Overexpression of Prunus DAM6 inhibits growth, represses bud break competency of dormant buds and delays bud outgrowth in apple plants

Most deciduous fruit trees cultivated in the temperate zone require a genotype-dependent amounts of chilling exposure for dormancy release and bud break. In Japanese apricot (Prunus mume), DORMANCY-ASSOCIATED MADS-box 6 (PmDAM6) may influence chilling-mediated dormancy release and bud break. In this study, we attempted to elucidate the biological functions of PmDAM6 related to dormancy regulation by analyzing PmDAM6-overexpressing transgenic apple (Malus spp.). We generated 35S:PmDAM6 lines and chemically inducible overexpression lines, 35S:PmDAM6-GR. In both overexpression lines, shoot growth was inhibited and early bud set was observed. In addition, PmDAM6 expression repressed bud break competency during dormancy and delayed bud break. Moreover, PmDAM6 expression increased abscisic acid levels and decreased cytokinins contents during the late dormancy and bud break stages in both 35S:PmDAM6 and 35S:PmDAM6-GR. Our analysis also suggested that abscisic acid levels increased during dormancy but subsequently decreased during dormancy release whereas cytokinins contents increased during the bud break stage in dormant Japanese apricot buds. We previously revealed that PmDAM6 expression is continuously down-regulated during dormancy release toward bud break in Japanese apricot. The PmDAM6 expression pattern was concurrent with a decrease and increase in the abscisic acid and cytokinins contents, respectively, in dormant Japanese apricot buds. Therefore, we hypothesize that PmDAM6 represses the bud break competency during dormancy and bud break stages in Japanese apricot by modulating abscisic acid and cytokinins accumulation in dormant buds.

