Identification and molecular characterization of an IDA-like gene from litchi, LcIDL1, whose ectopic expression promotes floral organ abscission in Arabidopsis

Unexpected abscission of flowers or fruits is a major limiting factor for crop productivity. Key genes controlling abscission in plants, especially in popular fruit trees, are largely unknown. Here we identified a litchi (Litchi chinensis Sonn.) IDA-like (INFLORESCENCE DEFICIENT IN ABSCISSION-like) gene LcIDL1 as a potential key regulator of abscission. LcIDL1 encodes a peptide that shows the closest homology to Arabidopsis IDA, and is localized in cell membrane and cytoplasm. Real-time PCR analysis showed that the expression level of LcIDL1 accumulated gradually following flower abscission, and it was obviously induced by fruit abscission-promoting treatments. Transgenic plants expressing LcIDL1 in Arabidopsis revealed a role of LcIDL1 similar to IDA in promoting floral organ abscission. Moreover, ectopic expression of LcIDL1 in Arabidopsis activated the expression of abscission-related genes. Taken together, our findings provide evidence that LcIDL1 may act as a key regulator in control of abscission.

Abscission is characterized by a coordination of biochemical events that take place in abscission zones (AZ) to shed vegetative or reproductive organs 1 . Currently it is well known that abscission involves multiple changes in cell structure, metabolism and gene expression, and the process is divided into four major steps 1,2 : (i) the differentiation of AZ at a specific position, (ii) the acquisition by the AZ of competence, (iii) the onset of the cell separation within the AZ, (iv) the differentiation of a protective layer at the plant's side. Over the past decades, a few genes have been found to play important roles in these four steps, which are mainly reported in model plants tomato and Arabidopsis [3][4][5][6][7][8][9][10][11][12][13][14][15][16] . In tomato, the MADS box transcriptional factors JOINTLESS and MACROCALYX form a complex to regulate the formation of the pedicel AZ together with the VHIID protein LATERAL SUPPRESSOR 11,17 . In Arabidopsis, two genes, BLADEON-PETIOLE 1 and 2 (BOP1, BOP2) which encode proteins belonging to a family containing BTB/POZ domains and ankyrin repeats, are necessary for AZ development. bop1bop2 double mutants retain all floral organs indefinitely and do not form cytologically distinct and active floral AZ cells 7 . NtBOP2, a homologue of the Arabidopsis BOP2 protein, also showed a specific role in the differentiation of the corolla AZ 16 . When AZ differentiation is completed, phytohormones are thought to be important abscission signals. In general, ethylene and jasmonic acid accelerate abscission, while auxin, gibberellins, and brassinosteroids inhibit abscission 1 . In addition, many studies have shown that the rate and degree of abscission are largely dependent on the balance between the levels of auxin and ethylene in AZ, especially on changes in auxin gradients 10,[18][19][20] . Once the abscission process is activated, many components start to function in the AZ. For the third phase, an IDA-HAESA (HAE)/HAESA-LIKE2 (HSL2) signaling system has been characterized during floral organ abscission in Arabidopsis 4,8,9,15,21 . Briefly, IDA encodes a peptide ligand that forms a complex with redundant receptor-like kinases HAE and HSL2, which presumably activates a MITOGEN-ACTIVATED PROTEIN KINASE (MAPK) cascade that acts to regulate downstream KNOX-like transcription factor BP/ KNAT1. Recently, it has been proposed that the involvement of this signaling module in abscission processes is conserved in other plant species 22,23 . However, so far, the characterization of key regulators in control of abscission has been reported mainly in the model plants tomato and Arabidopsis; these factors are largely unknown in perennial woody plants, especially in important horticultural cultivated fruit crops, such as litchi.
