and downregulated in transformed Ntmyb12 petal protoplasts compared to transformed empty vector petal protoplasts. In contract, the transcript level of NtDFR were significantly downregulated in transformed oeNtMYB12 petal protoplast and upregulated in transformed Ntmyb12 petal protoplasts, compared to transformed empty vector petal protoplasts (Supplementary Figure S4). These results suggest that NtMYB12 was involved in regulating flavonoid and proanthocyanin biosynthesis pathways.

The previous work showed NtMYB12 coexpression with NtFLS and NtLAR (Yang et al. 2021). With the petal development from stage 1 to stage 4, petal color accompanies with colorless, yellow, fain yellow, then white, the color pigment flavonoids (Lutin and Maringen) content were firstly up accumulated at stage 1, then went down from stage 2 to stage 4 (Ren et al. 2017; Yang et al. 2021), thus, NtMYB12 might directly regulates NtFLS, NtDFR, NtLAR, and NtANS gene expression in flavonol and (pro)anthocyanin biosynthesis. In order to clarify whether NtMYB12 physically regulated NtFLS, NtDFR, NtLAR, and NtANS gene expression, the promoters of NtFLS, NtDFR, and NtLAR gene were successfully cloned and registered in database by using genome walking assay, since the genome sequence of Narcissus was not created up to now. Expectedly, the promoter sequences of NtFLS, NtLAR, and NtDFR contain many MYB-motives (Table S2). Dual luciferase report assay (Fig. 4A) exhibited that NtMYB12 activated LUC activity in fluorescence imaging and in luciferase activity measurement (Fig. 4B-D) in tobacco leaf cells, 100-folds luciferase ratio of LUC/REN was significantly high in the cotransformed oeNtMYB12 and NtFLS promoter relative to NtFLS promoter alone transform (Fig. 4B). 40-folds luciferase ratio of LUC/REN was significantly high in the cotransformed oeNtMYB12 and NtLAR promoter relative to empty promoter alone (pGreen) transform (Fig. 4C). However, the luciferase ratio of LUC/REN did not significantly change in cotransformed oeMYB12 and NtDFR promoter relative to empty promoter alone (Fig. 4D). It indicated that NtMYB12 physically activated NtFLS and NtLAR gene expression but did not affect NtDFR expression in tobacco leaf cells. Especially, when NtFLS promoter was disturbed to three fragments (Fig. 4E), NtMYB12 can also activate various MYB binding element regions driving LUC gene expression in tobacco leaf cells (Fig. 4F). Then, we performed above dual luciferase activity assay in vivo narcissus tepal protoplast. The same results were exhibited, except oeNtMYB12 repressed NtDFR promoter driving LUC gene expression in narcissus tepal cells (Fig. 4G-H). Further, we detected NtMYB12 directly targeted and regulated NtFLS, NtLAR and NtDFR expression in vivo. The protoplasts of narcissus petal were prepared, collected, and transformed with the oeMYB12-GFP plasmid and alone GFP plasmid by PEG assay. Chromatin immunoprecipitation was performed with antibody against GFP, qPCR product was detectable in the promoter region of NtFLS, NtLAR, and NtDFR genes in IP. Calculation of the enrichment fold of IP relative to INPUT showed that NtMYB12 was significantly enriched in P1, P2, and P3 region of NtFLS gene and P2 region of NtLAR gene, as well as in the P1 region of NtDFR gene, but did not significantly changed in P1 region of NtLAR and P2 region of NtDFR gene. ACTIN gene was used as an inner control; GFP plant was used as an IP control (Fig. 4I-J). It indicates that NtMYB12 physically bound to the promoter of NtFLS, NtDFR, and NtLAR gene and activated NtFLS, and NtLAR expression but repressed NtDFR gene expression in flavonoid biosynthesis.

