Introduction

Anthocyanins are a kind of pigments that contribute to the red, blue, and purple colors of many plant organs and tissues in a wide range of species (Field et al. 2001; Honda and Saito 2002). Their functional roles include increasing tolerance of environmental stresses, enhancing resistance to herbivores and pathogens and attracting pollinators and seed dispersers (Simmonds 2003; Treutter 2005; Page et al. 2012). Moreover, anthocyanins have important pharmaceutical properties and their consumption is known to reduce the risk of heart disease, cancer, diabetes, and degenerative conditions such as Alzheimer’s disease (Zhang et al. 2008; Sun et al. 2013).

Anthocyanin biosynthesis, which represents an important branch of the flavonoid pathway, has been widely studied in many plants, including Arabidopsis thaliana, Petunia hybrida, as well as in species that bear fleshy fruits, such as apple (Malus domestic), strawberry (Fragaria ananassa), grape (Vitis vinifera), and pear (Pyrus communis) (Beld et al. 1989; Cavallini et al. 2014; Chagné et al. 2013; Maksym et al. 2013; Yang et al. 2013; Thilo et al. 2014). The key anthocyanin biosynthetic genes have also been cloned from the leaves of crabapple (Malus sp.), which is a useful model system for the study of anthocyanins due to the range of vivid leaf colors exhibited by different varieties, and the activities of the corresponding enzymes have been studied (Tian et al. 2011; Shen et al. 2012).

One of the most important of these enzymes, anthocyanidin synthase (ANS), is a 2-oxoglutarate (2OG) iron-dependent oxygenase that catalyzes the stepwise conversion of leucocyanidins to anthocyanins (Fig. 1), thereby completing the transformation of colorless to color compounds (Saito et al. 1999; Gong et al. 1997; Shimada et al. 2005). In the anthocyanin biosynthetic pathway, ANS is located downstream of the dihydroflavonol reductase DFR enzyme (Fig. 1) and is known to affect pathway efficiency and influence flower color depending on the species (Smith et al. 2012). The coding sequences of ANS enzymes have been identified in plant species, and several studies have shown that the expression pattern of the ANS gene corresponds with the accumulation of anthocyanins, leading to the supposition that ANS represents a key committing step in anthocyanin biosynthesis (Gong et al. 1997; Saito et al. 1999; Shimada et al. 2005). In a promoter analysis of the Forsythia intermedia ANS gene FiANS, it was shown that the lack of gene expression was the major reason for the absence of anthocyanins in the petals (Rosati et al. 1999). Moreover, in Gentiana triflora and apple reduced ANS expression resulted in fewer anthocyanin biosynthesis and a much weaker organ color (Nakatsuka et al. 2005a, b; Szankowski et al. 2009). However, the mechanism of anthocyanin accumulation and the function of ANS in Malus plant have yet to be determined.

Fig. 1
figure 1

The anthocyanin biosynthetic pathway. Enzymes involved in anthocyanin biosynthesis: PAL, phenylalanine ammonium lyase-1; C4H, cinnamate-4-hydroxylase; CHS, chalcone synthase; CHI, chalcone isomerase; F3H, flavanone 3-hydroxylase; DFR, dihydroflavonol 4-reductase; ANS, anthocyanidin synthase; GT, UDPG-flavonoid glucosyl transferase

Full size image

Malus crabapple is a valuable ornamental apple germplasm for studying mechanism of plant pigmentation and the molecular basis of the diversity of color in plant organs due to its considerable color diversity and its notably high resistance to abiotic stress. In this current study, we investigated the role of the crabapple ANS gene, McANS, in petal pigmentation using three crabapple cultivars, the petals of which are either dark red (‘Royalty’), exhibit a transition from red to pink (‘Radiant’) or vary from pink to white (‘Flame’) during petal development. Specifically, we examined the relationship between the expression levels of McANS and anthocyanin content and we also overexpressed McANS in tobacco to verify its proposed biochemical activity. The results support the hypothesis that McANS is responsible for petal coloration, and suggest that the transcription level of this gene may be a critical factor for the red coloration of crabapple flowers.

