Flowering time is an important ornamental trait for chrysanthemum (Chrysanthemum morifolium, Dendranthema x grandiflorum) floricultural production. In this study, CmNRRa, an orthologous gene of OsNRRa that regulates root growth in response to nutrient stress in rice, was identified from Chrysanthemum and its role in flowering time was studied. The entire CmNRRa cDNA sequence was determined using a combinatorial PCR approach along with 5′ and 3′ RACE methods. CmNRRa expression levels in various tissues were monitored by real-time RT-PCR. CmNRRa was strongly expressed in flower buds and peduncles, suggesting that CmNRRa plays a regulatory role in floral development. To investigate the biological function of CmNRRa in chrysanthemums, overexpression and knockdown of CmNRRa were carried out using transgenic Chrysanthemum plants generated through Agrobacterium-mediated transformation. CmNRRa expression levels in the transgenic plants were assayed by real-time RT-PCR and Northern blot analysis. The transgenic plants showed altered flowering times compared with nontransgenic plants. CmNRRa-RNAi transgenic plants flowered 40–64 days earlier, while CmNRRa-overexpressing plants exhibited a delayed flowering phenotype. These results revealed a negative effect of CmNRRa on flowering time modulation. Alteration of CmNRRa expression levels might be an effective means of controlling flowering time in Chrysanthemum. These results possess potential application in molecular breeding of chrysanthemums that production year-round, and may improve commercial chrysanthemum production in the flower industry.
Chrysanthemums (Chrysanthemum morifolium, Dendranthema × grandiflorum) are one of most economically important perennial flowering plants, with floricultural (cut flower), ornamental crop (pot and garden flower) and, for some cultivars, medicinal uses (Da Silva, 2003). To meet the year-round market demand, regulation of flowering initiation and development is a highly desired goal in Chrysanthemum breeding. The flowering time is affected by many variables, including environmental growth factors such as temperature (Cockshull, 1979; Damann and Lyons, 1996; Karlsson and Hanscom, 1992; Runkle et al., 1998; Yuan et al., 1995), photoperiod (Damann and Lyons, 1996; Runkle et al., 1998), light intensity (Cockshul and Hughes, 1971; Cockshull, 1979; Li et al., 2003; Rajapakse and Kelly, 1995), nutrient status (Graves et al., 1977; Graves and Sutcliff, 1974), as well as phytohormones (Graves et al., 1977; Harada and Nitsch, 1959; Jeffcoat and Cockshul, 1972; Smith and Kamp-Glass, 1990; Tompsett and Schwabe, 1974). Flowering can thus be induced artificially at specific times by altering environmental factors and externally applying phytohormones, but the process is labour and energy intensive.
With advances in genetic engineering, genetic transformation techniques now provide a promising way of introducing new traits by regulating the expression levels of exogenous and endogenous genes in plants. Several flowering-related genes have been identified in Chrysanthemum, including the MADS-box-related genes CDM111 and CDM44 involved in floral development (Shchennikova et al., 2004). As another example, DFL, a homologue of the FLORICOULA/LEAFY genes that were first characterized as floral meristem identity genes in snapdragon and Arabidopsis (Ahearn et al., 2001; Coen et al., 1990; Weigel et al., 1992), was cloned from Dendranthema lavandulifolium (Ma et al., 2008) and found to be strongly expressed in inflorescence bract, petal and stamen primordial tissues during inflorescence development. However, application of the flowering-associated genes to manipulation of chrysanthemum flowering has rarely been reported. Khodakovskaya et al. (2009) introduced the cytokinin biosynthesis isopentenyl transferase gene, driven by a tomato LEACO1 gene promoter, into the chrysanthemum cultivar ‘Iridon’. The resulting transgenic plants exhibited an increased number of flower buds, with delayed flower bud development and decreased flower bud diameters. Shulga et al. (2011) recently reported that overexpression of Chrysanthemum AP1-like genes under short-day conditions induced early flowering in transgenic chrysanthemum plants.
