Elucidation of molecular and hormonal background of early growth cessation and endodormancy induction in two contrasting Populus hybrid cultivars

Background Over the life cycle of perennial trees, the dormant state enables the avoidance of abiotic stress conditions. The growth cycle can be partitioned into induction, maintenance and release and is controlled by complex interactions between many endogenous and environmental factors. While phytohormones have long been linked with dormancy, there is increasing evidence of regulation by DAM and CBF genes. To reveal whether the expression kinetics of CBFs and their target PtDAM1 is related to growth cessation and endodormancy induction in Populus, two hybrid poplar cultivars were studied which had known differential responses to dormancy inducing conditions. Results Growth cessation, dormancy status and expression of six PtCBFs and PtDAM1 were analyzed. The ‘Okanese’ hybrid cultivar ceased growth rapidly, was able to reach endodormancy, and exhibited a significant increase of several PtCBF transcripts in the buds on the 10th day. The ‘Walker’ cultivar had delayed growth cessation, was unable to enter endodormancy, and showed much lower CBF expression in buds. Expression of PtDAM1 peaked on the 10th day only in the buds of ‘Okanese’. In addition, PtDAM1 was not expressed in the leaves of either cultivar while leaf CBFs expression pattern was several fold higher in ‘Walker’, peaking at day 1. Leaf phytohormones in both cultivars followed similar profiles during growth cessation but differentiated based on cytokinins which were largely reduced, while the Ox-IAA and iP7G increased in ‘Okanese’ compared to ‘Walker’. Surprisingly, ABA concentration was reduced in leaves of both cultivars. However, the metabolic deactivation product of ABA, phaseic acid, exhibited an early peak on the first day in ‘Okanese’. Conclusions Our results indicate that PtCBFs and PtDAM1 have differential kinetics and spatial localization which may be related to early growth cessation and endodormancy induction under the regime of low night temperature and short photoperiod in poplar. Unlike buds, PtCBFs and PtDAM1 expression levels in leaves were not associated with early growth cessation and dormancy induction under these conditions. Our study provides new evidence that the degradation of auxin and cytokinins in leaves may be an important regulatory point in a CBF-DAM induced endodormancy. Further investigation of other PtDAMs in bud tissue and a study of both growth-inhibiting and the degradation of growth-promoting phytohormones is warranted. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-02828-7.

growth cessation and dormancy development in vegetative buds of contrasting Populus cultivars differentially sensitive to low night temperature.

Background
The synchrony of the plant with its environment enables adapted temperate perennial plants to avoid injury. In these northern, temperate regions, growth cessation is a necessary pre-requisite to cold acclimation and subsequent freezing stress resistance [1,2]. The growth cycle is regulated by dormancy and in turn, dormancy is governed by both inherent but also environmental factors. While shortening photoperiod has long been known as the most important driver to woody plant dormancy induction [3][4][5][6], the temperature has also been and is increasingly recognized as a strong mediator of this response [7,8] for a review see Tanino et al. (2010) [9]. With global warming, more attention is being paid to temperature and its impact on the dormancy cycle. In this regard, research on forest and agroforest tree species have increasingly highlighted the impact of temperature on dormancy [10][11][12][13] and photosynthetic capacity [14]. In North America, with its wide adaptation and fast growth, Populus hybrids are the major agroforestry tree of choice in managed lands. Evaluating the impact of future climate change on Populus dormancy cycle is important to select cultivars which are better adapted to uctuating temperatures.
Dormancy in temperate trees is divided into three phases: paradormancy, endodormancy and ecodormancy [15]. Bud dormancy is de ned as 'the temporary suspension of visible growth of any plant structure containing a meristem' [15]. Paradormancy is de ned as growth cessation controlled by physiological factors within the plant but external to the affected structure, endodormancy is de ned as growth cessation controlled by physiological factors internal to the affected structure, and ecodormancy represents growth cessation controlled by environmental factors external to the plant [16]. Thus, the various types of dormancy in plants constitute a vast eld of study. However, because of the impact of the autumn dormancy induction period on other components of the annual growth cycle [17], and the demonstration that temperature mediates timing and depth of dormancy in Populus hybrids [13], in this paper, we will focus on these two aspects.
