Effect of exogenous spermidine on floral induction, endogenous polyamine and hormone production, and expression of related genes in ‘Fuji’ apple (Malus domestica Borkh.)

Flower bud formation in ‘Fuji’ apple (Malus domestica Borkh.) is difficult, which severely constrains commercial production. Spermidine (Spd) plays an important role in floral induction, but the mechanism of its action is incompletely understood. To investigate the effect of Spd on flowering, 6-year-old ‘Fuji’ apple trees were treated with 1 × 10−5 mol L−1 Spd to study the responses of polyamines [putrescine (Put), Spd and spermine (Spm)], hormones [gibberellins (GA3) and abscisic acid (ABA)], and polyamine-, hormone- and flowering-related genes. Spd application promoted flowering during floral induction by increasing MdGA2ox2 (gibberellin 2-oxidase) through GA3 reduction and increasing MdNCED1 and MdNCED3 (9-cis-epoxycarotenoid dioxygenase) through ABA enrichment during 60 to 80 days after full bloom. The flowering rate as well as the expressions of flower-related genes, except for MdLEY (LEAFY), also increased, thereby promoting flowering. In addition, spraying with Spd significantly increased the contents of endogenous polyamines except for Spm in terminal buds by increasing the expressions of polyamine-associated genes. We hypothesize that the contribution of Spd to flowering is related to crosstalk among polyamines, hormone signals, and related gene expressions, which suggests that Spd participates in the apple floral induction process.

remodeling, RNA processing, protein activation, cell proliferation, cell signaling regulation, membrane stabilization, and modulation of cell death 12,14,15 . Furthermore, numerous studies have shown that PAs are indispensable in advanced plant growth and development processes, including floral bud differentiation, flowering regulation, reproductive organ development, sex differentiation, and fruit growth and senescence [16][17][18] . Five key biosynthetic enzymes involved in PA biosynthesis in higher plants have currently been investigated: arginine decarboxylase (ADC), ornithine decarboxylase (ODC), S-adenosylmethionine decarboxylase (SAMDC), Spd synthase (SPDS), and Spm synthase (SPMS) 19 . The significant physiological functions of endogenous PAs in floral induction and flower primordial and organ formation have been confirmed in numerous laboratory studies of horticultural crops 20,21 . The triamine Spd, the most important of the three above-mentioned PAs, is involved in many physiological processes. In Polianthes tuberosa 20 , Spd and Cad are markers of flower bud differentiation. Experiments on Cucurbita pepo L. 22 and Ginkgo biloba 23 have also found that Spd predominates during flower bud differentiation. When Spd reaches a certain level, it can promote the initiation of a flowering gene, thereby triggering the synthesis of a special protein and finally the formation of the flower primordium 24 . As a consequence, exogeneous Spd was used in the present study.
Gibberellins (GAs) and abscisic acid (ABA) are significant hormones in plant flowering pathways. In Arabidopsis, GAs play a pivotal role in floral induction, which involves a flowering regulatory pathway independent of GA 25 . Studies have shown that GA has different roles in flowering regulation in different plants. In Arabidopsis, for example, GA accelerates the transition from vegetative development to the first inflorescence stage of reproductive development 26 ; in contrast, GA has a negative effect on flower bud formation in woody trees such as mango 27 and apple 28 . Hedden and Thomas 29 have reported that the terminal steps of GA biosynthesis are mainly catalyzed by gibberellin 20-oxidase (GA20ox) and gibberellin 3-oxidase (GA3ox), whereas GA catabolism is facilitated by gibberellin 2-oxidase (GA2ox). ABA controls many aspects of development, such as seed dormancy, root and bud structure, flowering time, and senescence [30][31][32][33] . The active effect of ABA on flowering time acceleration and drought-escape response has recently been demonstrated in Arabidopsis and rice 32,34 . In transgenic Arabidopsis, NCED is reportedly adjusted by positive feedback through ABA 35 . ABFs are a family of ABA-responsive element binding factors. In the presence of ABA, SnRK2 can phosphorylate its downstream target ABFs (i.e., ABF1, ABF2, ABF3, and ABF4) 36,37 .
Given the above relationship of PAs and hormones to floral induction, we studied the molecular mechanism underlying the physiological response of 'Fuji' apple to exogenous Spd treatment. To advance understanding of PA and hormonal regulation of flowering, we examined the responses of endogenous PAs and hormones, floral genes, and genes related to PA and hormone synthesis following spraying with Spd before the cessation of shoot growth. We then used the results to propose how Spd regulates floral development in conjunction with floral and related genes. The ultimate goals of this research were to elucidate the effect of PAs on flower formation from a physiological-molecular perspective, further enrich basic biological knowledge of fruit tree flowering, and provide a theoretical basis for artificial regulation of apple flowering time.