Introduction Perennial woody plants in temperate zones synchronize their annual growth patterns with seasonal environmental changes. This allows plants to avoid injuries from environmental stresses, such as cold conditions in the winter. Dormancy is a control mechanism that enables woody perennials to adapt to seasonal environmental changes. Bud dormancy refers to the inability of the meristem to resume growth [1]. Lang [2] and Lang et al. [3] defined plant dormancy as "the temporary suspension of visible growth of any plant structure containing a meristem" and classified the fruit tree bud dormancy states as paradormancy, endodormancy, and ecodormancy. Both paradormancy and endodormancy are states induced by the perception of environmental or endogenous signaling cues, but they differ regarding whether they originated solely from meristem-containing tissue (endodormant) or from a structure distinct from the one undergoing dormancy (paradormant). A certain amount of chilling exposure is critical for inducing the shift from endodormancy to ecodormancy. Ecodormancy is a state brought about by the limitation of growth-promoting factors, such as warm conditions and the availability of water and nutrients. Although Lang's definition has been widely adopted by researchers of plant bud dormancy, recently accumulated data suggest this terminology may need to be revised [4,5]. For example, discriminating between paradormancy and endodormancy, and endodormancy and ecodormancy, is problematic for fruit tree species of the genus Prunus, such as peach (P. persica) and Japanese apricot (P. mume), in which lateral flower and leaf buds are often used for dormancy research. In these fruit tree species, whether buds are paradormant, endodormant or ecodormant has been estimated according to the competency of bud break, which is often based on the mean time to bud break or bud break percentage in forcing condition [4,6,7,8,9]. Without the experiments such as incubating branch or single node cuttings in forcing condition, we cannot precisely distinguish paradormancy, endodormancy and ecodormancy in field trees.
The genetic and molecular regulation of bud dormancy has been extensively studied in the model woody perennial, poplar (Populus spp.), and much has been learned, as reviewed [1,10,11]. Additionally, molecular networks regulating the dormancy of various woody species, including horticulturally important fruit tree species, have been characterized based on omics studies of specific plant species [4]. In a previous study of Japanese apricot, which belongs to the family Rosaceae, Yamane et al. [12] applied an RNA subtraction technique to identify the genes expressed preferentially in endodormant buds (no bud break under forcing conditions) compared with paradormant (higher bud break frequency under forcing conditions compared with that of endodormant buds) and ecodormant buds (bud break frequency greater than 50% under forcing conditions). This study identified a MADS-box gene with dormancy-associated expression. This gene was similar to the Arabidopsis thaliana StMADS11 clade, which includes SHORT VEGETATIVE PHASE (SVP) and AGAMOUS-LIKE24 (AGL24) [12]. In peach, a mutant that continues to grow and fails to enter dormancy under dormancy-inducing conditions has been identified. This mutant, which is known as evergrowing (evg; USDA PI442380), was first identified in southern Mexico [13]. Sequencing and expression analyses of the evg locus identified six StMADS11 (SVP/AGL24)-clade MADS-box genes that may be associated with terminal bud formation [14]. These genes were named as DORMANCY-ASSOCIATED MADS-box 1-6 (DAM1-6) genes. The gene that Yamane et al. [12] detected in Japanese apricot RNA subtraction study appears to be an ortholog of peach DAM6, and was named PmDAM6.
In the Japanese apricot genome, six tandemly arrayed PmDAM genes (PmDAM1-6) have been identified [15,16]. A seasonal expression analysis of the PmDAM genes in a reverse transcription-quantitative PCR (RT-qPCR) [15], genome-wide transcriptomic analyses involving the Japanese apricot expressed sequence tag (EST) dormant bud database [17], and a 60 Kmicroarray analysis [18] revealed that PmDAM genes are preferentially expressed in dormant leaf buds, in summer (PmDAM1-3) and autumn (PmDAM4-6). Moreover, the expression levels of these genes are down-regulated during the dormancy release toward bud break of the lateral leaf buds. Furthermore, RT-qPCR and microarray analyses indicated that the expression levels of all six PmDAM genes are down-regulated following a prolonged exposure to artificial cold conditions [15]. Kitamura et al. [19], Zhang et al. [20], and Zhong et al. [21] also demonstrated that PmDAMs are expressed in endodormant flower buds in November and their expression levels are down-regulated during dormancy release and bud break. Meanwhile, Zhao et al. [22] examined flower buds, and highlighted the functions of PmDAMs in flower development as well as dormancy induction. The peach and Japanese apricot DAM6 expression levels in buds are negatively correlated with increasing bud break competency (bud-burst frequency in forcing conditions). In peach, DAM6 expression was negatively correlated with the time required for terminal bud break [23]. A negative correlation between peach PpDAM6 expression and the time required for bud break under forcing conditions was also reported for lateral leaf [24] and flower [25,26] buds. Among the other rosaceous perennials and temperate fruit trees, the down-regulated expression of the DAM-like genes during dormancy release and bud break has been reported for the lateral buds of raspberry (Rubus idaeus L.) [27], the lateral leaf and flower buds of pear (Pyrus pyrifolia) [28,29], and the terminal buds of apple (Malus × domestica). In apple, Porto et al. [30] proposed that the 'Golden Delicious' apple genome contains four DAM-like genes (MdDAM1-4). Expression analyses suggested that the MdDAM genes undergo seasonal expression-level changes, with an up-regulation during endodormancy [30][31][32][33]. The overexpression of MdDAMb, which is a SVP/DAM-like gene and up-regulated during ecodormancy [34], delays bud break [35], suggesting that DAM genes may affect bud break in both apple and Prunus spp.
Most studies regarding Prunus DAM genes have focused on expression levels, and have revealed a correlation between DAM expression levels and dormancy depth (days to bud break under forcing conditions). Among the DAMs in Japanese apricot, PmDAM6 is the most promising candidate regulator of dormancy because its expression is correlated with the endodormancy release and ecodormancy release (increased bud break frequency under forcing condition). Moreover, the results of our earlier comprehensive transcriptome analysis suggested that PmDAM6 showed the biggest difference in the expression-level fold change by artificial chilling exposure [17,18]. Thus, validating the function and clarifying the downstream factors regulated by PmDAM6 are necessary to characterize the mechanism controlling dormancy in Japanese apricot buds. To elucidate the biological functions of PmDAM6, hybrid poplar (Populus tremula × Populus tremuloides; clone T89) plants constitutively expressing PmDAM6 under the control of the cauliflower mosaic virus 35S promoter (35S:PmDAM6) were generated. Additionally, their phenotypes were compared with those of control plants, which were either wild-type poplar or poplar transformed with the empty vector [15]. When grown under long photoperiod (LP) conditions (16-h light/8-h dark), the shoot growth of 35S: PmDAM6 poplars was inhibited. Additionally, 35S:PmDAM6 poplars set terminal buds earlier than the control poplars. Thus, PmDAM6 may function as a growth inhibitor in poplar and induce bud set, which is a dormancy-related trait. We then attempted to obtain Japanese apricot overexpressed or RNA interference (RNAi)-suppressed transformants. However, we obtained an insufficient number of lines for a meaningful analysis of phenotypes [36]. Therefore, in this present study, we used transgenic apples to functionally validate the role of PmDAM6 because Japanese apricot is phylogenetically closer to apple than to poplar. Moreover, easily transformed apple lines are available. Additionally, the existence of MdDAMs, which are phylogenetically similar to Prunus DAMs, imply that similar DAM-controlled downstream factors and pathways might be conserved in apple and Prunus species. Thus, apple transformants may be useful for a detailed analysis of the biological functions of Prunus DAMs. On the basis of the observed phenotypes and hormone contents in the dormant buds of transformed PmDAM6-overexpressing apple lines, we herein discuss the biological functions of PmDAM6 and present a working hypothesis regarding how PmDAM6 regulates dormancy and bud break in Japanese apricot.