We previously identified an NRR gene (nutrition response and root growth) in rice involved in root growth regulation in response to nitrogen and phosphorous nutrient deficiency (Zhang et al., 2012). We also found that overexpression of NRRa in rice led to significantly delayed heading (Y. Zhang, G. Zhang, N. Xiao, L. Wang, Y. Fu, Z. Sun, R. Fang, X. Chen, unpublished data). In this study, we isolated and characterized an NRRa orthologous gene, designated as CmNRRa, from the most commonly used floricultural Chrysanthemum (Chrysanthemum morifolium cultivar ‘Xiahuang’ and ‘Hanfengche’). Among various tissues of Chrysanthemum ‘Xiahuang’, the strongest CmNRRa expression was detected in flowering buds, indicating a possible role of CmNRRa in floral development. A reverse genetics strategy was then used to overexpress or knockdown CmNRRa in transgenic Chrysanthemum ‘Xiahuang’ plants. Compared with nontransgenic plants, transgenic plants overexpressing or suppressing the expression of CmNRRa displayed significantly altered flowering times. Using this transgenic approach, we induced a long period (40–64 days) of early flowering by altering endogenous CmNRRa expression level. These results possess great application potential in the flower industry.
Characterization of CmNRRa
Combinatorial PCR was used to construct CmNRRa cDNA. Forward (JuS1, JuS2) and reverse (JuR1, JuR2) degenerate primers (Table S1) were designed based on conserved regions of rice NRRa cDNA sequences. Four degenerate primer pairs were used for PCR amplification of cDNAs from two Chrysanthemum morifolium cultivars ‘Xiahuang’ (DJ) and ‘Hanfengche’ (XJ). A 630-bp partial CmNRRa cDNA fragment from XJ was obtained with the JuS2 and JuR1 primer pair (Figure 1a), and 5′ and 3′ RACE experiments were then performed to generate the full-length CmNRRa1 cDNA sequence. From the 5′ RACE experiment, a 542-bp fragment was amplified using the Ju-5race-innp primer (Table S1) paired with the 5′ RACE inner primer provided by the RACE experiment kit (Figure 1b); sequencing of this 5′ RACE fragment revealed a 46-bp stretch upstream of the CmNRRa ATG translation start codon (Figure 2). 3′ RACE yielded 300-bp-long PCR product; sequencing of several clones indicated that the 3′ untranslated sequence of CmNRRa extended at least 231 bp after the stop codon (Figures 1c and 2).
The full-length CmNRRa cDNA sequence (accession number: HE855687) was generated from combinatorial PCR and 5′ and 3′ RACE products (Figure 2). Based on the CmNRRa1 sequence, a pair of primers were designed (Table S1) and used to amplify CmNRRa2 sequence from DJ cultivar. When the two 1074-bp CmNRRa cDNA sequences from two cultivars were compared, the nucleotide sequence identity was 98.7%, and the deduced amino acid sequence identity was 97.2%. Amino acid sequence identity between CmNRRa1 and OsNRRa was 40.2% and that between CmNRRa2 and OsNRRa was 40.8%.
CmNRRa expression pattern in Chrysanthemum
Leaf, stem, root, flower bud and flower head (i.e. disc floret, receptacle, sepal and peduncle) tissue samples were harvested from Chrysanthemum ‘Xiahuang’ plants, and the CmNRRa2 expression levels in these tissues were determined by real-time RT-PCR using primers S770 and R920 (Table S1). It was quite clear that the CmNRRa2 expression was significantly higher in flower buds and peduncles than in the other tissues (Figure 3), strongly suggesting an important role for this gene in floral development.
Establishment of the Chrysanthemum transgenic system
During the process of Chrysanthemum leaf disc regeneration, calli induction and adventitious bud regeneration were affected by the addition of the antibiotics hygromycin (Hyg), carbenicillin (Carb) and cefotaxime (Cef). Before transformation, leaf sensitivity in the culture solution was monitored for different concentrations of Hyg, Carb and Cef. When Hyg concentration was varied (0, 10, 20, 30 and 40 mg/L), callus browning was observed and few adventitious buds regenerated at 30–40 mg/L of Hyg, with a suitable concentration of Hyg determined to be 20 mg/L (Figure S1a). Similarly, a series of concentrations (0, 100, 200, 300, 400, 500, 600 and 700 mg/L) of Carb and Cef were screened (Figures S1b–c), and growth inhibition on shoot regeneration was significantly higher with Cef (300 mg/L) than with Carb (500 mg/L). The proper concentration of Carb was set accordingly at 500 mg/L.
Because the 20 mg/L of Hyg used during the Agrobacterium-mediated transformation still inhibited calli differentiation (only 2.8% differentiated), the incubated leaf disc was first cultured on 10 mg/L of Hyg until adventitious buds formed, and the amount of Hyg was then increased to the screening concentration of 20 mg/L. The elongated adventitious bud was further transferred to the 1/2MS+BA 2.0 mg/L+NAA 0.2 mg/L+Hyg 20 mg/L+Carb 500 mg/L medium for root development. With this transgenic system, 42 transgenic plants for CmNRRa-OX, 40 for CmNRRa-RNAi1 and 31 for CmNRRa-RNAi2 were obtained and used for the experiment.