Materials and methods

Plant material

The three crabapple cultivars, ‘Royalty’ (ever-red flowers), ‘Radiant’ (buds are red and flowers are pink), and ‘Flame’ (buds are pink and flowers are white), were used in this study. These 8-year-old crabapple trees grafted onto Malus ‘Balenghaitang’ stock were planted in the Crabapple Germplasm Resources Nursery, Beijing University of Agriculture (40.l″ north latitude, 116.6″ longitude). Petals samples from the cultivars were collected at 6 days (stage I), 3 days (stage II), and 1 day (stage III) before full bloom. The leaves were collected at 3, 6, and 9 days after bud-burst, and fruits were collected at 50 and 90 days after full bloom. Tobacco (Nicotiana benthamiana) plants were maintained in a glasshouse (temperature 20–22 °C, 16 h light, and humidity 60 %). All the materials were stored at −80 °C until they were analyzed.

Isolation of the McANS full-length cDNA

Based on the MdANS sequence (Genbank Accession Number AF117269) sequence at NCBI, a pair of oligo nucleotide primers was designed to isolate the McANS cDNA sequence. Reaction conditions were 95 °C, 5 min followed by 35 cycles of 30 s at 94 °C, 30 s at 60 °C, and 90 s at 72 °C. Primer sequences used for isolation of the McANS cDNA are listed in Table S1.

Expression vector construction and tobacco transformation

The coding sequence (CDS) of the McANS gene was amplified by PCR using the ANSF1 and ANSR1 primers using cDNA prepared from ‘Royalty’ petal RNA as a template. The primers for the McANS CDS amplification were designed with SpeI and KpnI restriction enzyme sites, which were used to ligate into the pBI121 vector (Tai et al. 2014). The A. tumefaciens strain LBA4404 containing pBI121-McANS were transformed to tobacco plant via the leaf disk method. Transgenic tobacco plants were selected by kanamycin resistance (Tai et al. 2014), While only the T2 progeny from the transgenic plants were used for further analysis. qRT-PCR for McANS and other anthocyanin biosynthetic genes expression was performed as previously described (Zhang et al. 2014). All primers are listed in Table S1.

Evaluation of flower color

To evaluate flower color, color components of the CIE L*a*b* (Commission Internationale de L Eclairage) coordinate, namely lightness (L*) and chroma, were measured with a handheld spectrophotometer (Konica Minolta CR-400, Minolta, Japan, Tokyo, Japan) immediately after flowers were collected. The color analysis methods were described by Shen et al. (2012).

Measurement of anthocyanin content

Anthocyanin concentrations in the petals of all three crabapple cultivars, as well as transgenic tobacco petals, were measured by high-performance liquid chromatography (HPLC) as described by Zhang et al. (2014). Each 1 g sample was powdered in a mortar and pestle in liquid nitrogen, then 1.0 ml of extraction buffer [70:27:2:1 (v/v/v/v) mix of methanol: water: formic acid: trifluoroacetic acid] was added and the mixture was stored overnight in the dark at 4 °C. The mixture was then centrifuged at 4 °C at 12,000×g for 15 min, and each supernatant was filtered through a 0.22 μm Millipore™ filter (Millipore Corp., Billerica, MA, USA). Detection was performed at 520 nm for anthocyanins. A 250 mm × 65 mm NUCLEODURH C18 column (Pretech Instruments, Sollentuna, Sweden), operating at 25 °C, was used to separate the compounds which eluted in a mobile phase consisting of 90 % (v/v) solvent A [0.1:2:97.9 (v/v/v) mix of trifluoroacetic acid: formic acid: water] and 10 % (v/v) solvent B [0.1:2:35:62.9 (v/v/v/v) mix of trifluoroacetic acid: formic acid: acetonitrile: water] at a flow rate of 0.8 ml min−1. The elution programme followed the procedure described by Wu and Prior (2005), with some modifications. Solvent B was initially at 30 % (v/v) for 5 min and increased linearly in six steps to 35 % (v/v) at 5 min, 40 % (v/v) at 10 min, 50 % (v/v) at 30 min, 55 % (v/v) at 50 min, 60 % (v/v) at 70 min, and 80 % (v/v) at 30 min.

Quantitative real-time PCR analysis of crabapple flowers and other organs

RNA Extract Kit (Aidlab, Beijing, China) was used to extract total RNA from flowers, leaves, and pericarps according to the manufacturer’s instructions. First strand cDNA was synthesized by using oligo dT followed by the manufacturer’s instructions (TaKara, Japan). Gene expression levels were determined using quantitative real-time PCR (RT-qPCR) with SYBR Green qPCR Mix (Takara, Ohtsu, Japan) and the Bio-Rad CFX96 Real-Time PCR System (BIO-RAD, USA) as was previously described (Tai et al. 2014). The qRT-PCR primers used in this study are described in Table S1.