CBF genes, rst described in Arabidopsis [26][27][28], are among the best-characterized plant transcription factors involved in plant abiotic stress tolerance, especially in cold acclimation. The expression of the dehydration-responsive element binding (DREB) protein/C-repeat binding factor gene (CBF) is rapidly induced by low temperature. The encoded proteins bind to the CRT/DRE (C-repeat/dehydration responsive element) regulatory DNA motif in the promoters of cold-responsive genes [29], thus inducing their expression, which results in an enhanced cold or frost tolerance [30]. CBFs have been described in a huge number of species, both mono-and dicots. Usually several or many gene family member have been identi ed in one species. Moreover, the number of CBF genes may vary even in the same species, in a genotype-dependent manner (copy number variation). Many CBFs were described in the monocotyledonous cereals (some 40 in bread wheat /Triticum aestivum/, 20 in barley /Hordeum vulgare/). Although fewer were described in woody species, their genome also encodes several (3-6) CBF genes. Their involvement in cold adaptation has also been con rmed. A recent review [24] summarizes the genetic regulation of cold hardiness in trees.
It is becoming more evident that CBF genes are also involved in dormancy regulation, especially in the development of endodormancy [31][32][33][34]. Benedict et al. (2006) studied the kinetics and tissue speci city of 4 CBFs identi ed from Populus balsamifera subsp. trichocarpa and concluded that CBFs are involved in dormancy development and that their differential expression ensures speci c roles for these 'masterswitches' in the different annual and perennial tissues [35].
The existence of CBF -DAM -dormancy 'pathway' has been suggested and, at least partially, shown by several studies in Japanese pear [36,37] and Japanese apricot [38]. In his review, Horvath (2009) proposed a theoretical model, 'which can be developed that could serve as a paradigm for further testing' [39]. Wisniewski et al. (2011) demonstrated that transgenic apple (Malus x domestica) plants, expressing a peach (Prunus persica) PpCBF1 gene showed not only an increased level of freezing tolerance, but also a modi ed response to short photoperiod, leading to the early onset of dormancy, early leaf senescence, and delayed bud break [31]. As a next step Wisniewski et al. (2015) analyzed CBFs, DAMs, RGLs, and EBB transcription factor genes, involved in the regulation of dormancy [32]. The expression of several apple DAM genes -already associated with dormancy development in woody Rosaceae plants -exhibited different expression patterns. CBF binding sites identi ed in the apple DAM promoters led to the suggestion of a regulatory model connecting CBFs and DAMs expression to endodormancy development [32]. DAM genes were rst identi ed in Prunus. A mutant peach, called 'evergrowing', was unable to enter endodormancy even when plants were exposed to short photoperiods or low temperatures [40][41][42].
DAM genes are members of the type II (MIKCc) subfamily of MADS-box transcription factors. Their sequences contain four major domains, the MADS-box (M), intervening (I-), keratin-like (K-), and Cterminal (C-) domain. These domains are responsible for DNA binding, protein dimerization, complex formations, and transcriptional regulation. A detailed structural and functional characterization of DAM genes can be found in the reviews published by  and Falavigna et al. (2019) [43,44]. In this latter publication, a model is proposed, introducing the molecular network of the regulatory genes involved in the dormancy cycle.
Expression patterns of the DAM genes were related to endodormancy and were mainly presented in the Prunus genus, among them peach (P. persica) ever-growing [45] and peach cultivars [46], Japanese apricot (P. mume) [47], and also in apple [48] or Japanese pear (Pyrus communis) [49,50]. DAM gene expression appears to be linked to the stage of dormancy (see Falavigna et al. (2019) for a review [44]). In most species, DAM gene expression is induced during the dormancy induction period but may also be involved in maintenance and release [46]. Based on amino acid sequence, poplar DAM1 and DAM2 expression were most closely associated with leafy spurge MADS 27-29 and unlike the other DAM genes, DAM1 and DAM2 were upregulated by dormancy inducing short-day conditions in poplar (Chen (2008) [51]as cited by Horvath et al. (2010) [52]). Interestingly in a later transcriptome study these results were not con rmed. Howe et al (2015) examined several DAM-like genes that were downregulated during endodormancy [53]. One of the examined genes was the Potri.002G105600, but in this study, the authors did sampling only once per month.
