Introduction

Senescence and abscission are the last stages of plant development. Senescence/abscission can participate in plant nutrient reuse and defense against biotic/abiotic stress (Schenk et al. 2005; Avila-Ospina et al. 2014; Koyama 2014). Flower senescence and abscission are regulated by the environment, pollination, and hormones. Pollination can induce the synthesis of ethylene, which binds to the ethylene receptor and initiates ethylene signaling, causing flower senescence (Woltering et al. 1995; Sekhon et al. 2012; van Doorn and Kamdee 2014). Flower senescence was delayed in the ethylene receptor1 (etr1–1) and ethylene insensitive2 (ein2–1) ethylene signaling mutant plants (Patterson and Bleecker 2004; Arora 2005). Auxin can also regulate flower senescence since flower senescence was delayed in auxin response factor1/2 (arf1/2) double mutants (Ellis et al. 2005; Kim et al. 2011; Brumos et al. 2014; Shi et al. 2015). The balance between cytokinin and ethylene also affects the senescence/abscission of flowers (Riefler et al. 2006; Hwang et al. 2012; Rogers 2013). Floral organ abscission occurs in the basal region of the silique called the abscission zone (Lewis et al. 2006; Aalen et al. 2013). When flower organs begin to separate from the plant, the abscission zone cells will begin to divide and differentiate. This step is regulated by BLADE ON PETIOLE1/2 (BOP1/2) genes (McKim et al. 2008). Floral organ abscission is deficient in bop1/2 double mutant plants. After abscission zone cell division/differentiation, digestive enzymes break down the pectin between organs and cause the separation of the floral organs from the plants. This process is controlled by genes such as INFLORESCENCE DEFICIENT IN ABSCISSION (IDA), and floral organ abscission was significantly deficient in ida mutant plants (Jinn et al. 2000; Butenko et al. 2003; Cho et al. 2008; Aalen et al. 2013).

The MADS box genes are plant-specific transcription factors that contain MIKC conserved domains (Theissen 2001). In plants, MADS box genes regulate flower development and transition from the vegetative phase to the reproductive phase. The M domain of MADS box genes can recognize the CArG box and bind to it (Huang et al. 1993). The I and K domains are responsible for protein-protein interactions. The C domain is the functional domain and may contain repressor or activator functions (Theissen et al. 1996). MADS box genes ABCDE have largely been studied in the frame of floral organ initiation. Group A genes regulate sepal and petal development, group B genes regulate petal and stamen development, group C genes regulate stamen and carpal development, group D genes regulate ovule development, and group E proteins interact with the A-D group proteins to perform their functions (Kater et al. 2006).

In addition to regulating floral organ formation, MADS box genes are also involved in the control of various plant developmental processes. For example, the MADS box transcription factor FOREVER YOUNG FLOWER (FYF) was reported to regulate flower senescence and abscission (Chen et al. 2011a). FYF is highly expressed in sepals/petals during early flower development, and its expression significantly decreased when flowers matured. Ectopic expression of FYF upregulated FYF UPREGULATING FACTOR1 (FUF1) and repressed downstream genes in ethylene signaling (EDF1/2/3/4) and abscission-associated genes (BOP1/2 and IDA), which resulted in delayed flower senescence and abscission (Chen et al. 2011a, 2015). Ectopic expression of the Oncidium orchid FYF orthologue OnFYF could also delay flower senescence/abscission in transgenic Arabidopsis plants (Chen et al. 2011b). In addition to FYF, there are other MADS box genes that are also involved in the control of floral organ senescence and abscission, such as AGAMOUS-LIKE 15/18 (AGL15/18) and SHORT VEGETATIVE PHASE (SVP) (Alvarez-Buylla et al. 2000; Fernandez et al. 2000; Adamczyk et al. 2007; Tao et al. 2012; Fernandez et al. 2014; Lee et al. 2014). Ectopic expression of these genes also caused significant delays in flower senescence and abscission. The mechanism for these MADS box genes in regulating flower senescence/abscission is still unclear.

Cattleya orchids are a popular floral crop in the international market since many cultivars with diverse flower shapes, colors, and sizes have been identified. The relatively short shelf life of Cattleya orchid flowers has caused a significant reduction of their value in the flower market. Thus, the development of a strategy to elongate the vase life for Cattleya orchid flowers is important and should immediately increase their market value. To further explore the potential role of the FYF orthologue in the regulation of Cattleya orchid floral senescence and abscission, we identified two FYF orthologues, CaFYF1 and CaFYF2, from Cattleya intermedia and functionally analyzed these genes in this study. We demonstrated that flower senescence and abscission were delayed in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX transgenic Arabidopsis plants. Our results showed that CaFYF1/2 also likely act as repressors in controlling floral organ senescence and abscission, similar to the role of Arabidopsis FYF. These results provide an applicable future strategy for the control of shelf life of Cattleya orchids by manipulation of the CaFYF1/2 function in this orchid. Furthermore, 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX also produced short sepals and petals in transgenic Arabidopsis plants, which were correlated with the upregulation of DELLA genes. Thus, this study showed that FYF orthologues could also affect flower organ development by negatively regulating GA response through activation of DELLA expression during flower development.

