The seeds of snapdragon, Antirrhinum majus ‘Maryland True Pink’ (Pan American Seed Co., West Chicago, IL, United States), were sowed in a mixed substrate of peat and vermiculite at a ratio of 2:1 (v/v) at 22°C and a 16/8 h light/dark cycle in a greenhouse at the Beijing University of Agriculture. The experiments were conducted when the snapdragon plants reached the full-bloom stage. The plants selected for experiment were placed in an artificial climate chamber under a 16/8 h light/dark and 22/19°C cycle with 40% relative humidity and 150 μmol⋅m^–2⋅s^–1 light intensity.

To record the effects of different light quality, flowers were kept in a growth chamber using LED light source at 22°C with 150 μmol m^–2⋅ s^–1 light intensity. Blue light (470 nm), red light (630 nm), green light (530 nm) were set for different light quality, and the flowers under dark were used as control. The flowers were sampled after 4 h exposure for different light quality. For further blue light exposure, monochromatic blue light (470 nm) and polychromatic light without blue light (EB) were used, and flowers were sampled after exposed for 12 h.

Headspace solid-phase microextraction (SPME) and gas chromatography-mass spectrometry (GC-MS) technologies were combined to collect and examine the floral volatile compounds of snapdragon. Briefly, using 10 ng⋅mL^–1 3-octanol as the internal standard, 1 g of lobes infected was placed in an airtight bottle equipped with a 100 μM Carboxen/polydimethylsiloxane fiber (SigmaAldrich) at 25°C for 5 min, followed by 40°C for 40 min. The trapped floral scent compounds were subsequently desorbed and transferred to an Agilent 5975C GC-MS (Agilent Technologies) equipped with an HP⁻¹MS fused-silica capillary column (0.25 mm diameter, 30 m length, and 0.25 μm film thickness) with helium as the carrier gas. The desorption program was designed as follows:

Isothermal column temperature at 40°C for 3 min, then the temperature was increased at a rate of 5°C min⁻¹ to 120°C for 1 min, and the temperature was then further increased at a rate of 10°C min^–1 to 180°C for 3 min, and then increased to 280°C for 1 min at a rate of 10°C min^–1. The volatiles were identified by comparing the mass spectra and retention time with the NIST 2008 mass spectra library and standard samples. Total ion chromatogram (TIC) was analyzed to detect flower volatiles.

Total RNA was isolated from snapdragon flowers using the TransZol Up Plus RNA Kit (TransGen Biotech, Co., Ltd., Beijing, China) and subsequently reverse transcribed into cDNA using the Evo MMLV RT for PCR Kit (Accurate Biotechnology, Co., Ltd., Hunan, China) according to the manufacturer’s specifications. The TF genes and the transcript levels of putative genes in different light conditions were identified according to the snapdragon reference genome (Li et al., 2019). SYBR Green-based real-time quantitative PCR (qRT-PCR) assay was performed. The expression level of AmUBI (ubiquitin) gene was used as an internal control (Manchado-Rojo et al., 2012). Relative gene expression was calculated using the 2^−ΔΔCT formula (Livak and Schmittgen, 2001). Three biological repeats were measured for each sample. All the primers used in this study are listed in Supplementary Table 1. Snapdragon exposed with blue, red, green light (150 μmol m^–2⋅s^––1 PPFD) and dark for 4 h for transcriptome sequencing (Novogene, Co., Ltd., Beijing, China). Sequencing libraries were generated using NEBNext Ultra RNA Library Prep Kit for Illumina^® (NEB, United States) following manufacturer’s recommendations. The libraries were sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated. Clean reads were obtained by removing adapter sequence, reads containing ploy-N and low quality reads from raw data. Reference genome¹ and gene model annotation files were downloaded from genome website directly. Index of the reference genome was built using Hisat2 v2.0.5 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.0.5. featureCounts v1.5.0-p3 was used to count the reads numbers mapped to each gene. All the TF genes and the transcript levels of putative genes in different light conditions were identified according to the snapdragon reference genome. In addition, according to the threshold (FPKM >1), the 18 TFs were screened. The heatmap was generated by TBtools.

The full-length CDSs of AmMYB24 and AmMYB63 without the stop codon were fused with GFP or mCherry by cloning into the pNC-Green-SubN vector and p2300-35S-H2B-mCherry vector. The recombinant constructs were transformed into Agrobacterium strain GV3101 and infiltrated into tobacco leaves. After infiltration, the tobacco plants were grown under a 16-h light/8-h dark cycle for 2 days. After incubation, subcellular localization was detected using a Confocal laser microscopy (Leica TCS SP5-II).