Analysis of transgenic plants showed CBF genes were also involved in the regulation of endodormancy. The ectopic expression of a peach PpCBF1 gene in apple resulted in short-day induced dormancy and increased cold hardiness [31], and affected the expression levels of apple MdDAM1 and MdDAM3 genes in buds [32]. Li et al. (2019) analyzed pear (Pyrus pyrifolia) CBF and DAM genes and found multiple CBF genes selectively regulate DAM genes and participate in endodormancy regulation [37]. Interestingly, this group found that "PpCBF1-PpDAM2 regulon mainly responds to low temperature during endodormancy regulation, with further post-translational regulation by PpICE3". In addition, the expression of ParCBF1 was found to be in close association with the decreasing ambient temperatures in apricot (Prunus armeniaca), and the expression levels of ParDAM5 and ParDAM6 changed according to ParCBF1 expression rates [54].
Molecular evidence also supports the CBF -DAM connection. The presence of CBF transcription binding sites was reported in the putative promoter regions of the leafy spurge DAM genes [52]. A model, illustrating a potential interaction between DREBs (CBFs) and DAMs was subsequently suggested [55]. An interaction of PpCBF2 protein with the promoter of PpMADS13-1 gene was shown in pear by transient reporter assay [56]. Also in pear, yeast one-hybrid and transient assays showed that PpCBF2 enhanced PpDAM1 and PpDAM3 transcriptional activity during the induction of dormancy [38].  showed that P. mume CBFs can bind to the PmDAM6 promoter via alternative binding sites and activate its expression [33,34]. Japanese pear PpCBFs were able to induce the expression of PpDAM1-1 and PpMADS13-3 genes in transient reporter assays [56]. Different biochemical methods revealed that pear PpCBF2 and PpCBF4 genes are able to bind to the promoter of PpDAM1 gene, activating its expression, and revealed that PpCBF1, PpCBF2, PpCBF3, PpCBF4 genes can activate PpDAM3 gene [37]. These results demonstrate the CBF -DAM signalling pathway is involved in endodormancy development and also demonstrate a certain level of CBF functional redundancy.
Herein we use a system of two contrasting Populus hybrid cultivars differing in growth cessation and dormancy acquisition which were previously distinguished based on low night temperature under short photoperiod. The hypothesis that growth cessation and dormancy induction is linked to leaf phytohormone levels, bud PtDAM1 and bud PtCBFs gene expression in Populus will be evaluated.

Results And Discussion
In our previous work [13] we studied the impact of temperature on growth cessation, dormancy development, and cold acclimation of four poplar cultivars. These temperature regimes changed the kinetics of dormancy development patterns with the 18/3 °C treatment inducing the widest separation of dormancy depth. Therefore, to elucidate if there is a relationship between the expression levels of CBF genes and dormancy in poplar, two cultivars differing in their dormancy acquirement based on night temperature responses were tested under short-day conditions in our current research.

Dormancy development Growth cessation
Growth cessation is the rst indication of dormancy induction [57] and was induced in both genotypes (

Dormancy Induction
While growth cessation was a more sensitive indicator, signi cant differences in the number of days to bud break were found between the two cultivars from day 40 (Fig. 2). Okanese buds took 10 days to break bud at Day 0, while at the end of the experiment (i.e. on the 50 th and 60 th day) this value was increased and levelled off at 13.5 and 14.1 days (respectively). Conversely, the duration of bud break was hardly changed in 'Walker' over the whole 60 days treatment period, just a slight uctuation was recorded ( Fig. 2). No difference was detected between the rst and the last days (days to bud break: 8.4 and 8.2 days, respectively).
The depth of dormancy was re ected by the parameter ΔDBB (Differences between the rst and last Days to Bud Break) and dormancy was not induced at all in 'Walker' (ΔDBB=0.2). Conversely in 'Okanese', the dormant state started to be induced after 30 days. At the end of the experiment, Okanese had a ΔDBB of 4.1 (Fig. 2). The data on growth cessation rate and the bud break analysis indicates that 'Okanese' reached a deeper dormant state than 'Walker'. Our results are consistent with the outcome of Kalcsits et al. (2009) [13], who reported characteristic differences between these two cultivars -'Okanese' was shown to be more capable of endodormancy development under the 18/3 o C day/night temperature treatment under 12 h and 10 h daylengths although a larger difference was found between the two cultivars (ΔDBB: 13.9) in that study.

