Classical physiological parameters such as water-, carbon-, nitrogen content as well as the fresh and dry weight in flower buds were useful to identify the annual duration and timing of the dormant phase and to define the beginning of ontogenetic development for ‘Summit’ sweet cherries [1]. However, these parameters did not allow identifying the date of endodormancy release - a very important parameter in phenological models. For this reason, several classical experiments with plant samplings were carried out to calculate the plant specific chilling requirement [1-6]. Recently, Chuine et al. [7] suggested carrying out these experiments for a wider range of tree species and varieties in order to capture this important parameter. These classical experiments can temporarily help to identify the chilling requirement or the date of endodormancy release for a wider range of tree species, but for an improved phenological modelling a deeper knowledge of dormancy induction, maintenance and release will be necessary.

Already Perry [8] announced in his review that different plant species and different genotypes within species vary in their dormancy phenomena and defined that a dormant plant having two attributes, (i) a period of markedly reduced growth rate with few, or in some cases no, cell divisions in the terminal/lateral meristems of the plant and (ii) a winter chilling requirement. The renewed growth after dormancy release is then related to the disappearance of inhibitors from the buds and increased concentrations of growth stimulating compounds. Generally, the transitions of the metabolism from an active growth phase to a phase of quiescence, and back to the active growth are gradual and also seasonally dependent. The reason for this is that the amount of cold or the chilling requirement varies with the weather conditions of the preceding season [8,9]. Changes in day length, light quality, and temperature provide the most important environmental cues for synchronising growth with seasonality. Xin and Browse [10] summarized in a review the most notable changes, which include changes or cessation of growth, reduction of tissue water content, transient increase in ABA levels, changes in membrane lipid composition, the accumulation of compatible osmolytes such as proline, betaine, soluble sugars and antioxidants. However, cold acclimation requires many changes in cell biology and metabolism, but the knowledge of the metabolic changes that contribute to these processes is still incomplete.

Autumnal bud development is a composite of bud formation, simultaneous acclimation to dehydration and cold, and acquisition of dormancy [9]. These three processes are interconnected with each other but happen at different levels: cells throughout the plant will acclimate to dehydration and cold for survival, whereas only the youngest derivatives of the meristem participate in bud formation, and yet only those meristem parts of the bud that will resume growth in spring, become dormant. Moreover, the simultaneous activity of various closely cellular, physiological, and morphological processes confounds the dissection of the underlying developmental programs and their respective signals. Only in a few cases processes could be separated and assigned to the action of a particular signalling route [9]. Buds in fruit trees are considered to be ‘utilizing sinks’ but not ‘storage sinks’ and thus may preferentially need metabolic substrates rather than storage compounds [11]. Metabolomic studies as yet do not only allow detecting the date of endodormancy release, but also help to better understand this complex physiological process. In this study four well-known metabolites are presented which can help to identify the timing of endodormancy release in sweet cherry buds of the cultivar ‘Summit’. Metabolites are organic compounds synthesized by many single enzyme-mediated chemical reactions. They are intermediates and products of plant metabolism for growth and development (primary metabolites), and e.g. for defense against herbivory (secondary metabolites). The plant hormone ABA has been shown to mediate the development of freezing tolerance, whereas several roles for sugars in protecting cells from freezing injury have been proposed Xin and Browse [10].

The aim of this study was to show whether temporal changes of these four metabolites during autumn and winter allow defining the date of endodormancy release of ‘Summit’ sweet cherries.

Experimental site

The study was conducted for 5 seasons (2011/12-2015/16) at the experimental sweet cherry orchard at Berlin-Dahlem (52.47 °N, 13.30 °E, h = 51 m a.s.l.), established in 2000. The orchard (980 m2) comprises 80 cherry trees (cultivars ‘Summit’, ‘Regina’ and ‘Karina’) growing in 8 rows with 10 trees each. Trees are grafted on Gisela-5 rootstocks and pruning was performed on demand. All investigations in this study are focused on the cultivar ‘Summit’, of origin in British Columbia. The soil type and long-term climatic conditions are given in [1].

Bud sampling

In order to analyse ABA and sugar content, weekly 3 flower-bud cluster per tree were taken, randomly over 4 trees (replications) during the cold season from beginning of October until April. After beginning of bud development sampling was conducted development-related at the stages ‘swollen bud’ (SB), ‘side green’ (SG), ‘green tip’ (GT), ‘tight cluster’ (TC), and ‘open cluster’ (OC). After freeze-drying and milling of the buds [1], the content of abscisic acid (ABA) and sugars (fructose, sucrose, glucose) were analysed during 5 seasons.

Analytic work

Extraction method: Freeze dried samples (50 mg) were extracted with 1.5 ml 60% methanol using an ultrasonification treatment (Sonorex RR 100, Bandelin electronic GmbH & Co. KG, Berlin, Germany) for 3 minutes followed by incubation at 4°C overnight (AEG Santo 60240 DT 28, Electrolux Hausgeräte GmbH, Nürnberg, Germany). The extracts were centrifuged at 9300 x g for 10 minutes at 4°C, the extraction repeated, and the supernatants pooled together and stored at -20°C until needed.