Phylogenetic analysis and sequence alignment

Homology analysis was determined and aligned with ANS protein sequences from other species using DNAMAN 5.2.2. A phylogenetic tree was generated using MEGA version 5.10 (Tamura et al. 2011), a minimum evolution phylogeny test, and 1,000 bootstrap replicates.

Statistical analysis

Statistical analyses were performed using Origin Pro 8 statistical software (OriginLab Corporation, One Roundhouse Plaza, Northampton, MA, 01060, USA). Microsoft Office PowerPoint 2003 was used for all graphics. Mean values ± SE are presented for each set of three biological replicate samples for each leaf stage or explants treatment. All data were analyzed using one-way ANOVA followed by Duncan’s SSR (shortest significant ranges) test to compare differences among the experimental sites with a probability (P) score >0.05 [Microsoft Excel 2003 and Data Processing System (DPS, Wiley-Blackwell) software 7.05].

Results

Cloning of the full-length McANS cDNA

We designed a pair of oligo nucleotide primers for McANS gene cloning based on the MdANS (AF117269) and related sequences in NCBI. Sequencing results revealed two alleles in crabapple, corresponding to the McANS-1 (FJ817488) and McANS-2 genes (KP742784), each of which has a full-length cDNA of 1074 bp encoding a predicted protein of 358 amino acids. The nucleotide and amino acids sequence similarity of McANS-1 and McANS-2 genes were 99 and 98 %, respectively, while the protein sequences of McANS-1 and McANS-2 shared 100 and 99 % similarity with MdANS (Malus × domestica, accession number AF117269), which has been shown to be participate in anthocyanin accumulation (Henry-Kirk et al. 2012) (Fig. 2a). The deduced amino acid sequence showed that McANS contains a 2-oxoglutarate (2OG) iron-dependent structure that is homologous to the conserved domain of ANS proteins (Xie et al. 2004).

Fig. 2
figure 2

Sequence characteristics of McANS and its relationship with ANS proteins from other species. a Protein sequence alignment of McANS with other known ANS proteins from different plant species. The 2OG-FeII-Oxy functional domain is shown in underlined, the His and Asp residues combining with Fe2+ are marked by a ‘circle’, and the conserved residues combining with 2OG are marked by a ‘square’. b Phylogenetic relationship between McANS and ANS proteins from other species. Phylogenetic and molecular evolutionary analysis was conducted using MEGA version 5.10 (using the minimum evolution phylogeny test and 1000 bootstrap replicates). The GenBank Accession numbers of the proteins are as follows: crabapple (FJ817488 and KP742784); Vitis vinifera (EU156063); Tulipa fosteriana cultivar ‘Shangnong Zaoxia’ (JX401218); Theobroma cacao (GU324350); Saussurea medusa (AY547342); Rosa hybrid cultivar (AB239791); Pyrus communis (JX403954); Prunus cerasifera (EF683132); Paeonia lactiflora (JQ070805); Medicago truncatula (EF544389); Malus x domestica (AF117269); Litchi chinensis (HQ402913); Lilium cernuum (KC013372); Gypsophila elegans (AY256380); Glycine max (EU334548); Gerbera hybrid cultivar ‘Tacora’ (AY997840); Fragaria x ananassa (JQ923457); Camellia sinensis (AY830416); Camellia japonica (AB524886); Camellia chekiangoleosa (JN944577); and Anthurium andraeanum (EF079869)

Full size image

In addition, we cloned the McANS gene sequences from the white petal cultivar ‘Flame’ and the pink petal cultivar ‘Radiant’. The McANS gene also had two alleles in these two cultivars and a sequence alignment showed that while the cDNA sequence of McANS-1 has three single-nucleotide differences between these three cultivars, the amino acid sequences were identical (Fig. S1).