Materials and Methods

Plant Materials and Growth Conditions

The seeds of Arabidopsis were germinated in murashige and skoog medium (Murashige and Skoog 1962) for 14 days under long-day conditions (16-h light/8-h dark) before being transferred to soil as described previously (Chen et al. 2015). Cattleya intermedia used in this study were maintained in the greenhouse of National Chung-Hsing University, Taichung, Taiwan.

Cloning of the cDNA for CaFYF1/2 from Cattleya intermedia

The cDNA sequences of CaFYF1/2 were identified from C. intermedia through next-generation sequencing (NGS) analysis of Cattleya flowers. For cloning cDNA of CaFYF1/2, the RNA was extracted from the flower of C. intermedia and RT-PCR amplification was performed. The full-length cDNA of CaFYF1 was amplified by PCR using 5′ primer, CaFYF1-F, and the 3′ primer, CaFYF1-R. The full-length cDNA of CaFYF2 was amplified by PCR using 5′ primer, CaFYF2-F, and the 3′ primer, CaFYF2-R. Sequences for the primers are listed in the Table S1 in the supplementary material. The amplified cDNA fragments of CaFYF1/2 were cloned into the linker region in binary vector pEpyon-22K (CHY Lab, Taichung, Taiwan) under the control of cauliflower mosaic virus (CaMV) 35S promoter and used for further plant transformation.

Construction of the CaFYF1/2+SRDX and CaFYF1/2+VP16 Constructs

For the 35S::CaFYF1/2+SRDX constructs, the cDNAs for CaFYF1/2 obtained by PCR amplification using the primers CaFYF1-F and CaFYF1-NS-R were cloned into the pEpyon-2aK plasmid upstream of the SRDX (LDLDLELRLGFA*) sequence, under the control of the CaMV 35S promoter as described previously (Chen et al. 2011a, 2015). For the 35S::CaFYF1/2+VP16 constructs, the cDNAs for CaFYF1/2 obtained by PCR amplification using the primers CaFYF2-F and CaFYF2-NS-R were cloned into the pEpyon-2bK plasmid upstream of the VP16-AD fragment, under the control of the CaMV 35S promoter as described previously (Chen et al. 2011a). Sequences for the primers are listed in the Table S1 in the supplementary material.

Plant Transformation and Transgenic Plant Analysis

Constructs described in this study were constructed and introduced into the Agrobacterium tumefaciens strain GV3101 and transformed into Arabidopsis plants using the floral dip method as described previously (Clough and Bent 1998). Transformants that survived in medium containing kanamycin (50 μg/ml) were further verified by RT-PCR analyses.

Real-time PCR Analysis

For real-time quantitative RT-PCR, the reaction was performed on a C1000 thermal cycler/Bio-Rad CFX96 touch real-time PCR detection system using SYBER Green Real-time PCR Master Mix (FastStar Universal SYBR Green Master ROX Roche). The amplification condition was 95 °C for 10 min, followed by 40 cycles of amplification (95 °C for 15 s, 55 °C for 15 s, 72 °C for 30 s, and then plate reading) and melted (65–95 °C with plate readings every 1 °C). Sequences for the primers used for real-time quantitative RT-PCR for CaFYF1, CaFYF2, IDA, BOP1, BOP2, ETR1, ETR2, ERS1, ERS2, EIN4, CTR1, EIN2, EIN3, EIL1, EDF1, EDF2, EDF3, EDF4, RGA, GAI, RGL1, and RGL2 were listed in the Table S1 in the supplementary material. The Arabidopsis housekeeping gene UBQ10 was used as a normalization control with the following primers: RT-UBQ10-F and RT-UBQ10-4-2. The transcript levels for orchid CaFYF1/2 genes were normalized using reference genes CaUBQ (primers RT-CaUBQ-1 and RT-CaUBQ-2) for Cattleya intermedia. All experiments were repeated at least twice for reproducibility. Data were analyzed using Gene Expression Macro software (version 1.1, Bio-Rad).

Ethylene Responses

For ethylene sensitivity test, mature wild-type and transgenic Arabidopsis plants were sealed in plastic chambers and gassed with air or air containing 10 ppm ethylene for 3 days under long-day conditions (16-h light/8-h dark) as described previously (Chen and Bleecker 1995; Chen et al. 2011a, 2015).

Results

Characterization of CaFYF1/2 from Cattleya Orchids

The Arabidopsis FYF gene can delay floral senescence and abscission by upregulation of FUF1 to repress the ethylene signaling downstream genes EDF1–4 (Chen et al. 2011a, 2015). Two FYF orthologues were identified from C. intermedia and named CaFYF1/2. CaFYF1 contains 660 nucleotides encoding 219 amino acids (Supplementary Fig. S1) and shows 44% and 70% identity to Arabidopsis FYF and Oncidium OnFYF, respectively, with 84% (49/58) and 97% (56/58) of the amino acids identical in the MADS box domain (Fig. 1(a)), respectively. CaFYF2 contains 675 nucleotides encoding 224 amino acids (Supplementary Fig. S2) and shows 40% and 57% identity to Arabidopsis FYF and Oncidium OnFYF, respectively, with 83% (48/58) and 93% (54/58) of the amino acids identical in the MADS box domain (Fig. 1(a)), respectively. Based on sequence comparison, CaFYF1 had a higher similarity to the FYF orthologues than CaFYF2. The amino acid sequence alignment for FYF, CaFYF1/2, and OnFYF and the sequences of several other FYF-like genes (Supplementary Fig. S3) from different plants were used to construct a phylogenetic tree (Fig. 1(b)). CaFYF1/2 were assigned to the orchid clade of FYF-like genes, and as expected, CaFYF1 is closer to the other orchid FYF orthologues than CaFYF2.