Apple transformation
In this study, we used the binary vectors p35S:PmDAM6 [15] and p35S:PmDAM6-GR, in which the PmDAM6 of the former vector was substituted by a fusion between PmDAM6 and the hormone-binding domain of a mouse glucocorticoid receptor [37]. In p35S:PmDAM6-GR transgenic apples, PmDAM6-GR was localized in the cytoplasm in the absence of glucocorticoid, and was recruited to the nuclei only when cells were treated with dexamethasone (DEX), which is a glucocorticoid compound. These vectors were introduced into Agrobacterium tumefaciens strain LBA4404, which were subsequently used to transform an apple rootstock cultivar, 'JM2' [38]. Apple transformations were conducted as described by Wada et al. [39]. Agrobacterium tumefaciens cells was cultured overnight on a shaker in 20 mL of ψB medium with 50 mg/L kanamycin at 28˚C. After a centrifugation, the pellet was resuspended in halfstrength Murashige and Skoog (MS) medium, and further diluted to an optical density at 600 nm (OD 600 ) of 0.8. Folded small leaf explants were soaked in the Agrobacterium solution, immediately sonicated (UT-105S; Sharp Corporation, Osaka, Japan) for 10 s, and incubated for 15 min. Then, the inoculated explants were incubated for 1 week on the co-cultivation medium N6+MS basic medium (pH 5.6) [38] containing 0.2 mg L −1 NAA, 5 mg L −1 BAP, 30 g L −1 sorbitol, 2.6 g L −1 phytagel (Sigma-Aldrich, St. Louis, MO, USA), and 20 μM acetosyringone, at 22˚C in the dark. The explants were transferred to a selective medium, which was the same as the co-cultivation medium except it contained 25 mg L −1 kanamycin and 50 mg L −1 meropenem trihydrate (MEROPEN; Dainippon Sumitomo Pharma Co., Ltd., Osaka, Japan) and lacked acetosyringone. After 1 month, explants were incubated at 22˚C under a 16-h light/ 8-h dark photoperiod and were subcultured monthly. Regenerated shoots were grown on proliferation medium [basic MS medium (pH 5.8) containing 0.1 mg L −1 indole-3-butyric acid, 0.5-1 mg L −1 6-benzylaminopurine, 30 g L −1 sucrose, 7 g L −1 bactoagar, 25 mg L −1 kanamycin, and 50 mg L −1 meropenem trihydrate]. The regenerated shoots were transferred to sterilized vermiculite moistened with the basic MS medium supplemented with 0.1 mg L −1 indole-3-butyric acid and 30 g L −1 sucrose (pH 5.8). When the plants rooted, they were moved to plastic pots containing sterilized vermiculite moistened with 1/1,000 Hyponex (Hyponex Japan, Osaka, Japan) and covered with a plastic bag. After several weeks of acclimatization, the plastic bags were removed, and the plants were transplanted to larger pots (10.5 cm diameter Y pot, Sakata Seed Corporation, Yokohama, Japan). All plants were cultivated in a growth chamber or greenhouse with standard procedures, including pesticide, fungicide sprays, and fertilizer treatments.
The success of apple transformations was confirmed by a DNA blot analysis as previously described [40]. Briefly, genomic DNA was isolated from leaves with a plant DNA isolation kit (NR-501; Kurabo, Osaka, Japan). Genomic DNA was digested with HindIII, run on a 0.8% agarose gel, and transferred to a Biodyne PLUS membrane (Pall, Port Washington, NY, USA). The membrane was hybridized with a digoxigenin-labeled PmDAM6 probe [36].
After an acclimatization step, apple plants from two transgenic lines, 35S:PmDAM6 (35S-2 and 35S-4) and 35S:PmDAM6-GR (GR21 and GR22), as well as wild-type (WT) control apple plants (n = 10-17 per line) were grown under LP conditions (22˚C with a 16-h light/8-h dark photoperiod) for 4 months beginning in December 2011. In May 2012, after all plants had stopped growing, leaves were collected for a subsequent analysis of PmDAM6 expression levels. Total RNA was isolated, and 1 μg was used as the template for a cDNA synthesis step, as described previously [15]. The cDNA solution synthesized from approximately 20 ng of total RNA was used for RT-qPCR, which was performed using LightCycler 480 (Roche, Basel, Switzerland) and a probe master mix (Roche). Primers and probes used for the amplification of PmDAM6 [15] are shown in S1 Fig. As a reference, an apple SAND (MDP0000185470 and/or MDP0000202305) [41] gene was monitored by RT-qPCR using SYBR Green Master mix (Roche) and gene-specific primers for MdSAND (S1 Table) for normalization. PCR was performed using a program of 45 cycles at 94˚C for 10 s, 55˚C for 20 s, and 72˚C for 1 s, with initial denaturing at 95˚C for 5 min. For SAND gene-specific qPCR, a dissociation curve analysis was performed to confirm that the fluorescence was only derived from gene-specific amplification. Three biological replicates were analyzed.

Observation of growth and bud dormancy states of 35S:PmDAM6 apple plants
In In December 2015, 5-year-old WT and 35S-4 (n = 3) plants grown in the greenhouse were moved to a cold room (5˚C in darkness). After a 30-day or 60-day cold treatment, the plants were exposed to forcing conditions (> 15˚C in a greenhouse). The number of days to bud break was determined for the terminal buds. This experiment was repeated in December 2016 using 6-year-old plants.
On 23 February and 8 March, 2017, the current year's shoots that were approximately 15 cm long (n = 5; 1 to 2 shoots each from 3 plants per genotype) were collected from 6-year-old 35S-4 and WT plants. The basal parts of shoots were soaked in water containing 1% Misakifarm (OAT Agrio, Tokyo, Japan), which is a cut-flower freshness preservation agent containing nutrients and fungicides, and then incubated under forcing conditions (22˚C with a 16-h light/8-h dark photoperiod). The basal parts of shoots were cut once per week to promote water uptake, and the solution was replaced weekly. The bud break rate (%) was recorded for 1 month.
For the phytohormone analysis, terminal buds were collected from greenhouse-grown 4-year-old plants on 20