To further explore the evolutionary relationships of McANS with ANS protein sequences from other species, a phylogenetic tree was constructed using MEGA 5.10. McANS-1 and McANS-2 clustered into the same subgroup with Camellia chekiangoleosa CcANS (JN944577) and Medicago truncatula MtANS (EF544389), both of which have also been shown to be involved in anthocyanin biosynthesis (Pang et al. 2007; Wang et al. 2014) (Fig. 2b). From this, we inferred that McANS may have a related function to that of other known ANS genes, and due to the high degree of similarity of McANS-1 with the functionally characterized MdANS from apple (Henry-Kirk et al. 2012), we focused on this Malus crabapple allele in this study.

Color variation and anthocyanin accumulation in three crabapple cultivars

To assess the relationship between the transcription level of McANS-1 and McANS-2 and crabapple flower color, we measured the content of anthocyanins, the level of McANS expression and petal color during different petal developmental stages in different crabapples cultivars (Fig. 3 and Fig. S2). The phenotypes of the petal in three Malus crabapple cultivars are shown in Fig. 3a. To evaluate petal color, the lightness value L*, the hue value a*, and the hue value b* were measured under natural light. This analysis showed that the ‘Royalty’ petals were red at all three developmental stages, while the petals of ‘Flame’ were pink at stage I and turned white during petal expansion. At stage I, the ‘Radiant’ petals were red and gradually turned pink with petal development (Fig. 3a). The brightness value of ‘Flame’ petals was significantly higher than that of ‘Royalty,’ while ‘Radiant’ petals had an intermediate brightness value.

Fig. 3
figure 3

Expression analysis of McANS in different petal developmental stages of the Malus crabapple cultivars ‘Royalty,’ ‘Radiant,’ and ‘Flame.’ Three stages were tested: (I) 6 days before full bloom; (II) 3 days before full bloom; and (III) 1 day before full bloom. a Flower phenotypes. b The relative expression profile of McANS. c Petal color analysis. The value ‘L*’ represents brightness, ‘a*’ represents a transition from greenness to redness as the value increases from negative to positive, ‘b*’ represents blueness/yellowness, ‘C*’ represents saturation and h* represents hue angle. Error bars indicate the mean ± SE of three replicate reactions. Different letters above the bars indicate significantly different values (P < 0.05) calculated using one-way analysis of variance (ANOVA) followed by a Duncan’s multiple range test

Full size image

HPLC analysis showed that the anthocyanin levels in the petals of ‘Royalty’ were notably higher than those in the petals of ‘Radiant’ or ‘Flame.’ During the three developmental stages, the anthocyanin content gradually decreased from 9.62 to 2.69 μg/g in the petals of ‘Royalty,’ and decreased from 4.29 to 0.19 μg/g in the petals of ‘Radiant,’ while anthocyanins were only detected at stage I of development in ‘Flame’ petals (Fig S2).

The abundance of McANS transcripts in ‘Royalty’ petals was approximately twofold and ninefold higher than in ‘Radiant’ and ‘Flame’ petals, respectively, at the first petal developmental stage, while there was a significant decrease in transcript levels during subsequent petal development in all the cultivars (Fig. 3b). Furthermore, an analysis of the correlation between anthocyanin content and the expression of McANS indicated a close relationship (Table S2). We also evaluated both the expression of McANS and anthocyanin levels in leaves and fruit peels (Fig. S3) and again saw a close correlation. Overall, we found that the anthocyanin content was consistent with McANS expression levels in the petals of three different crabapple cultivars, which suggests that the expression level of McANS may be a determining factor in the degree of petal coloration. These results also indicate that the function of McANS is spatially and temporally conserved in different tissues and among crabapple cultivars.

The transcriptional levels of anthocyanin biosynthetic genes in petals

To determine the effect of McANS expression on the regulation of other genes involved in anthocyanin biosynthesis, we analyzed the expression of a subset of genes associated with steps in the biosynthetic pathway that are upstream or downstream of McANS (Fig. 4). We observed that their expression levels generally declined with the development of petals in all three cultivars, and that the transcript levels of McCHS (chalcone synthase), McF3′H (Flavonoid 3′-hydroxylase), McDFR (dihydroflavonol 4-reductase), and McUFGT (UDPG- flavonoid-3-O-glucosyltransferase) correlated with the expression of McANS and the abundance of anthocyanins during petal development (Fig. 4). This was particularly notable for the upstream gene McCHS and the downstream gene McUFGT (Fig. 4).