Fig. 1
figure 1

Comparison of the protein sequences and phylogenetic analysis of FYF proteins in plants. a Comparison of the protein sequences of FYF (Arabidopsis), OnFYF (Oncidium), and CaFYF1/2 (Cattleya). The amino acid residues conserved in FYF are highlighted in black; amino acid residues similar to FYF are highlighted in gray, and dashes were introduced into the sequence to improve the alignment. b Phylogenetic tree analysis of plant FYF genes. Based on the protein sequence, CaFYF1/2 (in red color) were assigned to the orchid group of FYF genes, which are separated from Arabidopsis FYF in the dicot group of FYF genes. Names of the plant species are listed behind each of the MADS box protein names. Amino acid sequences for protein used in phylogenetic tree analysis were retrieved from the NCBI server (http://www.ncbi.nlm.nih.gov/) and analyzed by DDBJ ClustalW (http://clustalw.ddbj.nig.ac.jp/). The bootstrap phylogenetic tree was shown by using TreeView. Numbers on major branches indicate bootstrap percentages for 100 replicate analyses

Full size image

The Expression of CaFYF1/2 Genes

To analyze the expression pattern of the CaFYF1/2 genes during C. intermedia flower development, mRNA isolated from the organs (dorsal sepal, lateral sepal, petal and lips) of flowers (Fig. 2(a)) at three different developmental stages, 0, 4, and 8 days after flower opening (DAO), was used for real-time quantitative RT-PCR. The results indicated that CaFYF1 was highly expressed in young (0 DAO) and mature (4 DAO) flowers, and its expression was significantly decreased in old flowers (8 DAO) during the late stages of flower development right before senescence occurs (Fig. 2(b)). CaFYF2 had a similar expression pattern to that of CaFYF1 in the flowers; it was also detected at a higher level in young and mature flowers than in old flowers (Fig. 2(c)). The expression pattern of higher expression in early than in late flower development for CaFYF1/2 was similar to that for Arabidopsis FYF, which was reported previously (Chen et al. 2011a).

Fig. 2
figure 2

Detection of CaFYF1/2 expression in flower organs during different developmental stages of Cattleya. a A Cattleya intermedia mature flowers consist of one dorsal sepal (DS), two lateral sepals (LS), two petals (P), and a lip (Lp). Bar = 20 mm. b–c Detection of CaFYF1 (b) and CaFYF2 (c) expression in the dorsal sepal (DS), lateral sepals (LS), petals (P), and lip (Lip) of mature flowers (MF) at three different developmental stages, 0, 4, and 8 days after flower opening, by using real-time quantitative RT-PCR. The results indicated that CaFYF1/2 were expressed in all four perianth organs, and the expression was higher in young mature flowers (0- and 4-day MF) and significantly decreased in old mature flowers (8-day MF). The transcript levels of CaFYF1/2 were determined using two to three replicates and were normalized using Cattleya UBIQUITIN (CaUBQ). The error bars represent the standard deviation. Each experiment was repeated twice with similar results

Full size image

Ectopic Expression of CaFYF1 and CaFYF1+SRDX Delayed Flower Senescence and Abscission in Transgenic Arabidopsis Plants

To analyze the function of CaFYF1, we constructed 35S::CaFYF1 and analyzed transgenic Arabidopsis plants. Ectopic expression of CaFYF1 in Arabidopsis caused early flowering (Fig. 3(a)), similar to that in the 35S::FYF transgenic plants (Chen et al. 2011a). In addition, the delay in flower senescence and abscission was also observed in these 35S::CaFYF1 plants. Based on the severity of phenotypic alteration, two classes of 35S::CaFYF1 plants were described. In medium-severe 35S::CaFYF1 plants, the delay in senescence and abscission of the flower organs was very similar to that observed in 35S::FYF flowers described previously (Chen et al. 2011a). The perianth organs of these 35S::CaFYF1 transgenic flowers did not senesce and abscise even after the maturation of the siliques (Fig. 3(b, c)). The siliques in these medium-severe plants elongated normally (Fig. 3(b, c)), as in wild-type and 35S::FYF plants (Chen et al. 2011a). In severe 35S::CaFYF1 plants, the senescence and abscission of the flower organs were also significantly delayed for longer than that of position 10 throughout the inflorescence (Fig. 3(d, e)). In addition, the flowers of these severe plants produced short sepals and petals (Fig. 3(e–h)), and some failed to open (Fig. 3(e, g)) during the entire flower development. Furthermore, these flowers were mostly male sterile, and the siliques failed to elongate (Fig. 3(d, e)) due to the indehiscent anthers of the stamens (Fig. 3(f, h)). The alteration of flower organ development in addition to the delay in senescence/abscission of the flower organs in these severe 35S::CaFYF1 flowers was also observed in the severe 35S::FYF+SRDX (Chen et al. 2011a) and 35S::FUF1 flowers (Chen et al. 2015). As shown in Fig. 3(i), relatively higher CaFYF1 expression was observed in the severe plants compared to the medium-severe plants.