Expression pattern of CBF genes
The expression patterns of six CBF genes were recorded over the whole experiment. Samples were collected from leaf and bud tissues every ten days, taking into account the circadian rhythms of many CBFs, in the same period of the day, i.e. 4 -6 hours after the start of the light period. The expression of each gene in a given time-point was normalized to the level measured at the beginning (i.e. on the 0 day) of the given treatment.
Differences in the kinetics and spatial localization of the overall CBF expression were observed between the two cultivars. The highest levels of CBF expression across the entire experiment were recorded in the bud tissue, isolated from 'Okanese' on the 10 th day (Fig. 3A) and on the rst day in 'Walker' leaf samples ( Fig. 3D). The expression levels in 'Okanese' poplar buds peaked at the 10 th day and were at least an order of magnitude (10-20 fold) higher than in 'Walker' buds, and at any other time during the experiment for all the CBFs (with the exception of PtCBF5). There was differential expression of bud PtCBFs between the two cultivars in that PtCBF1 and PtCBF5 showed the highest and lowest expression in 'Okanese' buds, respectively, while the reverse was observed in 'Walker'. In 'Walker' leaves on the 1 st day, PtCBF1 and PtCBF2 expression levels were roughly equivalent to 'Okanese', however, 'Walker' leaf expression of PtCBF3, PtCBF4, PtCBF5 and PtCBF6 spiked on the 1 st day and were 150 -200 times greater than 'Okanese' (Fig. 3C, D).
The expression patterns of the unique CBF genes are described in detail in the Supplemented Fig 1. In bud tissue, PtCBF2, PtCBF3 and PtCBF5 were induced only in the beginning of the experiment, on the 10 th day, while PtCBF1 and PtCBF6 were induced not only at the beginning but also at the end of the treatment, on the 50 th and 60 th (PtCBF1) or on the 40 th day (PtCBF6). The induction level was always an order of magnitude higher in the 'Okanese' buds compared to 'Walker' for each CBF. The repression of CBF genes was more pronounced in the 'Walker' buds. A repressed period was recorded in the middle of the experiment for PtCBF1, PtCBF5 and PtCBF6 genes in 'Walker' buds, while only one repressed stage was found in 'Okanese' in the mid period of PtCBF3 expression (Supp. Fig. 1A, D and E).
By contrast in leaf tissue, two induction waves could be observed in the leaf samples for all CBFs in 'Walker': the rst was at the beginning (on the 1 st and 20 th day), while the second was at the end (50 th and 60 th day). Induction waves were also found in 'Okanese' leaves, but in the opposite direction, since repression of all CBFs was detected in the period 1 st -10 th and 30 th -40 th and nally on the 60 th day. It is interesting to note the differential responses between the cultivars in leaf tissues in that CBF induction was found in 'Walker' leaves, while repression was found almost in every case in 'Okanese' leaves (Supp. Fig. 1F-K). Thus, these two cultivars had similar but opposite PtCBF expression under dormancy inducing conditions based on buds or leaves.
Differences in the CBF expression kinetics and levels measured in the meristematic (bud, stem) and leaf tissues were studied in several cases in woody plants, among them poplar. Benedict  PtCBFs are cold-inducible in leaves, while only two (PtCBF1 and PtCBF3) were cold induced in the stem [35]. Under a short-expression period (24 hours), they concluded, 'the perennial driven evolution of winter dormancy led to the development of speci c roles for abiotic stress response regulators, such as the CBFs, in annual and perennial tissues'. CBF expression was followed in leaf and leaf bud tissues in Prunus mume during one year by   [33]. They also found a differential gene expression pattern for all six CBFs studied, with speci c induction kinetics. In that study, all six CBFs were induced in vegetative buds, in the cold period (November -January); PmCBF4, PmCBF5 and PmCBF6 being the most intensively expressed. These three CBFs were also the most induced in the leaf tissues. But interestingly, in leaves, the highest expression for all 6 CBFs was recorded during the warmest period, from June to July. This nding is in accordance with our results, i.e. that the CBF expression was much more intense in leaves of the non-dormant cultivar, may indicate that their role in the development in dormancy is organ-speci c. Six PmCBFs in 7 different organs were determined in P. mume [33]. The induction levels were high in stems, moderate in ower buds, leaf buds, and leaves, poor in owers, fruits, and seeds.