Fig. 4
figure 4

Expression analysis of genes involved in the anthocyanin biosynthesis pathway. The analysis was performed using petals at three different developmentals from three different Malus crabapple cultivars (‘Royalty,’ ‘Radiant,’ and ‘Flame’). Error bars indicate the standard error of the mean ± SE of three replicate reactions. Different letters above the bars indicate significantly different values (P < 0.05) calculated using one-way analysis of variance (ANOVA) followed by a Duncan’s multiple range test

Full size image

Over-expression of McANS-1 in tobacco

To further investigate the potential role of McANS in petal coloration, we introduced the McANS-1 gene under the transcriptional control of the constitutive cauliflower mosaic virus 35S promoter into tobacco (N. benthamiana) plants, and characterized two independent T2 transgenic lines. Unlike the petals of the wild type plants, those of the transgenic tobacco plants transformed with the 35S:: McANS-1 construct presented a light red coloration, especially in OE-McANS-1-A (Fig. 5a). The ‘L*’ value was similar among the control and transgenic plants, and the ‘h*’ value was significantly lower in transgenis plants than that in control plant petals. Meanwhile, the ‘a*’ values were 2.0-fold and 3.0-fold higher in transgenic line A and B than that in control plant petals, respectively, which is representative of a light red color (Fig. 5c).

Fig. 5
figure 5

Phenotypic analysis of McANS overexpressing transgenic tobacco flowers and expression profiles of McANS. a Typical flower phenotypes of control (wild type) lines (CK) and McANS over-expressing transgenic tobacco lines (OE-McANS-1-A and OE-McANS-1-B). b Total anthocyanin content and relative expression of McANS in petals of control (CK), OE-McANS-1-A and OE-McANS-1-B lines. c Analysis of petal color control from control (CK), OE-McANS-1-A and OE-McANS-1-B lines. A spectrophotometric colorimeter was used to measure color changes in petals and HPLC was used to analyze the total anthocyanin content in the petals of the transgenic tobacco lines. The value ‘L*’ represents brightness, ‘a*’ represents a transition from greenness to redness as the value increases from negative to positive, ‘b*’ represents blueness/yellowness, ‘C*’ represents saturation and h* represents hue angle. Error bars on each symbol are mean ± SE of three replicate reactions

Full size image

To confirm the identity of the anthocyanins that were synthesized as a result of transformation of tobacco with McANS-1, petals were collected and soluble anthocyanins were analyzed by HPLC. The anthocyanin content was higher in all the transgenic lines and the levels in the 35S::McANS-1-A line were twice those of the control lines (Fig. 5b), which is consistent with the flower phenotype. qRT-PCR results confirmed a high level of transcription of McANS-1 in the transgenic lines, and the expression of McANS-1 was not detected in the control plants. The transcript expression levels of endogenous anthocyanin biosynthetic genes were similarly assessed in the McANS-1-overexpressing lines A and B, and we observed an increase in the expression of NtCHS, NtF3′H, NtDFR, and NtUFGT (Fig. S4). We also established that there was a correlation between this increase and elevated anthocyanin levels in the different transgenic lines. Taken together, these results indicate that over-expression of McANS promotes the accumulation of anthocyanins in planta, and further support the original hypothesis that McANS is key factor in determining the levels of anthocyanin accumulation.

Discussion

Anthocyanin pigments are important for plants pollination, seed dispersal and as insect and animal attractants, which were produced by a branch of the flavonoid pathway (Treutter 2005; Page et al. 2012). In addition, anthocyanin pigments are important for the commercial fruit crops as quality traits; not only for the colors that they provide but also for their potential nutritional value in the human diet (Fraser et al. 2013). In this study, we investigated the molecular mechanism of the red color formation in crabapple, and specifically the role of the anthocyanidin synthase gene McANS.

We isolated two McANS alleles (McANS-1 and McANS-2) that are almost identical in sequence to ANS genes from other apple species (e.g., Malus domestic) and the deduced corresponding proteins similarly show a high degree of sequence identity. Phylogenetic analysis placed McANS in a group of ANS proteins that have been shown to be involved in anthocyanin synthesis, with the same subgroup evolutionary relationship being to an ANS sequence (CcANS) from the flowering plant Camellia chekiangoleosa, and the greatest evolutionary distance in this study to the herbaceous plant Camellia japonica. Transcriptome analysis of C. chekiangoleosa has previously led to the suggestion that CcANS is involved in anthocyanin biosynthesis (Wang et al. 2014), from which we inferred that McANS may have a similar function.