Observation of bud dormancy status of DEX-treated 35S:PmDAM6-GR apple plants
In November 2013, after the trees had shed their leaves, 3-year-old GR21, GR22, and WT plants (n = 3-5) were pruned, leaving only one main shoot from the current year, and moved from the greenhouse to a cold room (5˚C in darkness). After 4 weeks, the plants were transferred to forcing conditions (22˚C with a 16-h light/8-h dark photoperiod) and grown for 4 months. The 20 μM DEX solution containing 0.1% Tween 20 and 0.1% ethanol or the control solution containing 0.1% Tween 20 and 0.1% ethanol were applied every 3 days to the soil of pot-grown plants being grown under forcing condition. The number of days to bud break for the terminal buds and the number of lateral buds that had burst were determined. The length of the shoots sprouting from terminal buds and bud set rate (%) were recorded after 4 months under forcing conditions. Plants were photographed on 7 March 2014.
After the 2013-14 season, uniform DEX applications were not possible owing to the increasing plant sizes. Therefore, we allowed the branches to take up the DEX solution by soaking the basal parts of branches in 20 μM DEX. This treatment method prevented us from observing phenotypes over long periods. Dormant shoots require relatively long incubation periods for the to uptake liquids taken up at the basal parts of cut branches to reach the terminal buds and influence phenotypes. Thus, we were unable to determine the effects of DEX on the dormancy stage. Consequently, we tested the effects of DEX on dormant buds during the bud break stage. On 6 March 2017, the current year's shoots that were approximately 15 cm long (n = 6-10, 1 to 2 shoots from 5 plants per genotype) were collected from 6-year-old GR21, GR22, and WT plants. The basal parts of shoots were soaked in 20 μM DEX or control solutions and incubated under forcing conditions (22˚C with 16-h light/8-h dark photoperiod). For the phytohormone analysis, terminal buds were collected after 0, 24, and 96 h for GR21 and after 0 and 96 h for WT under forcing conditions, immediately frozen in liquid N 2 , and stored at -80˚C until analyzed. After a 7-day incubation, the bottom parts of shoots were cut to promote water uptake, and solution was replaced by a fresh solution. After a 15-day incubation, the shoots sprouting from buds were photographed, and the blade lengths of the three most developed leaves sprouting from the buds were measured.

Investigation of the phytohormone contents of the terminal buds of transgenic apple plants
Terminal bud samples were freeze-dried and ground to a fine powder using a Multi-beads shocker (Yasui Kikai Co., Osaka, Japan). For 35S:PmDAM6 apple plants, plant hormones were extracted from 20-mg samples in a methanol-formic acid buffer (methanol:water:formic acid = 15:4:1) containing a deuterium-labeled hormone internal standard [i.e., 40 ng d 6abscisic acid (ABA); 4 ng d 5 -indole acetic acid (IAA); 0.4 ng d 5 -trans zeatin (tZ), and d 6 -isopentenyl adenine (iP)]. Hormones were extracted overnight at 4˚C in darkness. Supernatants were purified with Oasis HLB columns (Waters Co., Milford, MA, USA) and then evaporated. The dry resulting pellets were resuspended in 1 M formic acid and then added to Oasis MCX columns (Waters). Acid and basic fractions were extracted in 100% methanol and 0.35% NH 3 in 60% methanol, respectively. Samples were evaporated, and the remaining pellets were resuspended in reconstitution solution (acetonitrile:water:formic acid = 85:15:0.1). Phytohormone contents were determined by liquid chromatography-triple quadrupole mass spectrometry (Waters; liquid chromatography system: Waters 2695, mass spectrometer: Quattro micro ARI Waters 2996).
For GR21 apple plants, phytohormone contents were analyzed as previously described [42] with minor changes. Briefly, harvested tissue samples were lyophilized and the exact dry weights (around 20 mg per sample) were determined. To avoid the effects of the ion suppression of salicylic acid (SA) on apple bud samples, we modified the solution used during the SA elution step. After sample purification by HLB and MCX columns, the acidic fraction was loaded onto an Oasis WAX column (Waters), and the main acidic fraction was eluted with 1% AcOH and 80% MeCN. A solution comprising 3% formic acid and 97% MeCN was then applied to the WAX column to elute SA. The eluates were evaporated to dryness, reconstituted in 1% AcOH and subjected to phytohormone quantification using an Agilent 1260-6410 Triple Quad LC/MS system (Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a ZORBAX Eclipse XDB-C18 column and an XDB-C8 Guard column (Agilent Technologies Inc.). In this analysis, IAA, tZ, iP, ABA, jasmonic acid (JA), JA-isoleucine (JA-Ile), and SA contents were determined.