Gene duplication and multiplication produced a large number of CBFs in many species. This redundancy makes possible the divergence of functionality, and the possibility for ne-tuning of adequate response for any environmental stimuli, such as stress. As mentioned above, 6 CBFs encoded in the P. mume genome exhibited different expression kinetics during the year: PmCBF1, PmCBF2, and PmCBF3 were upregulated in the stem tissues not only in the cold period but also in late spring [34]. Additionally, low temperature up-regulated 8 CBFs in Prunus mume which subsequently induced all six DAM genes resulting in dormancy development [36]. Under natural dormancy induction conditions, 3 out of 4 CBFs showed similar expression trends in Pyrus pyrifolia bud tissues, while PpCBF1 showed a different induction kinetic [37]. During an arti cial chilling test, PpCBF1 was the only CBF highly expressed, while PpCBF2 was repressed intensively, and the levels of PpCBF3 and PpCBF4 were undetectable.
These results show that although CBF expression kinetics may be similar, differences in the individual expression patterns can be distinguished. Shortening the light period by 2 hours/day to account for the variance in nature (at the same temperature regime) may have caused a moderate functional polymorphism in our experiment. PtCBF4 was detectable only in 'Walker' leaves, while PtCBF1 and PtCBF6 were the most intensively expressed genes in 'Okanese' buds. Whether they have different functions, as was suggested for PpCBF4 [37] in pear, is still unclear. It is also remarkable that PtCBF5 was the only gene which was not induced during the CBF-burst on the 10 th day in 'Okanese' buds but was the most intensively up-regulated in 'Walker' buds on the 1 st day. Therefore, we assume PtCBF5 is not related to dormancy development.
Leaf samples of Populus balsamifera ssp. trichocarpa genotypes originating from northern and southern populations were examined [58]. A growth chamber study showed all PtCBF genes were induced by cold, indicating functional redundancy. On the other hand, under eld conditions, a more diverse gene expression pattern was described. The expression of PtCBFs increased as the growing season progressed, but among the six genes, only PtCBF3 was marginally differentially expressed across latitudes. In our experiment, leaf samples also showed a certain level of functional polymorphism, but the most common outcome of the two systems is that in leaves, no dormancy dependent expression pattern was found, such a relation was present only in the bud tissues.

PtDAM1 identi cation and its expression kinetics
DAMs (Dormancy-Associated MADS-Box) are well-characterized genes in perennial plants, associated with various components of the dormancy cycle but particularly dormancy induction. DAM sequences had already been published in woody plants, all containing K-box and SRF-TF motifs [33,59,60]. The P. trichocarpa genome has been sequenced [61], however, it is still poorly annotated. We have found 151 candidates for the DAM genes. From these, we suggested the XP_024452024.1 protein entry as a putative PtDAM1 product (Fig. 4, Fig. 5). Howe et al. (2015) studied transcriptome changes during endodormancy induction by microarray in P. trichocarpa and found several DAM-like SVP genes were differentially expressed but were downregulated during endodormancy [53]. Since sampling was conducted on a once per month basis, it is not clear if upregulated peaks were missed.
Having identi ed a PtDAM1 gene in Populus, we decided to evaluate its potential role in dormancy development, using cultivars known to be differentially responsive to night temperatures. Therefore, primers were developed to study the encoding PtDAM1 gene expression. Compared to the rst sampling day, mild up-regulation of the identi ed putative XP_024452024.1 (is corresponds to older versions as MADS7, Potri.002G105600) sequence was recorded in 'Okanese' leaves through the experiment, while lower induction was found in 'Walker'. PtDAM1 was repressed from the middle of the experiment (Fig. 6) and the expression of PtDAM1 was almost unchanged throughout the 60 days in leaf tissues. The bud tissues showed much more pronounced induction than the leaf tissues. In more dormant 'Okanese', the maximum expression (2.8-fold) was recorded on the 10 th day then the induction gradually declined. Repression was recorded in both cultivars at the end of the treatment. PtDAM1 induction in buds was weaker in the rst half of the treatment in 'Walker' (1.1 -1.6-fold induction) which did not enter endodormancy.