Fig. 3
figure 3

Ectopic expression of CaFYF1 delays flower senescence and abscission in transgenic Arabidopsis plants. a The 35S::CaFYF1 transgenic Arabidopsis plants (right) flowered earlier than wild-type (WT) plants at the same stage (left). b Inflorescences of a 35S::CaFYF1 medium-severe plant (35S::CaFYF1-M). The flower organs remain attached to the base of the siliques in these 35S::CaFYF1 transgenic flowers (arrowed). c Flowers along the inflorescence of the 35S::CaFYF1 medium-severe plant (35S::CaFYF1-M) from (b). The numbers indicate the position of the flowers. d Inflorescences of a 35S::CaFYF1 severe plant (35S::CaFYF1-S) containing unopened flowers (arrowed) with short sepals and petals. The senescence and abscission of the flower organs were also delayed. e Flowers along the inflorescence of the 35S::CaFYF1 severe plant (35S::CaFYF1-S) from (d). The numbers indicate the position of the flowers. f–h Close-up of the flowers with short sepals (s) and petals (p) and indehiscent anthers (arrowed) of the stamens from (e). i Detection of CaFYF1 expression in one wild-type Columbia plant (WT) and two 35S::CaFYF1 plants with severe (35S::CaFYF1-S) and medium-severe (35S::CaFYF1-M) phenotypes. The expression of CaFYF1 was clearly higher in 35S::CaFYF1-S than in 35S::CaFYF1-M plants. CaFYF1 expression was undetectable in untransformed wild-type plants. The transcript levels of CaFYF1 were determined using two to three replicates and were normalized using UBIQUITIN10. The error bars represent the standard deviation. Each experiment was repeated twice with similar results

Full size image

To further analyze the function of CaFYF1, we constructed 35S::CaFYF1+SRDX (C-terminal fused the strong repressor motif SRDX (LDLDLELRLGFA)) and analyzed the transgenic plants. Similar to 35S::CaFYF1 plants, 35S::CaFYF1+SRDX Arabidopsis plants also showed early flowering (Fig. 4(a)) and delayed flower senescence and abscission phenotypes (Fig. 4(b, c)). The senescence and abscission of the 35S::CaFYF1+SRDX flower organs were also significantly delayed for longer than that of position 20 throughout the inflorescence (Fig. 4(b, c)). Similar to the severe 35S::CaFYF1 plants, the 35S::CaFYF1+SRDX flowers also produced short sepals and petals (Fig. 4(c, d)), and the siliques failed to elongate throughout development (Fig. 4(b, c)) since male sterility caused by the indehiscence of the anthers was observed (Fig. 4(d)). The similar phenotypes observed in 35S::CaFYF1 and 35S::CaFYF1+SRDX Arabidopsis plants indicated that CaFYF1 acts as a repressor in regulating flower senescence and abscission.

Fig. 4
figure 4

Ectopic expression of CaFYF1+SRDX delays flower senescence and abscission in transgenic Arabidopsis plants. a The 35S::CaFYF1+SRDX transgenic Arabidopsis plants (right) flowered earlier than the wild-type (WT) plants at the same stage (left). b Inflorescences of a 35S::CaFYF1+SRDX plant. The flower organs remain attached to the base of the flowers in the 35S::CaFYF1+SRDX transgenic flowers (arrowed). c Flowers along the inflorescence of the 35S::CaFYF1+SRDX from (b) contained short sepals/petals and no further silique development. The senescence and abscission of the flower organs were delayed. The numbers indicate the position of the flowers. d Close-up of the flowers with short sepals (s) and petals (p) and indehiscent anthers (arrowed) of the stamens from (c)

Full size image

Ectopic Expression of CaFYF2 and CaFYF2+SRDX Delayed Flower Senescence and Abscission in Transgenic Arabidopsis Plants

With a similar strategy to CaFYF1, 35S::CaFYF2 and 35S::CaFYF2+SRDX were also constructed, and the transgenic plants were analyzed to explore the function of CaFYF2. Similar to the 35S::CaFYF1 plants, 35S::CaFYF2 Arabidopsis plants also showed early flowering (Fig. 5(a)) and delayed flower senescence and abscission phenotypes (Fig. 5(b, c)). These 35S::CaFYF2 flowers also produced short sepals, petals with the indehiscence of the anthers (Fig. 5(c–e)), and failure of silique elongation throughout development (Fig. 5(b, c)). When the 35S::CaFYF2+SRDX plants were examined, they showed very similar phenotypes to the 35S::CaFYF2 plants, with early flowering (Fig. 6(a)) and delayed flower organ senescence and abscission (Fig. 6(b–e)). Both medium-severe 35S::CaFYF2+SRDX plants with elongated siliques (Fig. 6(b, c)) and severe 35S::CaFYF2+SRDX plants with short sepals/petals, indehiscent anthers, and undeveloped siliques (Fig. 6(d–g)) were observed. As shown in Fig. 6(h), relatively higher CaFYF2 expression was observed in the severe 35S::CaFYF2+SRDX plants than the medium-severe plants. These results indicated that similar to CaFYF1, CaFYF2 also acts as a repressor in regulating flower senescence and abscission.