Genomic DNA was isolated from snapdragon leaves. To amplify the promoter sequences of AmOCS, AmMYS, AmDXS, and AmDXR, specific primers were designed according to the Snapdragon Genome Database (see text footnote 1). Finally, the 2,000 bp upstream sequences of the intiation codon ATGs of AmOCS, AmMYS, AmDXS, and AmDXR were individually cloned into the pClone007 Vector Kit (TsingKe Biological Technology Beijing, China) and then the sequences were confirmed. The promoter sequences and the potential binding sites of regulators were further analyzed online by SOGO.²

AmOCS, AmMYS, AmDXS, and AmDXR promoters were cloned into the pNC-Green-LUC vector. The cDNA of AmMYB24 and AmMYB63 were cloned into the pNC-Cam1304-MCS35S vector, which carries the CaMV 35S promoter, to be used as the effector (Yan et al., 2019). The recombinant reporter vectors were transformed separately into Agrobacterium tumefaciens strain GV3101 with the plasmid pSoup19, and the recombinant effector vectors were transformed into GV3101. Then, the effector, reporter, and control constructs were co-transfected into Nicotiana benthamiana leaves. After incubation for 48 h, the leaves were exposed to a 16/8 h light/dark cycle. Relative LUC activity was calculated via normalization to LUC activity (Liu and Axtell, 2015), the results are shown as LUC/REN ratio. The data are presented as averages of three independent biological replicates.

For bimolecular fluorescence complementation (BiFC) assays, the open reading frame of AmMYB24 was cloned into pNC-BiFC-Ecn and the full-length coding region of AmCRY1 was inserted into pNC-BiFC-Enn vector. The recombinant plasmids (pNC-BiFC-Ecn-AmMYB24 and pNC-BiFCEnn-AmCRY1) and the empty vectors were transferred into GV3101, and then combinations of the plasmids were co-transformed into Nicotiana benthamiana leaves. Confocal laser microscopy (Leica TCS SP5-II) was used to observe the YFP signals.

For yeast two-hybrid (Y2H) assays, the open reading frames of AmMYB24 and AmCRY1 were, respectively, cloned into the pGADT7 and pGBKT7 vectors, and then Y2H assays were performed using the Matchmaker Gold Y2H system user manual (Takara, Japan). Combinations of recombinant plasmids were transferred into the yeast stain Y2H GOLD. The positive clones cultivated on SD/-Leu/-Trp agar medium plates were transferred to SD/-Leu/Trp/-His/-Ade medium plates, the working concentration of X-α-Gal was 40 mg/L. The results were observed after 3 days.

The open reading frames of AmMYB24 and AmMYB63 were amplified and inserted into the pGADT7 vector. Four promoters of AmOCS, AmMYS, AmDXS, and AmDXR were ligated into the pNC-AbAi vector. Three tandem copies of the MYBCORECYCB1 motifs in the promoter of AmOCS were also cloned into separate pNC-AbAi vectors. BstBI was used to linearize the recombinant pNC-AbAi vectors prior to transforming it into Y1H-Gold for integration. To confirmed that the plasmid integrated correctly, the Matchmaker Insert Check PCR Mix 1 (Takara, Japan) was used for a colony PCR analysis. Y1H assays were performed using the Matchmaker Gold Y1H system’s user manual (Takara, Japan). The empty vector pGADT7 with pNC-AbAi-AmOCS/AmMYS/AmDXS/AmDXR and pNC-AbAi-MYBCORECYCB1 were used as negative controls. The pGADT7 and p53-AbAi were used as the positive control. The clones were transferred to SD/-Leu/Ura medium containing 300 mg/mL AbA (Aureobasidin A) and cultivated at 28°C for 2 days after selection on the SD/-Ura plates.

The pNC-TRV2 and pNC-MCS35S vectors were used to generate AmMYB24 VIGS lines and overexpression lines. In VIGS assay, virus-induced gene silencing plasmid was constructed by inserting the fragment comprising the last 200 bp of coding sequence (390 –590 bp) of AmMYB24 and AmCRY1 (1,821–2,021 bp) into a pNC-TRV2 vector using 5 × In-Fusion HD Enzyme Premix (Takara Bio Inc., Japan). For overexpression, the full-length AmMYB24 was inserted into pNCMCS35S vector. The VIGS construct, pNC-TRV2- AmMYB24, pNCTRV1 and the plant over-expression construct, pNC-35S-AmMYB24, were introduced into Agrobacterium tumefaciens GV3101 (Weidi Biotechnology Co., Shanghai, China) for snapdragon infection. The Agrobacterium cells were then harvested and resuspended in inoculation buffer (10 mol⋅L⁻¹ MgCl2; 10 mol⋅L⁻¹ MES; 200 μmol⋅L⁻¹ acetosyringone) to an O.D. of 1.0, and incubated at room temperature for 3 h in the dark. To prevent the petals from mechanical or physical injury, the flowers at development S2 were infected by a disposable sterile syringe without a needle. After Agrobacterium infiltration, plants were placed in the dark for 12 h, and then cultured for 16 h under light and 8 h under dark. After 5–7 days, samples were collected for data analysis.