Results
Effects of exogenous Spd on flowering rate. To study the effect of exogenous Spd on apple flowering, the flowering rate of short shoots was investigated in the second year after treatment. As shown in Fig. 1, the flowering percentage was higher in trees treated with Spd than in the control.
Effects of exogenous Spd on PA contents. After applying exogenous Spd, we simultaneously analyzed the contents of the three endogenous PAs, namely, Put, Spd, and Spm. Except at 40 DAF, the Put content of buds treated with Spd was significantly higher than that of the control until 60 DAF, with no significant differences observed thereafter ( Fig. 2A). At the early floral induction stage (40 DAF), Spd levels were significantly decreased in treated buds relative to the control and then significantly steadily increased (Fig. 2B). Spm levels, which declined over time in both groups, were significantly lower in the Spd-treated group compared with the control at 30 and 70 DAF (Fig. 2C).
As demonstrated by the results shown in Fig. 1, Spd application significantly affected the contents of endogenous PAs in all tested trees during at least one time point. Effects of exogenous Spd on flowering-associated gene expressions. To better understand molecular events associated with significant changes in the degree of flowering, we analyzed the expressions of genes controlling flowering in terminal spur buds. MdFD transcription was significantly up-regulated at 40 DAF after treatment with Spd (Fig. 4A). The expression pattern of the floral integrator gene MdFT was very similar to that of MdSOC1, with significant up-regulation and peak expression of both genes observed at 50 DAF in the Spd-treated group (Fig. 4B,C). In contrast, the expression of MdLFY remained lower than in the control, with exogenous Spd significantly down-regulating MdLFY transcription at 40 and 50 DAF compared with the control (Fig. 4D).
These results (Fig. 4) indicate that exogenous Spd application significantly affected all flowering-related genes at one or more time points, thus suggesting that these genes responded to Spd levels in buds.    www.nature.com/scientificreports www.nature.com/scientificreports/ from 60 to 80 DAF (Fig. 5B,C). A significant increase in MdODC1 transcript levels was observed in Spd-treated shoots from 40 to 60 DAF (Fig. 5D). In addition, the expression of MdSAMDC2 in Spd-treated buds mainly increased during bud development and peaked at 70 DAF (Fig. 5E).

Effects of exogenous
Effects of exogenous Spd on expressions of GA-and ABA-related genes. We analyzed two important synthetic genes in the gibberellin biosynthesis pathway. MdGA20ox2 and MdGA3 oxidase (GA3ox) expressions in Spd-treated buds were initially significantly higher than in control buds, with MdGA20ox2 and MdGA3ox subsequently down-regulated at 70 and 80 DAF (Fig. 6A,B). In addition, the expression level of MdGA2ox2, a gene involved in the inactivation of GA catabolites, was also high. In particular, MdGA2ox2 exhibited significantly up-regulated expression from 40 to 60 DAF in response to Spd treatment (Fig. 6C). These data indicate the possible existence of a negative feedback loop maintaining a GA steady state 29 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Several ABA signaling pathway genes were also analyzed, including NCED, a family of ABA biosynthetic enzymes 38 . NCED1 and NCED3 are key regulatory factors in ABA biosynthesis. In addition, ABF3 is a major transcription factor in the ABA signaling pathway. Notably, expression levels of MdNCED1 and MdNCED3 substantially increased in response to exogenous Spd. MdNCED1 expression peaked at 40 DAF, and MdNCED3 levels were highest at 80 DAF (Fig. 6D,E). In contrast, expression of MdABF3 decreased, but the difference was not significant. (Fig. 6F).