Fig. 5
figure 5

The 35S::CaFYF2 transgenic Arabidopsis plants showed flower senescence and abscission. a The 35S::CaFYF2 transgenic Arabidopsis plants (right) flowered earlier than wild-type (WT) plants at the same stage (left). b Inflorescences of a 35S::CaFYF2 plants contained flowers (arrowed) that showed delayed senescence and abscission of the flower organs. c Flowers along the inflorescence of the 35S::CaFYF2 plant in (b) that contained short sepals and petals and no further silique development. The numbers indicate the position of the flowers. d–e Close-up of the flowers with short sepals (s) and petals (p) and indehiscent anthers (arrowed) of the stamens from (c)

Full size image
Fig. 6
figure 6

Ectopic expression of CaFYF2+SRDX delays flower senescence and abscission in transgenic Arabidopsis plants. a The 35S::CaFYF2+SRDX transgenic Arabidopsis plants (right) flowered earlier than the wild-type (WT) plants at the same stage (left). b Inflorescences of a 35S::CaFYF2+SRDX medium-severe plant (35S::CaFYF2+SRDX-M). The flower organs remain attached to the base of the siliques in the 35S::CaFYF2+SRDX transgenic flowers (arrowed). c Flowers along the inflorescence of the 35S::CaFYF2+SRDX medium-severe plant (35S::CaFYF2+SRDX-M) from (b). The numbers indicate the position of the flowers. d Inflorescences of a 35S::CaFYF2+SRDX severe plant (35S::CaFYF2+SRDX-S) containing flowers (arrowed) with short sepal/petal and without further silique development. The senescence and abscission of the flower organs for these flowers were delayed. e Flowers along the inflorescence of the 35S::CaFYF2+SRDX severe plant (35S::CaFYF2+SRDX-S) from (d). The numbers indicate the position of the flowers. f–g Close-up of the flowers with short sepals (s) and petals (p) and indehiscent anthers (arrowed) of the stamens from (e). h Detection of CaFYF2 expression in one wild-type Columbia plant (WT) and two 35S::CaFYF2+SRDX plants with severe (35S::CaFYF2+SRDX-S) and medium-severe (35S::CaFYF2+SRDX-M) phenotypes. The expression of CaFYF2 was clearly higher in 35S::CaFYF2+SRDX-S than in 35S::CaFYF2+SRDX-M plants. In untransformed wild-type plants, CaFYF2 expression was absent. The transcript levels of CaFYF2 were determined using two to three replicates and were normalized using UBIQUITIN10. The error bars represent the standard deviation. Each experiment was repeated twice with similar results

Full size image

Ectopic Expression of CaFYF1+VP16 and CaFYF2+VP16 Promoted Flower Senescence and Abscission in Transgenic Dominant-Negative Arabidopsis Plants

In this study, a dominant-negative form of either CaFYF1 or CaFYF2 was generated by fusing the activation domain VP16 to the C-terminal of CaFYF1 (35S::CaFYF1+VP16) or CaFYF2 (35S::CaFYF2+VP16), and the transgenic plants were analyzed. In contrast to the 35S::CaFYF1 and 35S::CaFYF1+SRDX Arabidopsis plants, the 35S::CaFYF1+VP16 plants showed early senescence and abscission of the flowers (Fig. 7(a, b)). The flower organ senescence and abscission were promoted as early as that at the position 1 flower in these 35S::CaFYF1+VP16 transgenic Arabidopsis plants (Fig. 7(b, c)). In addition, the 35S::CaFYF1+VP16 flowers failed to open, and no further silique elongation was observed during the entire flower development. Similar to 35S::CaFYF1+VP16 plants, 35S::CaFYF2+VP16 plants also showed early senescence and abscission of the flowers (Fig. 7(d, e)). Compared to wild-type flowers, the flowers of the 35S::CaFYF2+VP16 transgenic Arabidopsis plants showed earlier flower organ senescence and abscission at positions 1–2 (Fig. 7(e, f)).

Fig. 7
figure 7

Ectopic expression CaFYF1+VP16 and CaFYF2+VP16 promotes flower senescence and abscission in transgenic Arabidopsis plants. a Inflorescences of a 35S::CaFYF1+VP16 plant. The flower organs showed early senescence and abscission (arrowed). b Flowers along the inflorescence of the 35S::CaFYF1+VP16 from (a) that showed early senescence and abscission of the flower organs. The numbers indicate the position of the flowers. c Close-up of the flowers at positions 1 (left) and 2 (right) from (b) that showed early senescence and abscission of the flower organs, which were easily detached by a gentle touch. s, sepals; p, petals. d Inflorescences of a 35S::CaFYF2+VP16 plant. The flower organs showed early senescence and abscission (arrowed). The numbers indicate the position of the flowers. e Flowers along the inflorescence of the 35S::CaFYF2+VP16 from (d) that showed early senescence and abscission of the flower organs. The numbers indicate the position of the flowers. f Close-up of the flowers at position 2 from (e) that showed early senescence and abscission of the sepals (s) and petals (p), which were easily detached by a gentle touch

Full size image

The 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis Plants Are Insensitive to Ethylene Treatment