Investigation of phytohormone contents in dormant leaf and flower buds of Japanese apricot
Japanese apricot 'Nanko' grown at the experimental farm of Kyoto University, Kyoto, Japan (34˚N, 135˚E) were used in this study. Leaf buds and flower buds collected monthly from September to March during the 2005-06 were immediately frozen in liquid nitrogen and stored at −80˚C until used for plant hormone analyses. Phytohormone contents were determined as previously described [43] without MS Probe modification. In this analysis, IAA, tZ, trans zeatin riboside (tZR), iP, isopentenyl adenine riboside (iPR), and ABA contents were determined.

Shoot growth and dormancy induction phenotypes of 35S:PmDAM6 and 35S:PmDAM6-GR apple plants
When the p35S:PmDAM6 binary vector was used to transform the apple rootstock cultivar 'JM2', the transformation efficiency was 0.25%, which was approximately 10 times lower than usual (approximately 1%-5%). We obtained two 35S:PmDAM6 lines, 35S-2 and 35S-4. A genomic DNA blot analysis using the PmDAM6 probe indicated that a single copy was inserted into each transgenic line (S1 Fig). Additionally, PmDAM6 mRNA was detected in both 35S-2 and 35S-4, with a higher expression level in 35S-4 than in 35S-2 (S1 Fig). Moreover, PmDAM6-GR mRNA was also detected in two 35S:PmDAM6-GR lines used in this study.
Shoot growth was inhibited in 2-year-old 35S:PmDAM6 plants under LP conditions ( Fig  1A) and in 3-year-old plants under SP conditions (Fig 1B). The shoots stopped growing in all 35S:PmDAM6 and WT plants terminated their shoot growth in even under LP conditions. The node numbers and internode lengths were significantly lower in the 35S:PmDAM6 plants than in the WT plants (P < 0.01) (S2 Table). There was no major difference between the 35S: PmDAM6 and WT plants regarding the exact timing of terminal bud set even under SP condition. When 35S:PmDAM6 and WT plants were exposed to a cool conditions (10˚C), leaf shedding was observed at around the same time. Reduced plant size of 35S:PmDAM6 plants compared to WT plants were observed even in 5-year-old plants (Fig 1C).
In contrast to the 2-and 3-year-old plants, 4-year-old 35S:PmDAM6 plants set terminal buds approximately 1 week earlier than WT plants (Fig 2). Inhibited shoot growth and earlier bud set was also observed in 4-year-old PmDAM6 activity-inducible transgenic lines, GR21 and GR22, when they were transferred to forcing conditions after bud break and treated with DEX continuously for 4 months (S3 Table). DEX treatment did not affect these traits in WT plants (S3 Table). The timing of leaf shedding for the 35S:PmDAM6 plants was similar to that of WT plants throughout the observation period.

Bud break phenotypes and the dormancy depth of transgenic apple plants overexpressing PmDAM6
After plants had adapted to the growing conditions in a greenhouse without additional heating (semi-field conditions) in the Kyoto, Japan climate, we determined terminal bud break dates  Table). There were fewer open lateral buds in 35S:PmDAM6 plants than in WT plants (Fig 3A and 3B, S4 Table). Furthermore, for 3-yearold 35S:PmDAM6-GR plants, terminal bud break occurred significantly earlier in DEXtreated plants than in the control plants (Table 1). Additionally, significantly fewer lateral buds underwent bud break in DEX-treated plants than in control plants ( Table 1) Fig 4).
The dormancy depth of terminal buds was estimated by counting the days to bud break under forcing conditions of 35S-4 and WT plants using 5-and 6-year-old plants in the 2015-16 and 2016-17 seasons, respectively. As shown in Table 2, dormancy was significantly deeper (P < 0.01) in 35S-4 plants than in WT plants after the same amount of chilling exposure. In other words, bud break competency was significantly repressed in 35S-4 plants relative to that in WT plants. A similar difference was observed between 35S-4 and WT plants during the following growing season (Table 2).

Phytohormone contents in the terminal buds of transgenic apple plants overexpressing PmDAM6 from dormancy through the bud break stage
The terminal buds of 4-year-old 35S-4 plants collected in February 2015 (late dormancy period) contained more ABA than the corresponding WT buds, whereas there were no significant differences in the cytokinins (CK) contents ( Fig 5A). Among 6-year-old plants, the ABA content of terminal buds was greater in 35S-4 plants than in WT plants during the late dormancy stage, especially at the beginning of March (Fig 5B). In contrast, the contents of two CKs, tZ and iP, were lower in 35S-4 plants than in WT plants in March (Fig 5B). Regarding the samples collected on 25 January 2014, tZ was undetectable in both 35S-4 and WT plants. The IAA level was lower in 35S-4 plants than in WT plants, although the difference was not

35S:PmDAM6-GR line# 21 (GR21)
35S  Table). The dormancy states of these buds were briefly compared by determining the days to bud break under forcing conditions. More than 50% of WT and 35-4 shoots collected on 8 March underwent bud break in 1 and 2 weeks, respectively, suggesting that the dormancy was deeper and bud break competency was lower for the 35S-4 plants than for the WT plants in March.
In the 2016-17 season, the dormancy states of 6-year-old DEX-treated GR21 and GR22 plants were also determined. Bud outgrowth was significantly retarded in DEX-treated GR21 and GR22 lines compared with the control plants (Fig 6). The DEX treatment did not affect the bud outgrowth rate or the phytohormone contents in terminal buds of WT plants (Fig 6). However, DEX-treated GR21 plants accumulated more ABA and less iP and tZ by 96 h after the treatment compared with the control plants (Fig 6). In particular, the iP content was significantly decreased by DEX at 24 h after the treatment. There were no significant differences between DEX-treated and control plants regarding the contents of the other phytohormones, IAA, JA, JA-Ile, and SA.