NtMYB12 exists a conserved motif with a consensus sequence SG7 and SG7-2. This motif within the R3 domain has been suggested to play important role in interaction with the bHLH protein. To test whether, NtMYB12 still have possibility to interact with bHLH protein for their function although it has no typical bHLH binding sites. To this end, we firstly searched for the candidate members of the MYB12-bHLH-WD40 (MBW) complex. Based on the previous transcriptome dataset of Narcissus tazetta (NBCI, PRJNA750844, https://phytozome.jgi.doe.gov) (Yang et al. 2021), the coexpression network of NtMYB12 was constructed by WGCNA, a common method for constructing gene co-expression networks (Peter and Steve 2008), several putative coregulators were exhibited (Fig. 5A). Of them, two candidate genes NtWD40-1 and NtbHLH1 were highly similar with the key regulatory members WD40 and bHLH of MBW complex in Petunia and Arabidopsis. The expression profile of three candidates during tepal development was detected by RT-qPCR and showed a similar expression profile (Supplementary Figure S5).

Next, we detected the interaction among the candidate proteins NtWD40-1, NtbHLH1, and NtMYB12 by yeast two hybrid and bimolecular fluorescence complementation assay. The results showed that yeast AH109 strain cotransformed plasmids AD-NtMYB12 with BD-NtWD40-1 or BD-NtbHLH1 grew well in the medium (SD minus amino acids of LEU, ADE, TRP, and HIS), while colony cotransformed an empty AD vector with BD-NtWD40-1 or BD-NtbHLH1 did not grow there, suggesting that NtMYB12 can interact with NtWD40-1 in yeast cells (Fig. 5B). We further detected the localization of NtMYB12 and NtWD40-1 or NtbHLH1 in the nucleus and cytoplasm of the transformed tobacco leaf epidermal cells (Fig. 5C). Next, we confirmed the interaction of NtWD40-1, NtbHLH1, and NtMYB12 in vivo narcissus protoplast by BiFC. Cotransforming plasmids NtMYB12-VcVN with NtWD40-1 VcVC or NtbHLH1- VcVC with histone 2B fused RFP (H2B-RFP, positive control for nuclear localization) plasmid in narcissus protoplast, mVenus fluorescence signal was appeared in the nucleus (Fig. 5D). However, when the disturbed NtMYB12S1 without R3 domain was cotransformed with NtWD40-1 and NtbHLH1, Venus fluorescence signal was disappearing in the protoplast, while the disturbed NtMYB12S3 that deleted C3 domain of NtMYB12 was cotransformed with NtbHLH1, mVenus fluorescence signal was appeared in both the nucleus and cytoplasm. H2B-RFP was used as a control for transformation efficiency (Fig. 5E). Furthermore, coimmunoprecipitation of cotransformed tobacco leaf with oeNtMYB12-HA, oeNtWD40-1-GFP, oeNtbHLH1-CFP plasmids or with oeNtMYB12-HA and oeNtbHLH1-CFP plasmid using HA antibody showed that NtWD40-1 and NtbHLH1 or NtbHLH1 can been detected in IP by using GFP antibody, suggesting NtMYB12 physically interacted with NtWD40-1 and NtbHLH1 (Fig. 5F). These results indicate that NtMYB12 can interact with NtWD40-1 and NtbHLH-1 to form MBW triple complex via R3 domain.