Similar expression trend for PmDAM1 gene was described in Japanese apricot (Prunus mume) bud tissue, but differently in the leaf samples [47]. In the vegetative buds, expression of PmDAM1 (as well as PmDAM2 and PmDAM3) was upregulated from June to July, i.e. long before the start of growth cessation, then expression started to decrease. We also showed an initial PtDAM1 induction in our system, well before the start of growth cessation, or dormancy development. We found no characteristic changes in leaf tissues, however, in Prunus mume, different kinetic patterns were described in this organ [47]. Two seasonal expression trends were shown for P. mume DAMs, PmDAM1 (together with PmDAM2 and PmDAM3) was rapidly up-regulated in spring, being gradually down-regulated in autumn. This difference in the expression in leaf tissue might be explained by the two different experimental systems. In other studies in peach (Prunus persica), differential DAM gene expression appeared to be related to dormancy induction or ful llment of the chilling requirement phases. Based on the ever-growing peach mutant system, Li et al. (2009) reported DAM1, DAM2 and DAM4 were the most likely candidates associated with growth cessation and dormancy induction [45]. Using the same system, Yamane et al.
(2011) showed under both eld and controlled environment conditions and in leaves and stems, DAM5 and DAM6 gene expression levels were up-regulated during endodormancy induction and downregulated during endodormancy release which appeared to be tied to chilling requirement satisfaction [62]. Furthermore, DAM5 and DAM6 gene expression levels were higher in high chill cultivars and reduced with chilling requirement satisfaction [63]. DAM5 and DAM6 genes were negative regulators of bud break.
In our study, due to the very small size of poplar buds and only limited capacity of growth chambers, hormone analysis was conducted only in leaf samples. Overall, phytohormonal response in 'Okanese' was different than in the 'Walker' poplar hybrid cultivar with most signi cant distinction for Ox-IAA, phaseic Acid, DAM1, cis-zeatin riboside-O-glucoside (cZROG) (Fig. 7, 8). Exposure of poplar plants to short photoperiod and low night temperatures was associated with down-regulation of ABA content in leaves of both genotypes (Fig. 8). However, an early (on the 1 st day) transient elevation of the ABA metabolite, phaseic acid, indicated enhanced ABA degradation in the 'Okanese' cultivar, suggesting a preceding short-term up-regulation of ABA content early after temperature drop. This assumption is supported by the report on transient up-regulation of ABA in cold-stressed wheat leaves [75]. The ethylene precursor ACC was elevated in both clones. Ruttink et al. (2007) showed ethylene rise preceded ABA during dormancy induction [64]. Jasmonate has been known to be involved in several stress responses [76]. Inactivation of the repressors of JA signaling pathway -jasmonate ZIM-domain (JAZ) proteins, which physically interact with ICE1 and ICE2 transcription factors, results in up-regulation of CBFs [77]. CBF genes promote gibberellin deactivation and thus growth inhibition [78]. In our study in leaf tissue, JA levels were suppressed in both genotypes during the entire experimental period, and more in 'Okanese'. However, JA level in leaves need not correlate with its content in buds. Moreover, JAZ inactivation may be achieved by their interaction with DELLA proteins [79,80], which accumulate at low temperature and are stabilized by gibberellin down-regulation. In contrast to JA, SA levels were increased at the beginning of the experiment, one week longer in 'Okanese'. This agrees with the positive effect of SA on plant cold tolerance [81]. After the 3 rd week, the SA content was unchanged in both cultivars, however, the concentration was lower in the less cold-hardy 'Walker'. Benzoic acid, the precursor of SA and other phenolic compounds, was elevated during the experiment; in 'Okanese' until dormancy initiation, in 'Walker' during the whole experiment. These changes demonstrate differences in hormonal dynamics between the clones during leaf senescence (Fig. 8).