Phenotypes of CmNRRa-OX and CmNRRa-RNAi transgenic plants
To identify the function of CmNRRa in Chrysanthemum, transgenic Chrysanthemum ‘Xiahuang’ plants were generated in which CmNRRa was either overexpressed (CmNRRa-OX) under the control of the cauliflower mosaic virus 35S (CaMV35S) promoter or suppressed by RNAi methods (CmNRRa-RNAi1 and CmNRRa-RNAi2) (Figure S2). CmNRRa expression levels in the resulting transgenic plants were determined by real-time PCR and Northern blot analysis (Figure 4). Compared with nontransgenic control plants that normally bloomed 216 days after transplanting (DAT), 22.50% (9/40) of CmNRRa-RNAi1 plants bloomed at 152–176 DAT, and 32.26% (10/31) of CmNRRa-RNAi2 plants bloomed at 169–176 DAT. In other words, CmNRRa-knockdown T0 transgenic plants bloomed 40–64 days earlier than their nontransgenic counterparts (Figure 5a–b). In contrast, 16.67% (7/42) of CmNRRa-OX plants showed at least a two-week delay of flower initiation (Figure 5c). To confirm the phenotypes of CmNRRa transgenic plants, the subsequent generation plants were propagated vegetatively by stem cuttings in a greenhouse. Consistent with the phenotype of the original transgenic plants, compared with the nontransgenic plants bloomed at about 234 DAT, the 43–57 day's early flowering phenotype was also displayed in the vegetatively propagated CmNRRa-RNAi plants, 25% (10/40) of CmNRRa-RNAi1 and 38.71% (12/31) of CmNRRa-RNAi2 plants bloomed at 177–191 DAT, while 59.52% (25/42) of the CmNRRa-OX plants showed delayed flowering. These results verified the function of CmNRRa in regulating flowering time in Chrysanthemum.
In this study, a full-length CmNRRa cDNA sequence was obtained using a combinatorial PCR approach together with 5′ and 3′ RACE experiments. Real-time RT-PCR analysis revealed high CmNRRa expression levels in Chrysanthemum flower tissues. The function of CmNRRa in regulating Chrysanthemum flowering time was then systemically investigated using transgenic plants. Our results demonstrated that low CmNRRa expression levels induced remarkably early flowering in Chrysanthemum; conversely, overexpression of CmNRRa resulted in delayed inflorescence initiation. This was consistent with the controlling of rice heading date by OsNRRa (Y. Zhang, G. Zhang, N. Xiao, L. Wang, Y. Fu, Z. Sun, R. Fang, X. Chen, unpublished data). Based on these data, it appears that the orthologs of NRRa may play crucial roles in regulating flowering time in both dicotyledonous and monocotyledonous plants.
There are more than 6000 cultivars with various characteristics in Chrysanthemum. ‘Xiahuang’ is a standard cut chrysanthemum with a single flower head, while ‘Hanfengche’ is a spray chrysanthemum with a branch of flowers. In this study, the floral organ-related CmNRRa genes CmNRRa1 (derived from ‘Hanfengche’) and CmNRRa2 (obtained from ‘Xiahuang’) showed a high sequence identity, indicating that CmNRRa is very conserved in Chrysanthemums. Flowering time can be changed by altering expression of only one gene, and this fact implicates great potential in breeding programmes of both ornamental plants and cereal crops.
Nutrition elements, especially the macronutrients like nitrogen, phosphate and potassium, play a crucial role in plant growth and development. Nutrient concentrations and requirements change greatly during the transition from vegetative to reproductive growth stages. Normal floral development requires proper nutrient levels, and an earlier or faster flower differentiation has in fact been correlated with earlier nutrient termination (Yen et al., 2008). Consequently, proper fertilizer application can effectively improve flower or fruit production (Lovatt, 1998), while reduction of mineral nutrient availability can accelerate flowering of Arabidopsis thaliana (Kolar and Senkova, 2008). But the effect of nutrient on flowering time in Chrysanthemum was not investigated, because it is a typical short-day plant.