In angiosperms, anthocyanin biosynthesis is regulated by conserved MYB/bHLH/WD40 (MBW) complexes, which formed by the interaction between MYB transcription factors, basic helix-loop-helix (bHLH) transcription factors, and WD40-repeat proteins (Ramsay et al., 2005; Xu et al. 2015). In Narcissus, whether MBW of MYB12-bHLH-1-WD40-1 is required for (pro)anthocyanin biosynthesis. To test this hypothesis, we detected the MBW activation on NtFLS, NtDFR, and NtLAR expression and metabolites by dual luciferase assay in tobacco leaf cells and narcissus petal protoplast (Fig. 6A). As above mention, NtMYB12 alone can activate NtFLS and NtLAR promoter driving LUC gene expression in narcissus and tobacco cells, but repressed NtDFR promoter driving LUC gene expression only in narcissus protoplast (Fig. 4F-G; Supplementary Fig. S6), when the combined plasmids of NtMYB12-NtbHLH1, NtMYB12-NtWD40-1 or NtMYB12-NtbHLH1-NtWD40-1 with NtFLSpro-LUC, NtLARpro-LUC, or NtDFRpro-LUC were co-injected in tobacco epidermal cells without promoter with LUC as negative control and 35S promoter with LUC as positive control (Supplementary Fig. S6), in which three fluorescence signals (NtMYB12-GFP, NtbHLH1-CFP and NtWD40-mCherry) were used as controls of three plasmid transformation efficiencies (Fig. 6B), the results showed that luciferase activity did not significantly change in the tobacco leaf cells coinjected by the plasmids of NtMYB12-NtbHLH-NtWD40 with NtFLSpro-LUC or NtLARpro-LUC compared to alone NtMYB12 (Fig. 6C-H), however, surprisingly, luciferase activity increased significantly 10-100 folds in that coinjected by the combined plasmid of NtMYB12-NtbHLH-NtWD40 and NtDFRpro-LUC or by the combined plasmids of NtMYB12-NtbHLH1 or NtMYB12-NtWD40-1 or NtbHLH1-NtWD40-1 with NtDFRpro-LUC compared to NtMYB12 alone, respectively (Fig. 6I-K), MBW significantly promoted NtDFRpro-LUC signals, in MBW complex, it seems that both NtbHLH1 or NtWD40-1 promote the NtDFR promoter driving LUC expression (Fig. 6I-K), suggesting MBW specifically controlled NtDFR expression, MYB12 alone activated NtFLS and NtLAR expression and repressing NtDFR expression but both NtbHLH1 and NtWD40-1 alone activated NtDFR expression.

We then tested whether the elevated MYB12 levels could lead to metabolite flux redirection in the oeMYB12 line, enabling the Chinese narcissus to adapt to its particular color trait. The products of flavonoid and anthocyanin biosynthesis pathways were measured in the oeNtMYB12 or the Ntmyb12 and WT. Then, the contents of flavonol (Dihydroquercetin DHQ, Quercetin, Rutin), proanthocyanin (Leucocyanidin, Catechin), and anthocyanin (Cyanidin) in the oeNtMYB12 and Ntmyb12 tissue were measured and shown that the contents of flavonol were significantly accumulated in the oeNtMYB12 tissue; in contrast, they were significantly declined in the Ntmyb12 tissue (Fig. 6L). Interestingly, overexpressing oeNtMYB12 alone did not significantly increase the contents of proanthocyanin (Leucocyanidin, Catechin) and anthocyanin (Cyanidin) in narcissus callus, but knockdown Ntmyb12 significantly declined the contents of anthocyanin, suggesting that NtMYB12 was required for flavonol and (pro)anthocyanin biosynthesis. But NtMYB12 needs other regulators together to regulate anthocyanin biosynthesis.

To further test NtbHLH1 and NtWD40-1 with NtMYB12 (MBW) triplexes were involved in (pro)anthocyanin biosynthesis, we co-injected series of combinations of oeNtbHLH1-NtWD40-1-NtMYB12 and oeNtMYB12 alone to narcissus leaves (Fig. 6A), the products of flavonoid biosynthesis and anthocyanin biosynthesis pathways then were measured, the microscope checked the fluorescence of GFP, mCherry, and CFP for three plasmids successfully expressed (Fig. 6B). Apparently, the contents of flavonol and proanthocyanin were not significantly accumulated in the triplexes combinations relative to oeNtMYB12 alone (Fig. 6L), however, that of leucocyanidin was significantly declined 6 folds while anthocyanin were significantly accumulated 2.5 folds in the triplex NtbHLH-NtWD40-NtMYB12 relative to oeNtMYB12 alone (Fig. 6L) in narcissus leaves.