Litchi (Litchi chinensis Sonn.), a famous fruit tree originating from South China, has been widely cultivated in Southeast Asia due to its delicious and nutritional fruits. The litchi tree is easily subject to massive fruit drop, leading to low yield and heavy economic loss 24,25 . For example, a medium-sized tree may produce close to 60,000 female flowers but less than 5% of flowers can develop into mature fruits 26,27 . There are three to four waves of physiological fruit drop throughout fruit development depending on cultivar. Wave I, wave II, and wave III of abscission occur around 1 week, 3 weeks, and 6-7 weeks after full bloom, respectively, but wave IV is cultivar-specific and occurs at 2-3 weeks before harvest 26,27 . Thus, to reduce unexpected fruit abscission in litchi, a number of studies have been conducted on endogenous hormones [27][28][29] and carbohydrates [30][31][32][33][34][35] , which are proposed to play vital roles in the regulation of fruit abscission. Recently, genome-wide digital transcript analysis further revealed that a range of fruitlet abscission-related genes are regulated by ethylene and carbohydrate stress 36,37 . In fact, massive transcriptomic and metabonomic data about abscission were also obtained in other fruit trees, such as citrus leaf and shoot tips 38,39 , apple young fruits 40,41 and olive mature fruits 42 . A deeper understanding of the mechanisms underlying abscission would be helpful not only for the regulation of fruit crop production and the improvement of current cultural and management practices, but also for the elucidation of new molecular markers to improve genetic breeding 43 .
Here we identified three IDA-like genes in litchi, of which LcIDL1 shows the closest homology to Arabidopsis IDA and is found to be involved in litchi male flower and fruitlet abscission. Ectopic expression of LcIDL1 in Arabidopsis revealed that LcIDL1 is able to function to induce floral organ abscission. Our data suggest that LcIDL1 may act as a key regulator in control of abscission.

Identification of key candidate genes regulating fruit abscission in litchi. Previously, in an effort
to identify the components of fruit abscission in litchi, we screened for genes that were differentially expressed during abscission from a RNA-seq database. Through GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis, a number of genes were identified as the candidate components involved in fruit abscission process. These genes were clustered into diverse metabolic processes and pathways, including carbohydrate metabolism, plant hormone synthesis and signaling, transcription factor activity and cell wall modification 36,37 . But which is a master one? We found Litchi_GLEAN_10054315 encoding an IDA (INFLORESCENCE DEFICIENT IN ABSCISSION)-like gene and being highly expressed before fruit abscission. Here we designated it as LcIDL1. It appears that key regulators, such as IDA/IDL, may have conserved function in control of cell separation in different plant species 4 , so LcIDL1 attracted much more attention in our lab.
LcIDL1 is an IDA-like gene and shows closest homology to Arabidopsis IDA. To acquire all the IDA-like gene members in litchi, we used the conserved domain of IDA to blast against the litchi genome database. We obtained two more IDA-like genes, Litchi_GLEAN_10027620 and Litchi_GLEAN_10009234, and we named them as LcIDL2 and LcIDL3. IDA protein alignment between litchi and other plant species was conducted based on their full-length protein sequences. As shown in Fig. 1A, all members of the IDL gene family are characterized by an N-terminal secretory signal peptide (SP), a variable region and a conserved C-terminal region called the PIP domain 4 . The N-terminal of SP contains a stretch of aliphatic residues, a typical motif of secreted proteins. A sequence consensus logo of the C-terminal conserved PIP domain showed that this domain consists of a PSGPS motif and four highly conserved residues [(R/K)(R/K)HN] (Fig. 1B).
In order to examine the relationship of litchi IDL genes with those in other species, a phylogenetic analysis using the full-length IDA/IDL protein sequences from Arabidopsis (At), tomato (Sl), lotus (Lj), soybean (Gm), maize (Zm), poplar (Pt) and wheat (Ta) was performed. As shown in Fig. 2, the IDL family is divided into two subgroups, one containing LcIDL1 and the other containing LcIDL2 and LcIDL3. In contrast to LcIDL2 and LcIDL3, LcIDL1 shows the closest relationship to Arabidopsis IDA.
To investigate the subcellular localization of the LcIDL1 protein, the full-length coding sequence of LcIDL1 fused with yellow fluorescent protein (YFP) was delivered into the Arabidopsis protoplasts. LcIDL1-YFP protein was observed to be localized to the cell membrane and cytoplasm (Fig. 3).