Consistent with a role in anthocyanin synthesis, McANS transcript levels were observed to decrease during the development of crabapple petals, and this was accompanied by a reduction in the accumulation of anthocyanins. This is consistent the expression level of McANS being a determining factor in petal color formation in the three crabapple cultivars investigated here. The expression of McANS also correlated with color fading during petal expansion in the ‘Radiant’ and ‘Flame’ cultivars (Fig. 3 and Fig. S2). This is consistent with a previous study showing that the expression of ANS correlates with anthocyanin accumulation in evergreen azalea (Rhododendron × pulchrum), where anthocyanin levels increased at the initial pigmentation stage, while the ANS transcripts gradually decreased during flowering (Nakatsuka et al. 2008). Similarly, the expression of ANS was reported to increase with flower development and color formation in C. chekiangoleosa (Wang et al. 2014).

Several studies have shown that CHS (Besseau et al. 2007; Liu et al. 2012; Tian et al. 2011) and UFGT (Kobayashi et al. 2002; Nakatsuka et al. 2005a, b) are two key enzymes in the anthocyanin biosynthetic pathway. Here, a positive correlation between the expression of McANS and the upstream gene McCHS and downstream gene McUFGT was observed. Moreover, the expression levels of endogenous NtCHS and NtUFGT genes in transgenic tobacco petals were significantly promoted by the overexpression of McANS. This further suggests that the McANS is involved in anthocyanin biosynthesis and at an intermediary step between CHS and UFGT. We propose that the elevated expression of ANS may provide a feedback signal to promote the expression of CHS and UFGT to increase metabolic flux through the pathway.

Several studies have shown a clear correlation between environmental factors, anthocyanin synthesis, and ANS transcript levels. For example, light exposure can increase anthocyanin accumulation, especially in fruit skins (Azuma et al. 2012). In grapevine berries, apple, and bilberries (Vaccinium myrtillus L.), different temperature treatments can also induce quantitative and qualitative changes in the anthocyanin content, and the expression of ANS genes was reported to be increased by high light exposure (Azuma et al. 2012; Uleberg et al. 2012). The interaction of DNA-binding R2R3 MYB transcription factors, MYC-like basic helix loop helix (bHLH) proteins, and WD40 proteins can accommodate anthocyanin synthesis via coordinated transcriptional control of the anthocyanin biosynthetic genes (Allan et al. 2008; Feng et al. 2010; Zhang et al. 2011). In apple, the expression pattern of the MYB factor MdMYBA suggests that it regulates the red coloration of the fruit skin, and this transcription factor has been reported to regulate MdANS expression (Ban et al. 2007). McMYB10 also plays a key role in the regulation of anthocyanin accumulation in the petals of crabapples, where high levels of McMYB10 expression were reported to result in a strong anthocyanin accumulation phenotype (Jiang et al. 2014).

Previous studies have suggested that the expression of ANS has an important influence on the color of many plant organs. For example, There is no detectable anthocyanin accumulation in the leaves collected from ANS-RNAi suppression lines in apple (Malus domestica; Szankowski et al. 2009). Conversely, in this current study, over-expression of the McANS-1 gene in tobacco increased the content of anthocyanins and the expression of endogenous anthocyanin biosynthetic genes (Fig. 4, Fig. S4). We hypothesize that high levels of McANS expression may provide a feedback signal to promote the anthocyanin biosynthetic pathway (Figs. 3, 4, 5 and S4), resulting in enhanced anthocyanin accumulation.

Conclusion

Crabapple represents an important ornamental and economically apple germplasm resource. In this study, we isolated and characterized an endogenous crabapple gene, McANS-1, and demonstrated that it has an important function anthocyanin biosynthesis in crabapple petals. The transcript abundance of McANS correlated closely with petal pigmentation levels and we conclude that variation in the expression levels of McANS is a key determinant of differences in organ color between crabapple cultivars. We propose that the characterization of this crabapple anthocyanin biosynthetic gene may provide important information for the breeding of crabapples with vivid colors, and future work will be focused on the transcriptional regulation of McANS.

Author contribution statement

YCY designed the research. JT and JZ prepared figures and photos and wrote the paper. YZH and JZ carried out experiments and statistical analysis of the data. XZ analyzed data. JZ and TTS generated plant material and performed the tobacco transformation.