The stable transformation of oeNtbHLH-NtWD40-NtMYB12 in Narcissus and Arabidopsis was further performed to observe anthocyanin biosynthesis. Although oeNtMYB12-NtbHLH-NtWD40 transformed narcissus fail to grow up, the oeNtMYB12-NtbHLH-NtWD40 transformed Arabidopsis plants clearly exhibited a purple plant phenotype (Fig. 7A-D) and 15-fold high anthocyanin accumulation in leaves and twofold anthocyanin amount in seeds compared to the oeNtMYB12 alone plants (Fig. 7E-F), indicating overexpressing NtMYB12-NtbHLH-NtWD40 complex promote anthocyanin accumulation in Arabidopsis.

In addition, we further compared the coexpression network of NtbHLH1 and NtWD40-1 with NtMYB12 with metabolism profile of flavonoid biosynthesis and (pro)anthocyanin biosynthesis during tepal and corona development of Narcissus tazett and tepal and corona of two narcissus species Narcissus tazett and Narcissus pseudonarcissu species. Referring to previous data (Yang et al. 2021), interestingly, with the development of Narcissus tazett flower, the color of tepal was shifted from yellow to white, although the color of corona did not changed during development, the content of flavonol and proanthocyaynin were firstly upregulated at stage 1, then was immediately down-regulated (Yang et al. 2021), and the gene expression of NtMYB12, NtbHLH1, NtWD40-1, NtFLS and NtLAR were decline in both tepal and corona (Supplementary Figure S5), while yellow tepal and corona of Narcissus pseudonarcissu did not change during flower development (Supplementary Figure S7A), the content of flavonol and proanthocyaynin and the gene expression level of NtMYB12, NtbHLH1, NtWD40-1, NtFLS, NtLAR did not significantly change in both tepal and corona of Narcissus pseudonarcissu compared to Narcissus tazett (Supplementary Figure S7B-C). It further supported that NtMYB12-bHLH-WD40 complex requires for (pro)anthocyanin biosynthesis manipulating a colorized Narcissus tazetta “Jinzhanyintai” tepal.

Narcissus tazetta cultivated variety “Jinzhanyintai” material was grown and collected as the description in previous research (Ren et al. 2017; Yang et al. 2021). Tepals were first collected at the fifth day (sample S1) as sheathed leaves emerged from the bulbs, then collected at another three stages (samples S2-S4) as umbel in spathe developed. All samples were immediately frozen in liquid nitrogen and stored at -80⁰C.

The transgenic callus material of Narcissus tazetta was constructed on the basis of the stable genetic transformation system of narcissus callus established previously (Patent CN202210162166.8). Briefly, narcissus ovary in S0 ~ S1 period was used as explants to induce callus after disinfection, and then Agrobacterium EHA105 carrying NtMYB12 plant overexpression vector (35Spro:NtMYB12-GFP-GUS) and NtMYB12-RNAi vector were used to transform narcissus callus after 3 days of co-cultivation, it was transferred to the screening medium containing 25 μg/mL kanamycin. After 27 days of screening, it was used for material identification and experimental analysis.

Tobacco Nicotiana benthamiana was planted in a long-day climate chamber. The long-day culture conditions were a photoperiod of 16 h/8 h, a temperature of 22 °C, and a humidity of 60%) was used for transient injection experiments.