LcIDL1 is an abscission-associated gene during the male flower and fruitlet drop in litchi. The expression level of LcIDL1 in pedicel AZ tissues of male flowers was determined using a real-time PCR assay. Generally, the male flowers will drop at the pedicel AZ within one week after the stamens protruding out. AZ tissues of 1st day (stage I), 3rd day (stage II) and 7th day (stage III) after stamens protruding out were harvested for analysis (Fig. 4A). The transcription level of LcIDL1 at stage III was increased significantly compared to the other two stages (Fig. 4B), which was consistent with the degree of flower abscission. Next, we examined the effect of two abscission-inducing treatments, girdling plus defoliation (GPD) and ethephon (ETH) 37 on the transcription of LcIDL1 during litchi fruitlet abscission. Both GPD and ETH treatments significantly promoted the fruitlet drop starting from the third day, the cumulative fruit abscission rate was up to 100% on the fifth day (Fig. 4C). The expression level of LcIDL1 in pedicel AZ tissues was significantly induced after GPD and ETH treatments, with about 5-fold and 4-fold higher in GPD and ETH-treated AZ tissues than that in control on the fourth day (Fig. 4D). Collectively, these data suggested that LcIDL1 is an abscission-associated gene in litchi. Previous work revealed that the cytosolic pH increase in AZ cells occurs concomitantly with the execution of organ abscission 45 . We further observed the pH-value in the abscission zone of the floral organs using a pH-sensitive indicator, BCECF 45 . The BCECF-treated floral organs were imaged using a confocal microscope. As shown in Fig. 5D, obvious green fluorescence was observed in the abscission zone of the position 5 flowers in the wild-type. In contrast, the 35S:LcIDL1 and 35S:AtIDA overexpression lines not only displayed much earlier BCECF green fluorescence which can be detected at P3, but also showed much stronger BCECF green fluorescence and maintained a relatively higher signal density prolonged to position 10 (Fig. 5D). These data confirmed that LcIDL1 is able to function to promote floral organ abscission in Arabidopsis. Phylogenetic analysis revealed that LcIDL1 is a close homologue of AtIDA. To further compare the role of LcIDL1 and AtIDA in floral organ abscission, we generated transgenic plants expressing LcIDL1 driven by 35S or AtIDA promoters under an ida-2 mutant background 9 . Homozygous ida-2 35S:LcIDL1 and ida-2 pAtIDA:LcIDL1 transgenic lines were selected and validated by real-time PCR analysis (Fig. 5A,B). As described previously 9 , ida-2, the T-DNA insertion allele of ida, is deficient in floral organ abscission (Fig. 6C). We showed that the floral organs of ida-2 35S:LcIDL1-1 and ida-2 35S:LcIDL1-2 dropped at position 5, which is earlier than the wild-type Col (position 9) and ida-2 mutant (Fig. 6A). Consistently, the BCECF green fluorescence in AZ cells of ida-2 35S:LcIDL1 plants was observed earlier, stronger and longer than that in Col plants (Fig. 6B).
Furthermore, we showed that ida-2 pAtIDA:LcIDL1 transgenic lines ida-2 pAtIDA:LcIDL1-1 and ida-2 pAtIDA:LcIDL1-2 displayed the same floral organ abscission process as the wild-type, with the floral organs dropped at position 9 (Fig. 6C). In addition, the BCECF green fluorescence in AZ cells of ida-2 pAtIDA:LcIDL1-1 and ida-2 pAtIDA:LcIDL1-2 transgenic lines was also recovered (Fig. 6D), which suggested that ectopic expression of LcIDL1 driven by the Arabidopsis IDA promoter completely rescued the ida-2 floral organ abscission phenotype. Collectively, these data confirmed that LcIDL1 is capable of functioning in place of AtIDA to induce floral organ abscission in Arabidopsis.