Seasonal changes in the phytohormone contents of Japanese apricot dormant buds
Our above described analysis suggested that overexpression of PmDAM6 represses bud break competency in 4-to 6-year-old transgenic apples and also increased ABA levels during dormancy and bud break stage and decreased cytokinin contents during bud break stage. To address whether PmDAM6 also increased ABA levels and decreased cytokinins in Japanese apricot leaf and flower buds, we aimed to clarify the relationship between PmDAM6 expression changes and phytohormone contents of Japanese apricot leaf and flower buds during endodormancy release toward bud break. Several previous studies reported PmDAM6 was upregulated during endodormancy and then down-regulated toward bud break in leaf buds and flower buds [4,15,18,19].
Previous studies revealed that Japanese apricot 'Nanko' flower and leaf buds gradually shift from endodormancy to ecodormancy in December and January and bud break occurs in February and end of March, respectively, under the climate conditions where samples were collected [15,19]. In the current study, the ABA contents in buds steadily decreased after peaking in October (Fig 7). Decreases in the ABA content were synchronized with the shift from endodormancy to ecodormancy in both flower and leaf buds and also from ecodormancy to bud break in leaf buds. This suggested that the ABA concentrations in buds are associated with the depths of endodormancy and ecodormancy. The IAA contents decreased in buds as the season progressed, but increased just before bud break (Fig 7). The iP levels increased slightly during endodormancy release and increased further during ecodormancy release until bud break (Fig 7). The tZ contents did not undergo any major changes until February and January in leaf and flower buds, respectively, and then rapidly increased just before bud break (Fig 7). The results suggested that PmDAM6 expression changes [15,19] may be positively correlated with ABA contents in Japanese apricot leaf and flower buds. Additionally, increase in cytokinins was detected in bud break stage when PmDAM6 expression was undetectable level in leaf and flower buds [15,19]. Therefore, the obtained results appeared not contradictory to our hypothesis that PmDAM6 may mediate increase and decrease in the abscisic acid and cytokinins contents, respectively, in dormant Japanese apricot buds.

Discussion
To investigate the role of PmDAM6 in dormancy regulation, we used apple, a relative of Japanese apricot, both belonging to Rosaceae, because there are easily transformable apple accessions [38]. However, the efficiency of the transformation of apple plants with the p35S: PmDAM6 construct was very low, suggesting that PmDAM6 may inhibit adventitious shoot   Prunus DAM6 represses bud break competency and delays bud outgrowth regeneration in apple. The low regeneration efficiency after the overexpression of SVP-like or DAM-like genes in apple was also reported for MdSVPa and MdDAMb [35]. Nonetheless, two PmDAM6-overexpressing transgenic lines were obtained. We also generated transgenic apple lines that expressed a fusion between PmDAM6 and the hormone-binding domain of a glucocorticoid receptor (PmDAM6-GR). In these transgenic apple plants, PmDAM6-GR was activated by DEX treatments, as developed by Aoyama and Chua [37]. This system allowed us to evaluate the function of PmDAM6 during dormancy release, while avoiding its effects on shoot growth prior to dormancy establishment.

PmDAM6 inhibited growth regardless of plant age and induced bud set in 4-year-old apple plants
The shoot growth of 35S:PmDAM6 apple plants was inhibited over two consecutive seasons under both LP and SP conditions. Differences in node numbers and internode lengths resulted in significantly different plant heights between WT and 35S:PmDAM6. Thus, PmDAM6 may inhibit shoot elongation by affecting both meristem activity and cell elongation. In response to a naturally-occurring cold conditions during the 2013-2014 season, 4-year-old 35S:PmDAM6 plants exhibited an earlier bud set compared with WT plants (Fig 2). The earlier bud set in 4-year-old PmDAM6-overexpressing apple plants was confirmed by the bud-set timing of the DEX-treated GR lines (S3 Table). Thus, dormancy induction might be affected by the overexpression of PmDAM6 in apple. In peach, DAM genes are thought to be responsible for terminal bud set [14]. In fact, peach DAM6 is up-regulated when terminal bud set is observed in the field [44]. In the perennial herbaceous species leafy spurge (Euphorbia esula), the DAM homologs DAM1 and DAM2 are associated with dormancy induction [45]. Indeed, early bud set has been observed in 35S:PmDAM6 poplar [15]. In contrast, transgenic apple plants overexpressing MdDAMb do not exhibit early bud set [35], although the phenotypes of 35S:MdDAMb plants were observed only during the first 2 years. In the current study, the 35S:PmDAM6 apple plants did not exhibit earlier bud set during the first 2 years. Dormancy induction and establishment may be controlled by a complex genetic mechanism that is affected by environmental conditions [46]. Although this study demonstrated that PmDAM6 promotes dormancy entry in apple, future studies should aim to clarify whether PmDAM6 also influences dormancy establishment as discussed below.