Discussion
Numerous studies have revealed that flower bud differentiation is affected by a combination of in vitro and in vivo factors. Evidence suggesting a close connection between PAs and the physiological events leading to flowering has been summarized by Galston and Kaur-Sawhney 39 , Kakkar and Rai 40 , and Galston et al. 41 . In the present study, we investigated the roles of PAs in flower development in 'Fuji' apple. We observed a correlation between the concentration of endogenous PAs and hormones and the expression of related genes during floral development.
Effect of exogenous Spd on flower formation. The application of Spd has been previously shown to accelerate flower bud formation 20 , a finding further confirmed by our experimental results (Fig. 1). After spraying with Spd, the flowering rate of short shoots increased, which suggests that Spd treatment is beneficial to flower bud formation. At the molecular level, plant flowering is regulated by complicated networks involving endogenous and environmental signals 42 . In our study, the transcriptional activities of floral integrators MdFD (Fig. 6A), MdFT (Fig. 6B), and MdSOC1 (Fig. 6C) were enhanced by exogenous Spd. A rise in transcription levels of an FT-encoding gene has been connected with floral induction in several species 43 . According to our results, Spd applied before floral induction can hasten flowering in 'Fuji' apple. To our surprise, we observed a reduction in MdLFY expression during floral induction in apple buds treated with Spd (Fig. 5D). We speculate that MdLFY is subject to other modes of regulation in apple, but this hypothesis requires further testing.
PA signaling regulates floral induction in Spd-treated buds. The significance of PAs in flowering development has received much attention 44 . Because PA biosynthesis is transcriptionally regulated, we analyzed the expressions of genes related to PA biosynthesis in Spd-treated shoots. As inferred from results displayed in Figs 2A-C and 5A-F, we discovered that exogenous Spd significantly impacted the contents of endogenous PAs and related gene expressions in all tested materials at some time point(s). According to our results, Spd boosted the level of endogenous Spd in buds during the late stage of floral induction (from 60 to 80 DAF); however, no significant difference was found in MdSPDS1 expression during floral induction, and MdSPDS2 expression even significantly declined. The increase in endogenous Spd levels in Spd-treated buds may be mainly due to the permeation of exogenous Spd in addition to the synthesis of new Spd. In plants, Put is biosynthesized from arginine via ADC or from ornithine by way of ODC 20 . We propose that the presence of Spd augmented the content of Put because exogenous Spd significantly increased MdADC1 and MdODC1 expression and even MdSAMDC2 transcripts. Conversely, we observed a depression in the Spm content of Spd-treated buds, a finding in accord with previous studies where exogenous Spd decreased endogenous PA contents during flowering 45 . We conclude that feedback inhibition caused by high levels of endogenous PAs takes place after application of exogenous Spd. This finding contrasts with the results of Huang 20 , who found that early floral initiation is associated with a rise in free Spd and a drop in free Put and Spm in P. tuberosa. These conflicting results suggest that the effects of PAs on flowering vary depending on species, organ, and physiological process. Our results collectively demonstrate that endogenous PA contents and the expressions of related genes change markedly and regularly after exogenous Spd application along with developmental stages of flower bud differentiation, thereby affecting flowering.

Expression of GA genes in Spd-treated buds.
To provide evidence for the influence of applied Spd on the metabolism of endogenous hormones in terminal spur buds, we measured endogenous hormone concentrations and transcript levels of hormone biosynthetic genes. Hormones synergistically or antagonistically play a pivotal regulatory role in plant growth and development 1 . GA is one of the most studied hormones in the flowering process in plants, and the mutual effect of PAs and GAs has been preliminarily researched in species such as Arabidopsis and tomato [46][47][48] . Alcázar et al. 46 have reported that a high concentration of endogenous Spd/Spm in 35S:AtADC2 Arabidopsis plants is accompanied by a significant drop in AtGA3ox3 and AtGA20ox1 expression and a decrease in GA production. In addition, a significant increase in GA2ox expression and a decrease in GA content have been found in E8:ySAMDC transgenic tomato fruits, which had three to four times more Spd compared with the control 48 . In our study, however, we discovered that Spd application dramatically augmented the GA content of buds by increasing transcript levels of the GA biosynthetic genes MdGA20ox2 and MdGA3ox during the initial process of floral induction (30 to 50 DAF) while simultaneously significantly increasing transcript levels of the GA catabolism gene MdGA3ox. This result suggests the existence of a negative feedback loop to maintain GA balance 29 . Next, we found that Spd application lowered the expressions of MdGA20ox2 and MdGA3ox and increased the expression of MdGA2ox2, which significantly restrained GA biosynthesis from 60 to 80 DAF. Hence, the effects of Spd on GA vary depending on physiological processes, and gene family members have diverse reactions to Spd stimulation. As a high GA content inhibits flowering 49 , we conclude that exogenous Spd promotes flower formation by reducing GA in the later stages of physiological differentiation. The specific mechanism remains unclear, however, and requires further research.