Ectopic expression of LcIDL1 in Arabidopsis activated abscission-related gene expression. A previous study reported a positive role of IDA in regulating abscission-related gene expression 14 . In the present work, the expression levels of a subset of abscission-related genes, such as cell-wall remodelling (TCH4/XTH22) 3 , cell expansion (EXO and EXL1) 46 , Ca 2+ binding (GAD4) 47 were tested with qRT-PCR in Col, 35S:LcIDL1-1 and ida-2. AZ regions undergoing cell-wall loosening and organ separation (positions 3-8) were harvested for analysis. As shown in Fig. 7, the expression level of TCH4, EXO, EXL1 and GAD4 was decreased in ida-2 compared with wild-type. In contrast, the transcription of these genes was significantly up-regulated in 35S:LcIDL1-1  LcEF-1a was used as an internal control. The Y-axis is fold-change, the expression levels are relative to control 0 d. Data shown are means ± SD. One-way ANOVA (Tukey-Kramer test) analysis was performed, and statistically significant differences (P < 0.05) were indicated by asterisks.
compared with the wild-type. These data suggested that LcLDL1 may induce floral organ abscission by activating abscission-related gene expression.

Discussion
To date, IDA-like genes have been found in Arabidopsis 15 , soybean (Glycine max), tomato (Solanum lycopersicum) 12 and citrus (sweet orange and clementine) 22 . However, only Arabidopsis IDA has been tested for subcellular localization and found localized in the cell membrane, supporting that IDA, as a signal peptide, can be secreted 4 . In our report, LcIDL1 can be detected in the cytoplasm (Fig. 3). AtIDA is, in contrast to LcIDL1, not found in the cytoplasm. One possible reason is that AtIDA and LcIDL1 do not function completely similarly, the other one is that different systems were used for signal detection, since AtIDA was bombarded into the onion cells and Arabidopsis protoplasts were transfected in the present assay. Thus further study on IDA delivered to Arabidopsis protoplasts can tell more.
That ethylene acts an important factor inducing organ shedding has been established for many years 48 . Exposure to exogenous ethylene has no impact on the inhibition of floral organ abscission and ethylene sensitivity in ida mature plants, suggesting that IDA acts downstream of ethylene in the pathway by which ethylene controls abscission 4 . Consistent with this, the promoter fusion construct IDA:GUS, which is primarily expressed in the floral AZ and floral organs, is restricted only to the nectaries in ethylene insensitive mutant etr1-1, indicating that ethylene acts at least partly through ETR1 to transcriptionally induce IDA expression which then instigates abscission. Here our report further supports that IDA/IDL genes seem to be influenced by an ethylene-response pathway. When treated with ethylene, the expression level of LcIDL1 significantly increased in AZ cells. Additionally, GPD treatment induced much higher expression of LcIDL1 (Fig. 4), we inferred that this is likely also associated with ethylene emission induced by GPD treatment 36 . Similar results were also obtained in Populus and oil palm, which showed that PtIDA and PtIDL1 were expressed more highly in leaves after shade treatment (a similar treatment to GPD resulting in carbohydrate stress), and EgIDA2 and EgIDA5 could be induced by ethylene 23 , further supporting that IDA functions in the ethylene-response pathway.
Genetic studies have already revealed a framework of cell signalling, membrane trafficking, and transcriptional networks in the later stages of floral organ abscission in Arabidopsis, of which IDA has been found as a critical regulator as ida mutant plants are deficient in floral organ abscission in Arabidopsis 4 , while opposite to what is observed in the ida mutant, Arabidopsis plants overexpressing IDA cause precocious floral organ abscission 6 . Interestingly, transgenic plants expressing LcIDL1 in Arabidopsis also caused earlier floral organ abscission, suggesting that LcIDL1 is able to function to induce floral abscission in Arabidopsis (Fig. 5). In addition, beyond our expectation, transgenic plants expressing LcIDL1 under the control of the 35S promoter in ida-2 also displayed earlier floral organ abscission (Fig. 6), suggesting a similar role of LcIDL1 to AtIDA. More importantly, transgenic plants expressing LcIDL1 under the control of the Arabidopsis IDA native promoter in ida-2 can completely restore the deficiency in floral organ abscission, further supporting the notion that LcIDL1 shares the same function with AtIDA of promoting floral organ abscission in Arabidospsis (Fig. 6). Similarly, that citrus CitIDA3 is able to function to promote floral organ abscission in transgenic Arabidopsis has also been reported, and ectopic expression of CitIDA3 could also complement the abscission deficiency of the ida mutant 22 . Taken together, these results indicate that IDA, as a regulator of abscission, is functionally conserved in different species.