The plant hormone ethylene is the key hormone that controls plant senescence and abscission. Since 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX all caused similar delayed flower senescence and abscission in transgenic Arabidopsis, we thus further investigated whether the CaFYF1 and CaFYF2 genes are involved in regulation of the ethylene signaling pathway. When mature wild-type plants were exposed to air containing 10 ppm ethylene in sealed plastic chambers for 3 days, the perianth organs clearly senesced and abscised from the flower as early as that at position 1 (Fig. 8(a–c)). In contrast, the ethylene-treated 35S::CaFYF1 (Fig. 8(d–f)), 35S::CaFYF1+SRDX (Fig. 8(g–i)), 35S::CaFYF2 (Fig. 8(j–l)), and 35S::CaFYF2+SRDX (Fig. 8(m–o)) plants were all phenotypically similar to the air-treated control plants without any signs of promotion of flower senescence and abscission. The perianth organs clearly displayed vigor and remained on the flowers in these ethylene-treated 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants (Fig. 8(f,i,l,o)). These results revealed that the delayed senescence and abscission in the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis flowers were unaffected by the ethylene treatment.

Fig. 8
figure 8

Analysis of the effect of ethylene on the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis plants. a Inflorescences of a wild-type (WT) plant after being exposed to air containing 10 ppm ethylene (C2H4) for 3 days. The flowers showed early senescence and abscission (arrowed) without further development. b Close-up of the flowers at the top of the inflorescence in (a) that showed early senescence and abscission (arrowed). c Flowers along the inflorescence in (a) that showed early senescence and abscission of the flower organs. The numbers indicate the position of the flowers. d, g, j, m Inflorescences of the 35S::CaFYF1 (d), 35S::CaFYF1+SRDX (g), 35S::CaFYF2 (j), and 35S::CaFYF2+SRDX (m) plants after being exposed to air containing 10 ppm ethylene (C2H4) for 3 days. The flowers were not senescent and abscised after ethylene treatment. e, h, k, n Close-up of the flowers at the top of the inflorescence from (d, g, j, m), respectively. f, i, l, o Flowers along the inflorescence from (d, g, j, m), respectively. The numbers indicate the position of the flowers

Full size image

CaFYF1/2 Inhibit the Downstream Genes EDFs in the Ethylene Response

To further investigate the function of the CaFYF1/2 genes in regulating ethylene signaling, we analyzed the expression of genes in the ethylene signaling pathway in the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis plants. We found similar expression levels for genes upstream of the ethylene signaling pathway (Alonso et al. 2003; Chen et al. 2005; Stepanova and Alonso 2005), such as the ethylene receptors ETHYLENE RESPONSE 1,2 (ETR1,2), ETHYLENE RESPONSE SENSOR 1,2 (ERS1,2), and ETHYLENE INSENSITIVE 4 (EIN4); the negative regulator CONSTITUTIVE TRIPLE RESPONSE 1 (CTR1); the positive regulator EIN2; and the transcription factors EIN3 and ETHYLENE-INSENSITIVE3-LIKE 1 (EIL1), in wild-type and 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis plants (Fig. 9(a)). In contrast, the expression levels of downstream genes of the ethylene signaling pathway (Alonso et al. 2003; Stepanova and Alonso 2005; Castillejo and Pelaz 2008), such as EDF1, 2, 3, and 4, were significantly downregulated in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis plants (Fig. 9(a)). These results clearly indicated that the delayed flower senescence/abscission in the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants was likely due to suppression of the downstream genes in the ethylene response by the CaFYF1/2 genes.

Fig. 9
figure 9

The gene expression for various transgenic plants. a The detection of expression of genes upstream (ETR1, ETR2, ERS1, ERS2, EIL1, EIN4, CTR1, EIN2, EIN3) and downstream (EDF1, 2, 3, 4) of the ethylene signaling pathway in wild-type (WT), 35S::CaFYF1, 35S::CaFYF1+SRDX, 35S::CaFYF2 and 35S::CaFYF2+SRDX flowers by real-time quantitative RT-PCR. b The detection of expression of EDF1, 2, 3, and 4 in wild-type (WT), 35S::CaFYF1+VP16 and 35S::CaFYF2+VP16 flowers by real-time quantitative RT-PCR. c The expression levels of BOP1, BOP2, and IDA in wild-type (WT) 35S::CaFYF1, 35S::CaFYF1+SRDX, 35S::CaFYF2 and 35S::CaFYF2+SRDX flowers were detected by real-time quantitative RT-PCR. d The expression levels of RGA, GAI, RGL1, and RGL2 in wild-type (WT) 35S::CaFYF1-S, 35S::CaFYF1+SRDX, 35S::CaFYF2 and 35S::CaFYF2+SRDX-S flowers were detected by real-time quantitative RT-PCR. e The expression levels of RGA, GAI, RGL1, and RGL2 in wild-type (WT) 35S::FYF and 35S::FUF1 flowers were detected by real-time quantitative RT-PCR. The number of the plants used in real-time quantitative RT-PCR analysis: WT (6), 35S::CaFYF1 (7), 35S::CaFYF1-S (6), 35S::CaFYF1+SRDX (8), 35S::CaFYF1+VP16 (6), 35S::CaFYF2 (7), 35S::CaFYF2+SRDX (8), 35S::CaFYF2+SRDX-S (6), 35S::CaFYF2+VP16 (6), 35S::FYF (6) and 35S::FUF1 (6). In real-time quantitative RT-PCR, transcript levels of these genes were determined using two to three replicates and were normalized using UBIQUITIN10. The expression of each gene in various transgenic plants was relative to that in the wild-type plant, which was set at 1. The error bars represent standard deviation