PmDAM6 inhibited lateral bud break, but accelerated terminal bud break in 3-year-old apple plants
When 3-year-old 35S:PmDAM6 apple plants were grown under semi-field conditions, the lateral bud break was inhibited (Fig 3), while the terminal bud break was advanced relative to that of the WT plants. Moreover, the DEX-treatment of GR plants also accelerated the terminal bud break and inhibited lateral bud break (Table 1). This phenotype was observed only in 3-year-old plants consisting of a single 1-year-old shoot. In contrast, older plants bearing multiple 1-year-old shoots exhibited delayed terminal bud break ( Table 2). The mechanisms underlying the PmDAM6-mediated inhibition of the lateral bud break and acceleration of the terminal bud break in 3-year-old plants, remain unclear. One possibility is that PmDAM6induced growth inhibition may be antagonized with a strong sink activity at the terminal buds that take up water and nutrients. Consequently, only the lateral buds exhibit inhibited bud break in PmDAM6 overexpressing lines. Alternatively, PmDAM6 may directly inhibit only the lateral buds and not the terminal buds. In fact, SVP-like in hybrid poplar, SVL induces TCP18/ BRC1 which suppress lateral bud break [47]. Another possible explanation is that PmDAM6 may enhance apical part dominance effects during the bud break stage in 3-year-old plants.
Faust et al. [48] conducted decapitation experiments using apple dormant shoots. They concluded that apical dominance effects occur during the later phase of apple dormancy, and that a prolonged chilling exposure releases both apical dominance and endodormancy. Recently, Yao and Finlayson proved that ABA restricts the outgrowth of lower buds and promotes correlated inhibition [49]. These findings raise the possibility that ABA is one of the first upstream factors regulating the apical dominance responses to the red:far red ratio [50]. As discussed below, the overexpression of PmDAM6 increased the days to bud break in forcing condition, implying that PmDAM6 enhanced the dormancy depth of terminal buds during the adult phase. Additionally, PmDAM6 increased the ABA contents in apple terminal buds. Accordingly, we can hypothesize that PmDAM6 inhibits bud break in 3-year-old plants only in the lower lateral buds, possibly by increasing ABA contents. This hypothesis will need to be tested in future studies. Nevertheless, it remains unclear why this phenotype was observed only in 3-year-old plants and not in older plants.

PmDAM6 represses bud break competency and delays bud outgrowth via the accumulation of ABA and decrease in cytokinins contents in terminal buds of 4-to 6-year-old apple plants
The overexpression of PmDAM6 delayed lateral and terminal bud break in apples after 4 years under semi-natural conditions ( Table 2, Fig 4). Additionally, the dormancy depth was increased owing to the overexpression of PmDAM6 in apple ( Table 2). This is consistent with the results of a previous study on transgenic kiwifruit overexpressing SVP2, in which it prevented premature bud break [51]. If PmDAM6 is an important regulator for dormancy release, then it may induce major physiological changes in dormant buds. Thus, we investigated the phytohormone contents of the terminal buds from transgenic plants. Among the phytohormones, ABA is important for winter bud dormancy in woody perennials and the axillary bud dormancy of herbaceous species [1,49,50,52,53]. Our data suggest that PmDAM6 enhanced dormancy and delayed bud break, while simultaneously inducing the accumulation of ABA in the terminal buds of transgenic apple plants (Fig 5). Moreover, the ABA concentration was higher during the endodormant period and decreased during endodormancy release in Japanese apricot buds (Fig 7). We previously reported that PmDAM6 expression is down-regulated during endodormancy release toward bud break [15,19]. These findings suggest the seasonal down-regulation of PmDAM6 expression coincides with a decrease in the ABA contents during endodormancy release and bud break (Fig 7) in Japanese apricot leaf buds. Therefore, we hypothesize that PmDAM6 contributes to enhance the dormancy of Japanese apricot buds by mediating an increase in ABA contents (Fig 8). Tuan et al. [54] reported that pear PpDAM1 up-regulates the expression of the gene encoding 9-cis-epoxycarotenoid dioxygenase (NCED) by binding to the CArG motif in the promoter. Additionally, high levels of hybrid poplar SVPlike, SVL expression induces NCED3 [47]. Thus, PmDAM6 may contribute to the accumulation of ABA in terminal buds by promoting ABA biosynthesis. Although Wu et al. [51] did not find major differences in the ABA concentrations of 35S:SVP2 and control kiwifruit plants, their RNA-Sequencing analysis suggests that ABA and dehydration response pathways are modulated by the overexpression of SVP2. Thus, some SVP/DAMs genes may have conserved functions regarding the inhibition of bud break in poplar and several temperate fruit tree species during dormancy release and the bud break stage. These functions may involve modulating ABA accumulation and/or signaling. During the bud break stage, the overexpression of PmDAM6 increased ABA and decreased cytokinins levels in the terminal buds of apple plants (Fig 5). To the best of our knowledge, this is the first report of a possible interaction between DAM genes and cytokinins in woody  [15,19] and seasonal changes in phytohormone contents (this study) in Japanese apricot leaf buds. The expression of PmDAM6 is up-regulated during endodormancy and down-regulated during the endo-and ecodormancy release stages. Additionally, PmDAM6 expression is undetectable during bud break [15,19]. Down-regulated PmDAM6 expression coincides with a decrease in ABA during the endo-and ecodormancy release stages and with an increase in CK levels during the ecodormancy release and bud break stages. We hypothesize that PmDAM6 maintains endodormancy by repressing bud break competency and increasing ABA levels and inhibits dormancy release by increasing ABA and decreasing CKs contents in dormant Japanese apricot buds.
https://doi.org/10.1371/journal.pone.0214788.g008 plants. Cytokinins generally facilitate cell division and elongation, as well as bud outgrowth [55]. In fact, auxin and cytokinins levels increase in buds during the late stages of ecodormancy release toward bud break in Japanese apricot, suggesting that they influence bud outgrowth rather than dormancy. Additionally, in Rosa hybrida, ABA and cytokinins may function antagonistically during the regulation of axillary bud break in response to light intensity [56]. We observed that ABA contents continue to decrease not only from endodormancy to ecodormancy but also from ecodormancy to bud break in Japanese apricot leaf buds (Fig 7). Recently, Kitamura et al. [8] suggested that PmDAM6 transcripts levels at ecodormant stage are associated with the days to bud break in forcing condition in one F 1 population of Japanese apricot. Therefore, PmDAM6 may affect bud break competency during the endodormancy release stage as well as the bud break stage. Our results presented herein suggest that PmDAM6 may retard bud outgrowth by mediating increasing ABA and decreasing cytokinins contents in buds during the bud break stage (Fig 8).