The auxin, indole-3-acetic acid (IAA), had varying levels across the 60-day treatment in both cultivars. However, the main IAA catabolite, Ox-IAA, had a more consistent response, being up-regulated in 'Okanese' and down-regulated in 'Walker', which indicates stronger IAA deactivation in 'Okanese' leaves. Dormancy initiation, associated with substantial suppression of growth rate, was accompanied by IAA downregulation, which was not observed in the non-dormant clone. Baldwin et al. (2000) showed that while the auxin naphthaleneacetic acid was not required for bud scale development, its absence was critical [82].
The whole cytokinin pathway was downregulated in Okanese compared to Walker: the precursors iPRMP and tZR increased only in Walker, the active form (iP) decreased only in Okanese, and the deactivated form iP7G was accumulating in Okanese and decreasing in Walker. Other compounds did not show any major changes between both trees.
Cytokinin analysis clearly showed that promotion of dormancy in 'Okanese' was associated with a general decrease of cytokinin biosynthesis and profound elevation of their deactivation products in leaves (Fig. 7, Fig. 9, Supple. Fig. 2). Collectively, these results provide new evidence that the degradation of growth-promoting phytohormones such as IAA and cytokinins may be an important mechanism of endodormancy induction.

The relation between PtCBFs and PtDAM1 expression, hormone level and the development of dormancy
A CBF-burst occurred on the 10 th day of the short photoperiod and low night temperature treatment in 'Okanese' bud tissues, while in 'Walker' CBF levels were an order of magnitude lower (Fig. 3). In 'Okanese' which was able to enter endodormancy (Fig. 2), CBF1 had the highest relative expression at the initiation of dormancy. PtDAM1 expression peaked in 'Okanese' exactly on the same 10 th sampling day (Fig. 6). By contrast, 'Walker' which did not attain endodormancy (Fig. 2) had a lower CBF expression on the 1 st day (Fig. 3), while PtDAM1 expression was also low and unchanged during the experiment (Fig. 6). Growth rate started to decline in both cultivars by the 3 rd week, but at a much faster rate in 'Okanese' (Fig. 1). These ndings support the possible relationship between PtCBF1, PtDAM1 induction and endodormancy development.
The dormancy-associated phytohormone, ABA, was surprisingly down-regulated in leaves of 'Walker' and even more downregulated in 'Okanese'. However, the concentration of the ABA degradation intermediate, phaseic acid, increased in 'Okanese' while it was reduced in 'Walker' and therefore, an ABA induction peak in 'Okanese' leaves may have been missed (Fig. 8). Recent evidence indicates a role of DAM1 in activating NCED3 through binding to its promoter and upregulating ABA biosynthesis in Japanese pear [83]. AcSVP2 may mimic ABA action [85]. They further indicated that SVP2 was mediated by ABA to decrease meristem activity and prevent premature bud break. DAMs also appear to play a regulatory role in the ABA signaling pathway [85]. Thus, there is increasing evidence that CBF and DAM gene actions are linked with phytohormonal concentration and action in dormancy. The reverse has also been demonstrated in that Knight et al. (2004) earlier showed ABA to upregulate CBF expression [86]. Singh et al. (2019) reported that SVL is the ortholog of SVP in aspen (Populus tremula x tremuloides), which mediates photoperiodic dormancy induction via callose synthase, operating downstream of ABA [74]. Singh et al. (2018) also showed ABA induced the expression of the DAM/SVL gene in hybrid aspen [87]. For an excellent recent review, see Liu and Sherif (2019) [25].
In a recent study, analysis of a transformant hybrid aspen (Populus tremula x tremuloides) showed that expression of SVL, a negative regulator of bud break, was down-regulated in hybrid aspen buds after low temperature treatment. It was noted that nonetheless, SVL is similar to DAM genes, clustering closer to SVP in Arabidopsis and apple than to hybrid aspen or peach DAM genes [74,87]. Interestingly, SVL induced the expression of callose synthase and negatively regulated the gibberellin pathway. Moreover, CBF14 and 15 upregulated the GA2ox5 gene which deactivates gibberellins in barley [88].