It has been suggested that pH changes in AZ cells are associated with abscission progress. Before onset of organ abscission, the alkalization of the cytosol in AZ cells will occur and stronger BCECF fluorescence will be detected in AZ cells of mutants that promote abscission. Mutants that inhibit abscission will maintain a low pH value and weak BCECF fluorescence in AZ cells 45 . In the present study, we showed that an earlier, stronger and longer BCECF fluorescence can be detected in transgenic plants overexpressing LcIDL1 which displayed significantly earlier floral organ abscission (Figs 5 and 6), indicating that there is a strong correlation between pH changes in AZ cells and the abscission process. Further, genes involved in pH changes in AZ cells need to be identified.
Once the abscission process is activated, the pectin in the cell walls between the AZ cell layers starts to degrade followed by an expansion in the size of the AZ cells. A concerted effort of research on many plant species has identified cell wall remodeling enzymes acting on structural polysaccharides leading to the hydrolysis of the middle lamella and cell walls of the AZ cells 1 . It has also revealed that IDA not only regulates floral abscission but also regulates lateral root emergence through changing the CWR enzyme activity 13 . In this study, higher expression levels of four abscission-related cell wall remodeling genes, GAD4, TCH4, EXO and EXL1, were detected in LcIDL1 overexpression lines (Fig. 7). In addition, our previous report showed that a cluster of cell wall degradation and loosening genes including β-1,3-glucanase, β-D-xylosidase, endoglucanase, xyloglucan endotransglucosylase/ hydrolase, pectinesterase, polygalacturonase, and β-D-xylosidase were induced in AZ cells during litchi fruit abscission. Collectively, it can be proposed that IDA/IDL regulates initial cell wall loosening and separation of AZ cells to induce abscission, probably by controlling the expression of cell wall remodelling genes.
In conclusion, our findings revealed that LcIDL1 is likely involved in the regulation of male flower and fruitlet abscission in litchi. Furthermore, LcIDL1 functions similarly as AtIDA to promote floral organ abscission in Arabidopsis, which may operate through regulating pH changes and cell wall remodelling genes in AZ cells.

Materials and Methods
Plant materials and treatments. For litchi, three 9-year-old litchi trees (Litchi chinensis Sonn. cv. Feizixiao) grown in an orchard in South China Agricultural University (Guangzhou, China) were randomly selected. Thirty fruit bearing shoots with similar diameter (about 5-8 mm) growing in different directions from each tree were tagged. Ten of them were treated with girdling (a ring of bark about 0.5 cm in width and cambium was removed from the branch base) followed by defoliation (removing all leaves above the girdle) at 35 days after anthesis (GPD treatment); Ten of them were dipped in 250 mg L −1 ethephon solution (containing 0.05% Tween-80 surfactant) for 1 min (ETH treatment), while the remaining untreated shoots were used as controls. Three out of each set of ten treated shoots were used to monitor fruit abscission dynamics for 5 days and the others were used for sampling. Cumulative fruit abscission rate (CFAR) was calculated according to our previous method 49 . Samples were collected at 0, 1, 2, 3 and 4 days after treatment. AZ was excised by cutting around 2 mm at each side of the abscission fracture plane. After separation, all tissues were quickly frozen in liquid nitrogen and stored AZ tissues of position 3-8 were collected for testing. AtUBQ was used as an internal control for qRT-PCR analysis. The Y-axis is fold-change, the expression levels are relative to wild type Col. Data shown are means ± SD. One-way ANOVA (Tukey-Kramer test) analysis was performed, and statistically significant differences (P < 0.05) were indicated by asterisks.
Scientific RepoRts | 6:37135 | DOI: 10.1038/srep37135 at − 80 °C for future analysis. Each tree was treated as a biological replicate. For the flower assay, male flowers were selected and flowers with stamens just spreading were designed as stage I, the 3rd day after stamens spreading was designed as stage II, and the 7th day after stamens spreading was designed as stage III. Male flower AZ tissues were sampled as mentioned above.