Expression of ABA genes in Spd-treated buds. A previous finding that ABA contents of leaves and
shoots change significantly during the induction phase of apple flowering 50 suggests that ABA participates in the regulation of flowering induction in perennial woody plants. NCED3 has different functions in different species. In Brassica napus, for example, BnNCED3 has been discovered to participate in both stress adaptation and plant development 51  www.nature.com/scientificreports www.nature.com/scientificreports/ increases 52 . The interplay between ABA and PAs during seed maturation and germination has also been thoroughly researched 53 . Further evidence highlighting the interaction between PA anabolism and the ABA signaling pathway has been obtained in grape, a perennial plant 46 . ABA content is positively correlated with the expression level of NCED 51 . Our data indicate that Spd application significantly enhances the expressions of MdNCED1 and MdNCED3, in turn increasing ABA levels during the late stage of floral induction. ABF3 has redundant functions in ABA signaling 54 . In our study, exogenous Spd decreased the expression of MdABF3. We thus infer that Spd regulates flowering by depressing MdABF3 expression. Results reported by Keumbi Hwang et al. 55 support the idea that ABA is involved in flowering regulation to speed the floral transition. Increased ABA is thus considered to play a crucial role in flowering regulation. Our results indicate that Spd promotes flowering to a certain extent and is closely related to ABA metabolic regulation at the transcriptional level.

Conclusions
The current study revealed the existence of an intricate regulatory mechanism that manages floral induction in 'Fuji' apple. The contribution of Spd to the promotion of flowering might be closely associated with the metabolism of hormones and floral integrators (Fig. 7). How PAs regulate flowering genes and influence 'Fuji' flowering, however, still needs further study. Analysis of flowering rate. The flowering rate was calculated as described by Zuo et al. 56 with sight modifications. Specifically, two large branches on each Spd-treated or control tree were labeled, and the flowering rate of terminal buds on their short shoots (<5 cm) were recorded. During full bloom, on April 10, 2019, the total number of terminal buds on short shoots and the number of floral terminal buds on the marked branches were counted. The flowering rate (number of floral buds/number of total buds) based on these data was then calculated.

Materials and Methods
Extraction and measurement of endogenous PAs in spur apical buds. PAs were estimated using high-performance liquid chromatography (HPLC) as described by Flores and Galston 57 with minor modifications. Briefly, frozen tissue (0.3 g) from each treatment was mixed with 3 mL cold 5% (w/v) perchloric acid in a centrifuge tube. The homogenate was chilled in an ice bath for 1 h and then centrifuged at 15,000 ×g for 30 min at 4 °C. Next, 1 mL of 2 M NaOH solution was combined with 1 mL supernatant, and 10 µL benzoyl chloride was added. The mixture was incubated for 20 min at 37 °C, and 2 mL saturated sodium chloride was then added to terminate the reaction. Benzoyl-PAs were extracted into 2 mL diethyl ether by centrifugation at 4,500 ×g for 5 min. 1 mL of the ether phase was collected and evaporated to dryness with a nitrogen blower. The dried extracts were dissolved in 500 µL of 60% methanol for further testing. www.nature.com/scientificreports www.nature.com/scientificreports/ The above-prepared benzoyl-PA extract was filtered through a 0.22-μm membrane filter and then eluted on a C18 reverse-phase column (6.0 mm × 150 mm; particle size 5 mm) at room temperature. The mobile phase was composed of HPLC-grade methanol, HPLC-grade acetonitrile, and ddH 2 O (58:2.5:39.5, v/v/v). The flow rate was 1.0 mL/min, the detection wavelength was 254 nm, and the column temperature was 30 °C. Put, Spd and Spm PA standards were used to construct standard curves. The PA analyses were repeated three times.
Quantitative analysis of endogenous hormones in spur apical buds. GA 3 and ABA in terminal spur buds were measured by liquid chromatography as described previously 58 with slight modifications. Specifically, 0.3 g of each treated fresh bud sample in 20 mL of 70% cold acetone was maintained on ice for 1 h. After centrifugation at 8,000 ×g and 4 °C for 10 min, the supernatant was collected and concentrated using a nitrogen evaporator. Next, extraction was performed by addition of 10 mL of petroleum ether, decolorization for 10 min, and removal of the ether phase. This procedure was carried out three times. Ten milliliters of ethyl acetate was added and collected after 10 min, and this step was repeated once more. The two collections of the ester phase were concentrated with a nitrogen blower until no solvent remained. The dried extracts were dissolved in 1.5 mL methanol and filtered through a 0.22-μm filter for future analysis.
The liquid-chromatography mobile phase consisted of HPLC-grade methanol and 0.1% formic acid. The flow rate was 1.0 mL/min, and the column temperature was 30 °C. GA 3 and ABA (Sigma) standards were used to construct standard curves. Three biological replicates of GA 3 and ABA analyses were performed. RNA isolation and cDNA synthesis from spur apical buds. Total RNA was isolated using a cetyltrimethylammonium bromide method 59 with minor modifications.