Dormancy is known to be induced primarily by temperature in some fruit species, such as apple and pear [89]. Increasing evidence highlights the role of temperature, especially in the case of northern woody cultivars. While the main regulator of growth cessation and dormancy induction in woody species is short photoperiod, it may be moderated by, and interact with temperature [17]. The increasing con rmation of direct regulation by cold-induced CBFs on DAM gene expression [34,37,56], Niu et al. (2016) provided evidence and proposed a model in which CBF induces DAM and DAM downregulates FT which then suppresses growth and stimulates the development of dormancy [38]. Liu and Sherif (2019) further outlined a model integrating multiple phytohormonal networks regulated by DAM [25]. Key among them was the direct suppression by DAM of cytokinins, gibberellins and direct activation by DAM of ABA and callose deposition. Our study provides additional evidence that cytokinin and IAA degradation may be an important regulatory mechanism to endodormancy induction.

Conclusion
In this study, the differences between the early induction of growth cessation and the depth of endodormancy between two tested poplar cultivars under short photoperiod and low night temperature treatment are associated with the differential expression levels of CBF1 and PtDAM1 genes in buds as well as degradation of growth-promoting phytohormones auxin and cytokinins in leaves. However, since other DAM genes were not examined, we cannot rule out the possibility of other DAM gene involvement.

Dormancy assessment
Dormancy development was measured using the bud-break method adapted from Kalcsits et al. (2009) [13]. In brief, small cuttings with two buds were collected from two pots from each genotype. For each genotype and every sampling time-point, 20 branches were cut, so the budburst on 40 buds was examined at given time-point. Cuttings were put in water in glass tubes and kept under LD conditions (18-h daylength) at continuous 22°C. Samples were collected in every 10 th day over the 60 days long experimental induction period. Bud-break was de ned as the point when the rst leaves started to emerge from the dormant bud, a longer time to bud-break indicates a higher level (i.e. deeper) dormancy. The depth of dormancy (ΔDBB) was calculated according to Kalcsits et al. (2009) as the difference between the days to bud break between the last and the rst sampling days [13].

Growth cessation assessment
The length of the growing branches was measured from the base to the apex every week. Seven pots with 4 branches were measured per genotype. Growth rates (cm*week -1 ) were calculated, and when the growth rates (almost) reached zero, the plants were considered to have stopped their growth period. For these examinations, we use different plants than for gene expression and hormone analysis. These plants were not wounded during the whole experiment.Gene expression studies The youngest fully expanded leaf and mid branch bud samples (about 3 plants per every sampling point altogether 12 leaves and buds were collected) for gene expression studies were collected 4 -6 hours after the start of the photoperiod and frozen immediately in liquid nitrogen and kept at -80°C till RNA extraction. Samples were homogenized by TissueLyser II (Qiagen) equipment (29 Hz, 1:30 min), twice.  [58]. The normalized relative gene expression levels were calculated by the ΔΔCt method [90]. Ct values were normalized to the Ct values of the housekeeping Pt18S rRNA gene (Table S1). Expression level, measured at a given time point, was compared to the expression level measured on the rst day for each genotype. The raw ΔΔCt values are included in the Supplemented Table S2.
The relative expression values (fold change) were converted to log 2 values, clustered and visualized with the Gitools software on the Supplemented Figure 2 [91].

Identi cation of PtDAM1 gene
For sequence analysis, the Populus trichocarpa reference genome assembly was retrieved from the NCBI Assembly server (https://www.ncbi.nlm.nih.gov/assembly) at proteome level (GCF_000002775.4). Pfam and Hidden Markov Model (HMM) based protein domain search was performed using hmmscan packages of HMMER 3.0 software [92]. The protein collection from the poplar proteome was aligned using a MUSCLE alignment method (Fig. 5) and inferred using Maximum-likelihood phylogenetic tree by MEGA6 software package [93]. Based on the Bayesian Information Criterion (BIC) the best-t, Jones-Taylor-Thornton (JTT+G) substitution pattern was chosen for the phylogenetic reconstruction. One thousand bootstrap pseudo-replicates were used to test the reliability of the inferred tree.

Hormone analysis
The youngest fully expanded leaf samples (ca 50 mg FW) were puri ed and analyzed according to Dobrev and Kamıńek (2002) and Dobrev and Vankova (2012) [94,95] Availability of data and materials All the relevant data are included in the manuscript and the supplemented materials.