All experiments were conducted in triplicate. The data were processed using SPSS 16.0. The mean values were compared via one-way ANOVA. The p-values were evaluated using the Tukey’s multiple range test. All the parameters used in Turke’s test were default parameters in SPSS. Stars indicate the level of significance, *0.01 < p < 0.05, and ^**p < 0.01.

Sequence data from this article can be found in the Snapdragon databases (see text footnote 1) under the following accession numbers: AmMYB24 (Am06g35820) and AmMYB63 (Am07g00460).

In our previous study, blue light exposure increased expression levels of AmOCS and AmMYS more than red and green light exposure, concurrently with the emission of ocimene and myrcene (Yang et al., 2022). We isolated terpenoid biosynthesis-related genes: AmDXS, AmDXR, AmOCS, and AmMYS, and exposed snapdragon flowers with monochromatic blue light and polychromatic light without blue light in order to determine their response to blue light. We found that after 12 h exposure with blue light, the relative expression level of AmOCS increased by 7.8-fold, but did not change considerably under dark and EB conditions (Figure 1A). Similar to AmOCS, the expression of AmMYS significantly increased after blue light exposure compared to dark and EB (Figure 1B). The expression of AmDXS and AmDXR also increased after blue light exposure (Figures 1C,D). The enhancement of AmMYS, AmDXS, and AmDXR expression was lower than that of AmOCS. These results confirmed the effect of blue light to enhance expression of terpenoid biosynthesis-related genes.

To investigate the regulatory basis for blue light induction of the expression of terpenoid biosynthesis related genes, we used the results of transcriptome sequencing (Yang et al., 2022) to screen TFs against the reference genome (see text footnote 1) (Li et al., 2019). Two MYB TFs were found from 18 TF genes related to MYB from the RNA-seq data for different light conditions (Figure 2A). These two TF protein sequences include typical R2R3 MYB domains and exhibit high homology to AtMYB24 and SiMYB63 (Figures 2B,C), so were named AmMYB24 and AmMYB63, respectively. The expression patterns of these two TFs in snapdragon flowers were detected. Expression of AmMYB24 was detected in all different parts of snapdragon flowers, with relatively high expression levels in the lower lobe, tube, glandular hair, and upper lobe (Figure 3A). Different from AmMYB24, relatively high expression of AmMYB63 was only found in the upper lobe and lower lobe, and the level in the upper lobe almost twofold higher than that in lower lobe (Figure 3B). Next, the expression levels of these two genes were detected at floral developmental stages from S1 to S4. The transcript expression of AmMYB24 first increased and then decreased with floral development, with highest levels during S2 and S3 stages. AmMYB63 showed a gradually increasing pattern in transcript expression from S1 to S4, with the highest level occurred at S4 (Figure 3C). In addition, to determine the site of action of AmMYB24 and AmMYB63 in cells, these two TFs were fused with GFP and introduced into N. benthamiana leaves. We observed significantly visible fluorescence in the nucleus (Figure 3D), suggesting the specific localization of AmMYB24-GFP and AmMYB63-GFP fusion proteins in the nucleus, consistent with these two TFs both acting as nuclear regulators. Additionally, both AmMYB24 and AmMYB63 showed high expression levels under blue light. Further, after exposure with blue light for 12 h, there was significantly increased expression of AmMYB24, with a more than 2.3-fold increase compared with control (Figure 3E). EB condition decreased the expression of AmMYB24 significantly, to a level far below the expression level under blue light. After 12 h exposure with blue light, there was no significant increase in the relative expression level of AmMYB63 compared with dark condition (Figure 3F). However, like AmMYB24, the expression level of AmMYB63 was repressed significantly in response to EB (Figure 3F).