Full size image

In contrast to the gene expression levels detected in the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants, the expression levels of EDF1/2/3/4 were clearly upregulated in the early senescent 35S::CaFYF1/2+VP16 dominant-negative plants (Fig. 9(b)). In addition, the expression of the senescence marker gene SAG12 (senescence-associated gene 12) (Noh and Amasino 1999) was also significantly upregulated in the 35S::CaFYF1/2+VP16 plants (Fig. 9(b)).

BOP1/2 and IDA Expression Is Downregulated in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX Arabidopsis Plants

During flower development, the BOP1/2 genes are required for the formation of the floral AZ (McKim et al. 2008), and the IDA gene functions to regulate flower organ abscission (Butenko et al. 2003; Cho et al. 2008; Stenvik et al. 2008). Both bop1/2 and ida mutants caused the defects in flower abscission but not senescence. Since the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants showed delayed flower organ abscission, we explored the relationships among BOP1/2, IDA, and CaFYF1/2. When the expression of BOP1/2 was examined, downregulation of these two genes was observed in the 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX flowers compared to the wild-type flowers (Fig. 9(c)). Similarly, a clear reduction of IDA expression was also observed in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX flowers (Fig. 9(c)). These results revealed that CaFYF1/2 could control flower organ abscission by suppressing the abscission-related genes BOP1/2 and IDA.

Expression Levels of the GA-Responsive DELLA Genes Are Upregulated in Plants Carrying 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX

In addition to the delay in flower senescence and abscission, the severe 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants showed additional morphological defects, including short sepals and petals (Figs. 3, 4, 5, and 6). GA was reported to promote the development and elongation of flower organs by suppressing the function of the DELLA proteins (Cheng et al. 2004; Tyler et al. 2004), which contain an N-terminal DELLA domain responsible for the activity of the DELLA proteins in response to GA (Peng et al. 1997; Silverstone et al. 1998; Dill and Sun 2001). In Arabidopsis, combinations of DELLA mutants, such as rga/rgl1/rgl2, have been reported to rescue the defects of floral organs to different degrees in ga1–3 mutants (Cheng et al. 2004; Yu et al. 2004; Tyler et al. 2004). By contrast, ectopic expression of a dominant mutant of the DELLA protein, rgl1△17, caused consistent repression of GA responses and underdeveloped flower organs (Wen and Chang 2002), similar to the ga1–3 mutants (Koornneef and van der Veen 1980).

To further clarify the relationship between CaFYF1/2 and the GA response during flower organ development, we analyzed the expression of DELLA genes, such as RGA, GAI, RGL1, and RGL2, in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants. The results clearly indicated that all four DELLA genes were upregulated in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants (Fig. 9(d)). These findings indicated that ectopic expression of CaFYF1/2 not only caused the delayed senescence/abscission of the flower organs but also affected the genes regulating the GA response and caused the short flower organ phenotype. Since similar alterations of the flower organ development were observed in the 35S::FYF (Chen et al. 2011a) and 35S::FUF1 flowers (Chen et al. 2015), the expression of these DELLA genes was also analyzed in 35S::FYF and 35S::FUF1 plants. Not surprisingly, similar upregulation of these four DELLA genes was observed in the 35S::FYF and 35S::FUF1 plants (Fig. 9(e)). Thus, FYF orthologues and the downstream gene FUF1 may also be involved in the regulation of the GA response during flower development.

Discussion

CaFYF1/2 of Cattleya Function in Regulating Flower Senescence and Abscission

We previously reported that FOREVER YOUNG FLOWER (FYF), a MADS box transcription factor, could regulate flower senescence and abscission in Arabidopsis (Chen et al. 2011a). We also found that ectopic expression of the Oncidium orchid FYF orthologue OnFYF could delay flower senescence/abscission in transgenic Arabidopsis (Chen et al. 2011b). To develop a strategy to elongate the vase life for Cattleya orchid flowers, we identified two putative FYF orthologues, CaFYF1 and CaFYF2, from C. intermedia and further performed functional analysis of their role in regulating floral senescence and abscission in this study.

The identity of CaFYF1 and CaFYF2 as FYF genes in C. intermedia was first supported by the sequence identity. CaFYF1 shows 44% and 70% identity to Arabidopsis FYF and Oncidium OnFYF, respectively, whereas CaFYF2 shows 40% and 57% identity to Arabidopsis FYF and Oncidium OnFYF, respectively. In the MADS box domain, CaFYF1 and CaFYF2 showed 97% (568/58) and 93% (54/58) amino acid identity to Oncidium OnFYF. The high sequence identity between CaFYF1/2 and the FYF orthologues from various plant species reveals that these two genes are putative FYF orthologues of C. intermedia.