The construction of NtMYB12-RNAi vector refers to the method of Pu et al. (2012). Construction of 35Spro:NtMYB12-GFP vector was used the pK7FWG2 vector as the backbone, the 35Spro:NtMYB12-GFP vector was constructed in two steps by GATEWAY cloning. Briefly, the full-length cDNA sequence of NtMYB12 (stop codon removed), was cloned to pENTR vector (Invitrogen) by homologous recombination with the D-TOPO® gatewayvector and sequenced correctly, The LR reaction was performed with TOPO-NtMYB12 and the expression vector pK7FWG2 to obtain the final vector pK7FWG2-NtMYB12 (35Spro:NtMYB12-GFP-GUS). Other plasmids were obtained by using Novozyme's "ClonExpress® II One Step Cloning Kit" (C112) kit to obtain the linearized vector and the target gene insert by homologous recombination with the adapter sequence in infusion cloning the manner. pCBIM-GFP-Flag (35S:GFP) was used as a positive control in narcissus protoplast transformation experiments; pK7FWG2 (35S: GFP) and modified pCAMBIA3301 (ACT2pro:GFP) were used for plant overexpression vectors. pRNAi-GG (Accession No. JQ085427.1) was used for the construction of the NtMYB12-RNAi vector. pGreen (T7pro:LUC) used for the construction of the dual luciferase vector was driven by the T7 promoter firefly luciferase gene (LUC)) expression as a reporter system, and Renilla luciferase gene (REN) expression driven by NOS promoter as an internal standard. pGADT7 (AD) and pGBKT7 (BD) were used for the construction of yeast two-hybrid vectors. pRTVcVN (Accession No. MH373677) 0.1, referred to as VcVN; pRTVcVC (accession number MH373678.1) (He et al. 2018), VcVC; pRTVnVN (referred to as VnVN) and pRTVnVC (referred to as VnVC) for bimolecular fluorescence complementation experiments (BiFC), where pRTVnVN and pRTVnVC as a negative control.

According to Takara's Genome Walking Kit (Code No.6108) manual, it is designed to clone the upstream sequences of NtFLS, NtLAR, NtDFR and NtANS CDs 5' end, respectively, and the primer sequences are shown in the Supplementary Table S2.

The isolation of Narcissus tazetta leaf and tepal cell protoplasts referred to the isolation methods of Arabidopsis and rice protoplasts (Zhang et al. 2011), with slight modifications. The tepals from 4-5 flowers at development stage 4 incubated and shacked in fresh enzyme solution at 60 rpm speed under dark condition for 4 h. After the solution became feculent, the same volume of W5 buffer was added, then followed the above method (Zhang et al. 2011) to collect the protoplasts. For the transformation of Narcissus tazetta leaf and tepal cell protoplasts, refer to the method of Yoo et al. (2007) and Zhang et al. (2011).

The VcVN-NtMYB12, VcVN-NtbHLH1, VcVN-NtWD40-1, VcVC-NtMYB12, VcVC-NtbHLH1, VcVC-NtWD40-1 plasmids were extracted by the "EndoFree Plasmid Midi Kit" (CW2105), then were co-transformed to the protoplasts of Narcissus tazetta mesophyll cells and observed the fluorescence under a laser confocal microscope (Leica SP8) after 16 h. Co-transfection of VnVN and VnVC empty vector was used as the negative control.

The dual luciferase experiment refers to the method of Rodriguez (2004), with slight changes. The constructed ACT2pro:NtMYB12-GFP, ACT2pro:NtWD40-1-mCherry, ACT2pro:NtbHLH1-CFP overexpression vector (as effector), NtFLSpro:LUC, NtLARpro:LUC, NtDFRpro:LUC vector (as reporter), GFP-tagged empty vector (as negative control for effector) and pGreen empty vector (as reporter) The positive control) was transferred into Agrobacterium GV3101 (pSoup-p19) competent cells, and the positive clones were screened and identified, and the Agrobacterium resuspended solution with various combinations of effector and reporter was mixed in an equal ratio of 1:1 and used for tobacco injection. the experimental group (e.g. ACT2pro:NtMYB12-GFP + NtFLSpro:LUC), the negative control (GFP empty + NtFLSpro:LUC) and the positive control group (pGreen) should be injected into the same leaf. The injected tobacco plants were cultured in the dark for 12-24 h. After 2-3 days, the GFP fluorescence signal was observed under a fluorescence microscope or a laser confocal microscope to illustrate the expression of effector protein.

The luciferase activity of the samples was qualitatively analyzed by using the plant in vivo imaging system (the Immunization Center of Fujian Agriculture and Forestry University). Subsequently, the relative luciferase activity of the samples was quantified using a microplate reader according to Huang et al. 2020. The fluorescence signal value of firefly luciferase/renilla luciferase (LUC/REN) was used to represent the relative luciferase activity.