To demonstrate the regulatory effects of AmMYB24 and AmMYB63, and determine if these two TFs directly target the terpenoid biosynthesis-related genes, the promoters (2,000 bp sequences) of AmDXS, AmDXR, AmOCS, and AmMYS were amplified and cloned to give four reporter vectors, pAbAiAmDXS, pAbAi-AmDXR, pAbAi-AmOCS, and pAbAi-AmMYS, respectively. As shown in Figure 4A, the Y1H assay results indicated that both AmMYB24 and AmMYB63 could bind to the promoters of AmOCS, AmMYS, AmDXS, and AmDXR, suggesting that these two putative TFs interaction with the promoters of these four genes. Based on yeast growth, AmMYB24 showed better binding affinity for these promoters than that of AmMYB63. In addition, the abilities of AmMYB24 and AmMYB63 to activate the promoters of AmOCS, AmMYS, AmDXS, and AmDXR were also performed by using a dual-LUC assay in N. benthamiana leaves. The reporter vectors were generated, and AmMYB24 and AmMYB63 were cloned into pNC-MCS35S vectors (Figure 4B) and co-transfected into N. benthamiana leaves (Figure 4C). The LUC/REN ratios revealed that AmMYB24 only activated AmOCS promoter in planta (Figure 4C). Differently, AmMYB63 represses the activities of the four promoters. Therefore, AmOCS was selected for the further validation of the regulatory effect. In the AmOCS promoter sequence, nine MYB binding sites (MBS) were identified using SOGO online (see text footnote 2), including two MYB2AT motifs, three MYBCORE motifs, two MYBST1 motifs, one MYBCOREATCYCB1 motif, one MYBGAHV motif (Figure 5A). To determine the binding site of AmMYB24 in the AmOCS promoter, five reporter vectors containing the LUC reporter gene driven by differently truncated versions of the AmOCS promoter were constructed (Figure 5B). Through dual-LUC assays, we found that there was no significant difference in the LUC/REN ratio when the promoter was truncated to P1, P2, P3, P4, or P5, but significantly higher LUC activity was detected when the promoter was full-length upstream of the ATG start codon, a level that was about four times higher LUC activity than control (Figure 5B). This was tested further by Y1H assay, which indicated that the sole MYBCOREATCYCB1 motif (TAACAAA) between positions −1,446 and −2,000 bp upstream of the ATG is most likely the target of binding of AmMYB24 to the AmOCS promoter (Figure 5C). These results revealed that AmMYB24 acts as a regulator of AmOCS by directly binding to the promoters of terpenoid biosynthetic genes, particularly AmOCS.

To validate the function of AmMYB24 in ocimene biosynthesis, the expression levels of AmMYS, AmDXS, AmDXR, and AmOCS and the emission of ocimene were investigated after transiently overexpressing and silencing AmMYB24 in snapdragon flowers. In snapdragon flowers overexpressing AmMYB24, the expression of AmMYB24 increased to about fourfold the expression level of WT, and AmOCS expression was remarkably up-regulated by 15-fold compared with the WT expression level (Figure 6A). The expression levels of both AmDXS and AmDXR were also enhanced by overexpressing AmMYB24 (Figure 6A). As expected, silencing AmMYB24 resulted in a significant decrease in the expression level of AmOCS, to a level about 1/5 that of WT (Figure 6B). As per Figure 6B, reduction in the expression levels of AmOSC, AmDXS, AmDXR, and also AmMYS in AmMYB24 silenced flowers. An obvious increase and decrease in the emission of ocimene was observed by TIC in flowers with AmMYB24 overexpression and silencing, respectively (Figure 6C). AmMYB24 overexpression resulted in a twofold increase in ocimene release compared with WT, and a 70% decrease in ocimene release was found in AmMYB24 silenced flowers (Figure 6D).

To identify upstream signaling events, we next investigated the potential role of CRY1 in the blue light induction of ocimene. An interaction between AmMYB24 and AmCRY1 was detected using Y2H assays (Figure 7B). Additionally, BiFC assay was used to assay the interaction between these two proteins (Figure 7A). The yellow fluorescence signals indicated interaction between AmMYB24 and AmCRY1 in N. benthamiana leaves. To validate the function in ocimene biosynthesis, after silencing AmCRY1 by VIGS, effects on gene expression and ocimene emission were detected. We found that the expression levels of AmMYB24, AmOCS, and AmDXR were reduced significantly by 60–90% in AmCRY1-silenced flowers. (Figure 8A). The emission of ocimene was also inhibited in AmCRY1 silenced flowers, with a 60% decrease in ocimene production (Figure 8B). These findings indicated that AmCRY1 can mediate ocimene biosynthesis by interacting with AmMYB24 in snapdragon flowers.