For Arabidopsis materials, Arabidopsis ecotype Col-0 was used in all experiments and the ida-2 mutant (SALK_133209) 9 , which is in the Col background, was obtained from the Nottingham Arabidopsis Stock Centre (NASC). For generation of LcIDL1 and AtIDA overexpression lines in Col or in ida-2, the full-length open reading frame (ORF) of LcIDL1 and AtIDA were subcloned into vector pCAMBIA1302 under the control of the 35S promoter with Clontech's In-Fusion system primers (Supplemental Table), then these constructs were transformed into Col or ida-2 plants following the floral dip method 44 . For generation of ida-2 pAtIDA:LcIDL1 transgenic plants, the full-length ORF of LcIDL1 was subcloned into vector pCAMBIA1302 with the 35S promoter replaced with the Arabidopsis IDA promoter (− 1 to − 1490 bp), then this vector was transformed into ida-2 plants following the floral dip method. The T3 homozygous transgenic plants were used for phenotypic analysis. All the Arabidopsis plants were grown at 22 °C under long-day (16 h light/8 h dark) conditions. To reduce variation, all genotypes tested in each experiment were grown together. Subcellular Localization Analysis. The coding sequence of LcIDL1 without the stop codon was amplified by PCR primers (listed in Supplemental Table) and then subcloned into the pSAT6-EYFP-N1 vector and fused in-frame with the Yellow Fluorescent Protein (YFP) sequence under the control of the 35S promoter. The fusion constructs were introduced into Arabidopsis protoplasts by using 40% polyethylene glycol (PEG) as described previously 50 . YFP fluorescence was observed with a confocal laser scanning microscope (LSM 7 DUO, ZEISS, Germany). The transient expression assay was repeated three times.
Quantitative RT-PCR Analysis. Total RNA was isolated from litchi AZ tissues or Arabidopsis leaves (20-day-old) using 1 mL Trizol reagent (Invitrogen). The first strand cDNA synthesis was generated using 2 μ g total RNA according to the manufacturer's instructions of the TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix Kit (TransGen, Beijing). 100 ng synthesized cDNA was used as a template to perform quantitative RT-PCR analysis. PCR reactions were performed in a total volume of 20 μ L, with 0.5 μ L for each primer (10 mM, final concentration 100 nM) and 10 μ L for SYBR Green PCR Supermix (Bio-Rad) on an ABI7500 Real-Time PCR System (Applied Biosystems). The PCR program included an initial denaturation step at 94 °C for 3 min, followed by 40 cycles of 5 s at 94 °C and 1 min at 60 °C. Each sample was quantified at least in triplicate and normalized using Ubiquitin 10 (UBQ) or EF-1a as an internal control for Arabidopsis and litchi 51 , respectively. The gene-specific primer pairs for quantitative Real-Time PCR are listed in Supplemental Table. All PCR reactions were normalized using a C t value corresponding to the reference gene. The relative expression levels of target gene were calculated with the formula 2 −ddCt 52 . Values represented the average of three biological replicates.
Phenotypic Analysis. Litchi and Arabidopsis flowers were carefully removed from the plant body and then were imaged using a stereoscope (ZEISS, SV11). The whole Arabidopsis plant and a single leaf placed on a black cloth were photographed by digital camera (D3200, Nikon). BCECF fluorescence analyses. BCECF fluorescence analysis was conducted according to a previous detailed method with some modification 45 . Inflorescences with flowers located at various positions along the inflorescence were carefully removed from the plant body and immersed in 10 μ M BCECF-AM (B1150, Thermo Scientific ™ ) solution under darkness for 20 min, then the inflorescences were rinsed four times with phosphate-buffered saline (PBS, pH 7.4) to remove excess BCECE-AM. Before imaging, flowers at different positions were excised separately from the inflorescences, and each flower's sepals, petals, and stamens were removed using forceps without damaging the carpels, receptacle, and peduncle. The images were snapped with a confocal laser scanning microscope (LSM 7 DUO, ZEISS, Germany). Samples were excited by both 488 nm and 633 nm light, then BCECF fluorescence and chlorophyll auto-fluorescence were detected through 494-598 and 647-721 filters, respectively. All BCECF experiments were repeated three times with different biological samples from different inflorescences, and representative images are presented.