The combination of constructs expressing the corresponding MYB12-HA-tagged proteins and NtbHLH1-CFP or NtWD40-1-GFP-tagged proteins were co-transformed into narcissus protoplasts. After incubation 16 h, the transfected protoplasts were harvested for the Co-IP and immunoblotting assays as described previously (Zheng et al. 2018). The HA-tagged proteins were immunoprecipitated with anti-HA antibody coupled beads (anti-HA affinity matrix, Roche). The immunoprecipitated proteins were then separated via 13.5% SDS-PAGE, blotted and analyzed by immunoblotting with anti-GFP (TransGen, HT801, 1:1,000 dilution) and Anti-HA-Peroxidase (Roche, 3F10, 1:500 dilution) antibodies.

According to the description of Huang et al. (2020). A little modification was as follow: The narcissus petal protoplasts transformed with oeNtMYB12-GFP and oeGFP plasmid were collected and isolated the chromatin, and the cross-linked DNA fragments ranging from 200 to 1000 bp in length were immunoprecipitated by an antibody against the GFP (TransGen, HT801). The enrichment of the selected promoter regions of target genes was determined by comparing the amounts in the precipitated and non-precipitated (input) DNA samples, by quantitative PCR using designed region-specific primers (Supplementary Table S1). The experiments were performed in triplicate. NtACTIN were used as controls in qPCR.

Tepals, cornas from S1 to S4 stages of Narcissus tazetta and transformed callus tissue, as well transformed Arabidopsis plants were collected for analysis the contents of flavonoids and (pro)anthocyanin. HPLC were performed to analysis the extracts of flavonoids metabolism according to the description by our previous publication (Ren et al. 2017; Yang et al. 2021). The flavonoids and anthocyanin standards including naringenin, dihydroquercetin, kaempferol, quercetin, lutein, Leucocyanidin, Catechin, and Cyanidin were purchased from Sigma-Aldrich (St. Louis, MO, USA), rutin and zeaxanthin were purchased from TCI (TCI Co., Ltd, Japan) and Sangon (Shanghai, China), respectively. All flavonoids and (pro)anthocyanin standards were prepared in MeOH/DMSO (1:1, v/v) and MTBE (containing 0.01% BHT), respectively, and stored at -20 °C before use.

The total RNA of Narcissus tazetta was extracted using the protocol described by Ren et al. (2017).

The Illumina RNA-seq library can be constructed with mRNA isolated from oeNtMYB12 and Ntmyb12 tissue. Transcriptomes were sequenced using the Illumina HiSeqTM- 2500 platform for high-throughput sequencing of short reads RNA seq. Clean data were obtained by removing reads containing adapters, reads containing poly-N, and low-quality reads. These clean reads were then mapped to the reference transcriptome sequence using software (Riechmann et al. 2000). Only reads with a perfect match or one mismatch were further analyzed and annotated based on the reference transcriptome. The mapped fragments for each gene were counted by feature Counts and fragments per kilobase per million mapped reads (FPKM) were calculated. Differential expression analysis between two samples was performed using the EBSeq R package (Ogata et al. 1996). The resulting P values were adjusted using the Benja-mini and Hochberg’s approach for controlling the false discovery rate (Jia et al. 2004). Genes identified by EBSeq with FDR 0.01, FC > = 2, were defined as differentially expressed.

Illumina RNA seq needs to assemble the offline data, Sequence assembly of high-quality sequencing data using Trinity software (Trapnell et al. 2010; Sharon et al. 2013). The detail procedure was reference to manual of Biomarker technologies company.

WGCNA is a common method for constructing gene co-expression networks (Peter and Steve 2008). Through WGCNA analysis was according to the description by Yang et al. (2021).

The new Duncan's multiple range test in one-way ANOVA was used; The Kruskal-Wallis test in the unpaired test was used when they were not uniform. The different significance represented the different letters. And p < 0.05 means a significant difference.

Not applicable.

All authors approve the manuscript and consent to publication of the work.

The authors declare no conflict of interest.