Several genes have been reported to be involved in floral developmental responses to nutrient application. The flowering time of the Arabidopsis phosphate deficiency mutant pho1 is delayed by 6 days compared with that of the wild-type plant (Poirier et al., 1991), and the low-phosphorus insensitive mutant lpi flowers 3–5 days earlier than the wild type (Sanchez-Calderon et al., 2006). The Arabidopsis MYB62 gene is associated with plant flowering, root development and phosphate stress (Devaiah et al., 2009). Moreover, Seligman et al. (2008) found that nitrate reductase plays a regulatory role in the flowering process by affecting endogenous nitric oxide content. Although there is increasing evidence linking nutrient response with plant reproductive growth, the underlying molecular mechanism is still unclear. We previously reported that the OsNRRa gene is involved in control of root growth in response to nutrient starvation (Zhang et al., 2012). OsNRRa expression was significantly decreased in rice seedling shoots subjected to nitrogen nutrient starvation, suggesting that low levels of OsNRRa represented a response of the nitrogen deficient pathway for regulating aerial shoot development and reproduction. Indeed, altering NRRa levels in rice led to variations in heading date compared with that in wild-type plants (Y. Zhang, G. Zhang, N. Xiao, L. Wang, Y. Fu, Z. Sun, R. Fang, X. Chen, unpublished data). In this study, the connection of CmNRRa with flowering time was revealed by a reverse genetics study. Thus, the function of plant NRRa resides in the coordination of nutrient stress signals with reproductive development. Further dissection of the genes directly regulated by NRRa will help to clarify the molecular mechanism.
Cloning of CmNRRa
Combinatorial PCR was conducted to clone CmNRRa cDNA from Chrysanthemum ‘Xiahuang’ and ‘Hanfengche’. The amino acid sequence of rice NRRa was first aligned with its homologous sequences, and degenerate primers (forward primer JuS1, JuS2; reverse primer JuR1, JuR1; Table S1) were then designed based on two conserved regions located at residues 63–75 in the N-terminal and 282–300 in the C-terminal part. Total RNA was extracted with TRIZOL Reagent (Invitrogen, Carlsbad, CA) from chrysanthemum leaves and further treated with TURBO DNase according to TURBO DNA-free Kit (Ambion, Austin, TX) instructions. cDNAs were produced using the Superscript III First Strand Synthesis System (Invitrogen). The four primer pairs JuS1 + JuR1, JuS1 + JuR2, JuS2 + JuR2 and JuS2 + JuR1 were employed to amplify the partial CmNRRa cDNA fragment using PCR with LA Taq DNA polymerase (Takara, Dalian, China). PCR conditions were 94 °C for 4 min to denature the template, followed by 30 cycles at 94 °C for 1 min/50 °C for 1 min 10 s/72 °C for 2 min, and a final extension step of 72 °C for 10 min followed by cooling to 4 °C. The amplified fragment was cloned into a pGEM-T vector (Promega, Madison, AL) and subsequently confirmed by sequencing.
5′ RACE and 3′ RACE
Based on the CmNRRa cDNA sequences obtained in the previous step, specific primers Ju-5race-outp, Ju-5race-innp, 3race-ju2 and 3race-ju3 (Table S1) were designed for 5′ and 3′ RACE experiments. The FirstChoice RLM-RACE kit (Ambion) was used for 5′ RACE experiments. The 3′ RACE experiment was performed as described in Zhang et al. (2012). Reverse transcription was carried out using leaf total RNA and an oligo (18) dT-adapter-1 primer and was then followed by two rounds of PCR performed using the first primer pair 3race-adapter2 + 3race-ju2 and then 3race-adapter3 + 3race-ju3 (Table S1). The amplified product from the second round of PCR was cloned into a pGEM-T vector and confirmed by sequencing. Based on the obtained 5′ and 3′ ends of CmNRRa cDNA sequences, the specific primers JU-P5 and JU-P3, which spanned the entire CmNRRa cDNA sequence, were designed to amplify cDNA from ‘Xiahuang’ and ‘Hanfengche’, respectively. The resulting 1074-bp PCR product was cloned into a pGEM-T vector and sequenced.
CmNRRa overexpression construct
Forward and reverse primers of JuOX5-xba and JuOX3-xba were designed (Table S1) according to the full-length CmNRRa cDNA sequences. A 914-bp CmNRRa1 cDNA fragment, which included the ORF region, was produced by PCR and confirmed by sequencing. The amplified fragment was digested with XbaI and ligated into the pCAMBIA-1300-D4S plant expression vector digested with XbaI following removal of the 5′ phosphates with calf intestinal phosphatase; the resulting 130-CmNRRa plasmid included the ORF sequence of CmNRRa1 cDNA inserted downstream of the CaMV 35S promoter.