PmDAM6 may be a growth inhibitor and act as a lateral bud break repressor in Japanese apricot buds
Our study demonstrated that PmDAM6 inhibits plant regeneration and growth, and promotes bud set, as well as represses bud break competency in transgenic apple plants. Additionally, PmDAM6 inhibits lateral bud break, as well as terminal bud break with the exception of those in young transgenic apple plants. Moreover, a working hypothesis regarding how PmDAM6 affects dormancy release and bud break in Japanese apricot has been developed based on the results of this study (Fig 8). Specifically, PmDAM6 might inhibit dormancy release and maintain dormancy by inducing the accumulation of ABA in dormant buds. Additionally, it may delay bud outgrowth by affecting the balance of endogenous ABA and cytokinins, which function antagonistically in the regulation of bud break and bud outgrowth. Our data provide a foundation for future analyses of PmDAM6 functions affecting the complex cross-talk between ABA and cytokinins that regulates dormancy and bud break. The putative role of PmDAM6 in repressing lateral bud break competency during dormancy revealed in this study appears to be consistent with the finding that PmDAM6 expression is up-regulated and down-regulated during "endodormancy" induction and release, respectively [15]. According to Sasaki et al. [15], "endodormancy" induction and release were estimated based on the decrease and increase in bud break frequency and/or the decrease and increase in the number of days to bud break under forcing conditions, respectively. Because all PmDAMs were expressed when bud break competency is repressed in lateral leaf buds [15], we hypothesize that all PmDAMs may share the conserved bud break repressor function as PmDAM6 does. However, SVP clade where PmDAMs belong to is supposed to be under positive selection [57], exact role of other PmDAMs should be elucidated by future genetic studies. Although the present study suggested the possible involvement of PmDAM6 in dormancy entry, it still remains unclear whether PmDAM6 also affects dormancy establishment. Considine and Considine [5] recently proposed changes to dormancy terminology, and suggested that bud dormancy can be further divided into either dormant (D) or quiescent (Q) states depending on the chromatin state in dormant meristem [i.e., heterochromatin (D state) or euchromatin (Q state)]. To verify the possible involvement of PmDAM6 for dormancy establishment, we may have to first determine whether D state cells exist in Prunus (and other fruit tree species in Rosaceae) dormant buds or not, then if any, the critical time points when the D state begins and ends in the leaf bud meristem.
In poplar, ABA mediates short-photoperiod induction of SVP-like, SVL expression during bud dormancy establishment [58]. In Rosaceae, dormancy of Prunus species is induced by the interaction of short photoperiod and low temperature [59]. In contrast, apple is typically not sensitive to short photoperiod for dormancy induction [60]. However, Wisniewski et al. [61] reported that peach CBF-overexpressing apple is sensitive to short photoperiod and induced bud set, which suggests that short photoperiod-induced dormancy pathway may be, at least in part, conserved in rosaceous fruit tree species. Additionally, our previous investigation indicated there is a close relationship between PmDAM6 transcripts levels and ABA contents in flower buds of two Japanese apricot cultivars with contrasting chilling requirement for dormancy release [19,62]. These results collectively suggested that Japanese apricot dormancy may be at least partially induced by short photoperiod and ABA-dependent, which appeared to be similar to the SVL-mediated poplar dormancy. If the PmDAM6-mediated dormancy release pathway is somehow similar to SVL-mediated dormancy regulation in poplar [47,58,63], ABA biosynthesis and/or signaling may be a putative upstream key factor for PmDAM6mediated regulation of dormancy and bud break in Japanese apricot.