The second line of evidence demonstrating that CaFYF1/2 are FYF genes is the expression pattern of the CaFYF1/2 genes during flower development. CaFYF1/2 showed a very similar expression pattern, with high expression in young (just open) and mature (4 days after open) flowers, which remained vigorous without signs of senescence and abscission. The expression of CaFYF1/2 was significantly decreased in old flowers (8 days after opening) during the late stages of flower development right before flower senescence occurs. The expression of CaFYF1/2 was higher in early than late flower development, indicating that their function in preventing senescence/abscission occurred during early flower development and was similar to that for Arabidopsis FYF, which was reported previously (Chen et al. 2011a). When CaFYF1/2 expression decreased during late flower development, this suppression was weakened and resulted in the occurrence of senescence/abscission of the flowers.

The functional similarity of CaFYF1/2 to FYF orthologues in regulating flower senescence/abscission was further demonstrated by transgenic analysis. Flower senescence and abscission were significantly delayed in 35S::CaFYF1/2 transgenic Arabidopsis plants. This result indicated that the function of the Cattleya CaFYF1/2 genes was related to the regulation of flower senescence and abscission, similar to that for FYF of Arabidopsis and OnFYF of Oncidium orchid. Furthermore, similarly delayed flower senescence and abscission in 35S::CaFYF1/2+SRDX transgenic Arabidopsis plants also suggested that CaFYF1/2 may play a role as a repressor, similar to Arabidopsis FYF, in controlling flower senescence and abscission. The delayed floral senescence and abscission of 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX transgenic Arabidopsis plants were strongly correlated with the downregulation of downstream genes in the ethylene signaling pathway, such as EDF1/2/3/4 (Alonso et al. 2003; Stepanova and Alonso 2005; Castillejo and Pelaz 2008). In contrast, the expression of EDF1/2/3/4 was upregulated in the 35S::CaFYF1/2+VP16 dominant-negative mutants, which showed the promotion of flower senescence and abscission. These results indicated that CaFYF1/2 is a repressor of the ethylene response in controlling floral senescence and abscission. This assumption was further supported by the insensitivity of 35S::CaFYF1/2 and 35S::CaFYF1/2-SRDX transgenic Arabidopsis plants to ethylene treatment. In addition to regulating the ethylene response, CaFYF1/2 also delayed flower abscission by repressing the abscission-associated genes BOP1/2 and IDA (Hepworth et al. 2005; McKim et al. 2008; Norberg et al. 2005; Butenko et al. 2003). The expression of BOP1/2 and IDA was clearly downregulated in 35S::CaFYF1/2 and 35S::CaFYF1/2-SRDX flowers. Thus, CaFYF1/2 from Cattleya orchids likely controlled flower senescence and abscission by negatively regulating the ethylene response and the abscission-associated genes, similar to other FYF orthologues identified from other plant species (Fig. 10).

Fig. 10
figure 10

A model for the function of the CaFYF1/2 genes in regulating flower senescence/abscission. In wild-type flowers, CaFYF1/2 controlled senescence/abscission of the flowers by negatively regulating (sideways “T”) the downstream genes EDF1/2/3/4 in the ethylene signaling pathway. The decrease (gray right-pointing triangle) of CaFYF1/2 expression during late flower development resulted in the senescence/abscission of the flowers. CaFYF1/2 also controlled the process of abscission of the flowers by negatively regulating (sideways “T”) BOP1/2 and IDA, which are involved in AZ formation and abscission initiation, respectively. In addition to regulating flower senescence/abscission, CaFYF1/2 and their orthologues also controlled flower organ elongation by negatively regulating the GA response through activation of (black arrows) the expression of the DELLA genes during flower development

Full size image

CaFYF1/2 Regulate Flower Organ Elongation Through Negative Regulation of GA Response

Notably, 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants also showed morphological defects in flowers, such as short sepals and petals. This phenotype was observed previously in Arabidopsis plants ectopically expressing Arabidopsis FYF (Chen et al. 2011a) or its downstream gene FUF1 (Chen et al. 2015). However, the mechanism causing this phenotype has not been investigated. DELLA proteins, which were suppressed by GA signaling, were reported to play a critical role in suppressing the development and elongation of flower organs (Cheng et al. 2004; Tyler et al. 2004). Interestingly, when the expression of the DELLA genes RGA, GAI, RGL1, and RGL2 was analyzed, they were clearly upregulated in 35S::CaFYF1/2 and 35S::CaFYF1/2+SRDX plants. We further confirmed that these four DELLA genes were also activated in 35S::FYF and 35S::FUF1 plants. Thus, this study provided a novel finding that FYF orthologues could also affect flower organ development by negatively regulating the GA response through activation of DELLA expression during flower development (Fig. 10). This effect should be reduced during late flower development once FYF expression decreased and caused the elongation of the flower organs. The ectopic expression of the FYF orthologues (35S::CaFYF1/2 and 35S::FYF) constitutively activated DELLA expression and suppressed the GA response, causing inhibition of flower organ elongation during all stages of flower development as shown in our results.

In summary, this research identified two FYF orthologues from C. intermedia that may have the same function as FYF as a repressor in delaying flower senescence and abscission. CaFYF1/2 and their orthologues may have additional functions in controlling flower organ elongation by regulating the GA response during flower development. The identification and functional analysis of CaFYF1/2 in this study provide a useful strategy for the control of shelf life as well as flower shape modification of Cattleya orchids in the future.