To knockdown CmNRRa in Chrysanthemum, we created two intron-spliced self-complementary constructs (CmNRRa-RNAi1 and CmNRRa-RNAi2) containing the sense and antisense sequences of the target fragment corresponding to the 5′ terminal (393 bp) and 3′ terminal (395 bp) of CmNRRa cDNA, respectively. The sense fragments were amplified from CmNRRa1 cDNA using primer pairs JU-1antS-EcoI + JU-1antR-Kpn and JU-2antS-EcoI + JU-2antR-Kpn (Table S1); after confirmation by sequencing, the products were digested with EcoRI/KpnI and cloned into pCAMBIA-HAN immediately downstream of the CaMV 35S promoter to generate the intermediate vector. The antisense fragments were amplified with primer pairs JU-1antS-Xba + JU-1antR-Bam and JU-2antS-Xba + JU-2antR-Bam (Table S1); the products were digested with XbaI/BamHI and cloned adjacent to the intron spacer region of the corresponding intermediate vector to create the RNAi vectors CmNRRa-RNAi1 and CmNRRa-RNAi2.
These plant expression plasmids were transferred into Agrobacterium tumefaciens strain EHA105 for transformation of Chrysanthemum.
The recombinant constructs were transformed into Chrysanthemum ‘Xiahuang’ via Agrobacterium-mediated transformation. Leaf discs from the Chrysanthemum plantlets, approximately 5 mm2 in size, were used for the regeneration system. For adventitious bud induction, leaf discs were cultured on MS+6-BA 2.0 mg/L+NAA 0.2 mg/L medium in a 26 °C growth chamber under a light intensity of 1500–2000 lx and a 16-h/8-h light/dark photoperiod. To determine the proper antibiotic for screening transformations, leaf discs were cultured for three weeks on MS+6-BA 2.0 mg/L+NAA 0.2 mg/L medium supplemented with a series of different antibiotics (concentrations of 0, 10, 20, 30 and 40 mg/L for hygromycin, and 0, 100, 200, 300, 400, 500, 600, 700 mg/L for carbenicillin or cefotaxime) and then examined for growth of adventitious buds.
For Agrobacterium-mediated transformations, leaf discs were precultured on MS+6-BA 2.0 mg/L+NAA 0.2 mg/L for 2 d. The fresh single clone harbouring the target plasmid was first cultured in LB medium rotated at 200 rpm for 12 h at 28 °C. After addition of 20 mg/L acetosyringone (AS), culturing was continued for about 3 h until an OD600 of 0.8 was obtained. Following centrifugation at 5000 rpm for 10 min and supernatant removal, the culture was re-suspended in 1/2 MS+20 mg/L AS and incubated with the precultured leaf discs for 10 min. Leaf discs incubated in 1/2MS+20 mg/L AS without Agrobacterium were used as the nontransgenic control. The remaining leaf discs were dried with autoclaved filter paper, transferred to MS+6-BA 2.0 mg/L+NAA 0.2 mg/L+AS 20 mg/L co-culture medium and cultured in the dark for 3 d. Leaf discs were then transferred to MS+6-BA 2.0 mg/L+NAA 0.2 mg/L+ Hyg 10 mg/L + Carb 500 mg/L screening medium. The screening medium was changed every two weeks until adventitious buds formed. Finally, adventitious buds were transferred to 1/2MS+BA 2.0 mg/L+NAA 0.2 mg/L+Hyg 20 mg/L+Carb 500 mg/L medium to enable root growth. The transgenic Chrysanthemum regenerants were transplanted into soil/vermiculite (1 : 1) and cultured in a greenhouse.
Real-time quantitative RT-PCR was performed with a SYBR Green Real-time PCR Master Mix (Toyobo, Osaka, Japan) and primers S770 and R920 (Table S1) to analyse the expression level of CmNRRa in transgenic plants and in leaf, stem, root, flower bud and flower head (disc floret, receptacle, sepal and peduncle) tissues. The 116-bp Chrysanthemum ACTIN gene was used as an internal control, with JUactin-F and JUactin-R primers (Table S1) designed based on a sequence of C. x morifolium (GenBank: AB205087.1).
Northern blot analysis
Total RNA (15 μg) extracted from transgenic plant leaves was used for RNA gel blotting (Sambrook and Russell, 2001). The 393-bp probe fragment was amplified from CmNRRa1 cDNA using the primer pair JU-1antS-EcoI + JU-1antR-Kpn.
Approximately 10-cm-long stem fragments from original transgenic plants were used as cuttings for subsequent generation propagation. The cuttings were first dipped into α-NAA (250 mg/L) for 3 s, transferred into soil/vermiculite (1 : 1) and cultured in a greenhouse.
This work was supported by grants from the Transgenic Program (2011ZX08010-002-003) and the State Key Laboratory of Plant Genomics (2012A0301-05).