Identification of ABA: The extraction of the phytohormone was performed as described above. The determination was adapted using the high-performance liquid chromatography and tandem mass spectrometry [12-14]. Chromatography was carried out on a Dionex Ultimate 3000 HPLC system (Thermo Fisher GmbH, Idstein, Germany) using a Phenomenex Luna C18 column (Phenomenex Inc., Aschaffenburg, Germany; 150 × 3 mm; 3 μm) at 40°C , connected to a C18 pre-column containing the same material with a flow rate of 0.32 ml/minutes. The eluents were A (distilled water with 0.1% formic acid) and B (methanol with 0.1% formic acid). The gradient was applied under the following conditions: 50% eluent B, 0-1 minutes; 50-75% eluent B, 1-12 minutes; 75% eluent B, 12-13 minutes; 75-50% eluent B, 13-13.1 minutes; 50% eluent B, 13.1-17 minutes. Run time was 18 minutes and the injection volume was 10 μl. Concentrations of 0.01-10 μg ml−1 of selected phytohormones in a mixture (abscisic acid, gibberellic acid, 6-benzyl adenine, indole-3-acetic acid, benzoic acid, acid, indole-3- butyric acid, naphthalene-acetic acid, Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) were used for external calibration. Negative ion mass spectra of the samples were recorded in the MRM modus using a 4000 Q Trap^® mass spectrometer (AB Sciex Germany GmbH, Darmstadt, Germany). The ion source temperature was set at 600°C and a dwell time of 120 ms for negative ionization were used.

Identification of sugars: The detection of sugars was performed by analysing the samples with a Shimadzu HPLC system (Shimadzu Europa GmbH, Duisburg, Germany) equipped with an evaporative light-scattering detection (Shimadzu ELSD-LT II at 40°C, Gain=3). An X-bridge Amide (particle size 3.5 μm, 150 mm of length and 4.6 mm of internal diameter; Waters GmbH, Eschborn, Germany) column with 12% carbon load was used. The extracts were centrifuged at 16000 x g for 2 minutes and 50 μl of the supernatant were used for HPLC analysis. Acetonitrile (80%) and 0.1% ammonia in water (20%) was used as eluent and a gradient with changing flow rate under isocratic conditions was applied: 0-30 minutes, 1-1.5 ml/minutes; 30-32 minutes, 1.5-1 ml/minutes; 32-35 minutes, 1 ml/minutes. External calibration was made using solutions of fructose, glucose and sucrose at 0.2-2 mg ml-1 each (dissolved in 60% methanol).

Statistical analysis

For all sampling dates, mean values and standard deviations of the ABA- and sugar contents were calculated for 4 replications. Significance of temporal changes of the metabolites was tested with the Tukey-HSD test (p≤0.05). In order to quantify the relationship between mean minimum temperature during endodormancy and mean sucrose content during ecodormancy the Pearson correlation coefficient was calculated.

Timing of para-, endo-, ecodormancy and ontogenetic development

The annual timing of para-, endo-, ecodormancy and ontogenetic development for ‘Summit’ in Figures 1-3 was indicated by different colours in the figure’s background, derived from Chmielewski and Götz [1]. The beginning of endodormancy was equated with the date of complete leaf fall (LF, BBCH 97). Its end (t₁) was derived from classical experimental studies with cherry twigs, which were cut off weekly from November until December and placed in 500 ml plastic flasks, filled with water, to observe the beginning of blossom in a climate chamber with 12 h light, temperatures of ∼20/∼15°C (day/night) and 70% relative humidity. When twigs started to bloom under controlled conditions (BBCH 60), it was the indication that the chilling requirement at sampling in the orchard was fulfilled. Thus, it was possible to time t₁. Beginning of ontogenetic development (t₁*) was defined as the date with an ongoing increase of bud’s water content, which was related to continuously rising air temperatures above the freezing point [1].

Figure 1: Weekly and development orientated changes of ABA content, 2011/12 - 2015/16. Black squares indicate timing of t₁ according to ABA content. LF: leaf fall, SB: swollen bud, OC: open cluster, t₁: end of endodormancy and t₁*: beginning of ontogenetic development, according to the classic approach, DOY: day of year.

Figure 2: Weekly and development orientated changes of sugar content (glucose, sucrose, fructose, total sugar), 2011/12 - 2015/16. LF: leaf fall, SB: swollen bud, OC: open cluster, t₁: end of endodormancy and t₁*: beginning of ontogenetic development according to the classic approach, DOY: day of year.

Figure 3: Weekly and development orientated changes of sucrose content, 2011/12 - 2015/16. Black squares indicate timing of t₁ according to SUC content. LF: leaf fall, SB: swollen bud, OC: open cluster, t₁: end of endodormancy and t₁*: beginning of ontogenetic development according to the classic approach, DOY: day of year.

Weather conditions in the seasons 2011/12 - 2015/16

With leaf fall (LF) of cherry trees, in the end of October or mid of November (mean 05/11, Table 1), air temperature continuously decreased during the dormancy phases. Mean air temperature reduced from 5.8°C during the relatively short endodormancy phase (LF-t₁: 22.8 d) to 2.2°C during the long period of ecodormancy (t₁-t₁*: 86.2 d). Beginning of ontogenetic development was contrasting between the season 2012/13 (t₁*= 85 day of year (DOY), 26 March) and 2013/14 (t₁*= 35 DOY, 4 February, Figure 1). Between LF and SB mean air temperature in 2013/14 was always higher compared to 2012/13, leading to an earlier beginning of ontogenetic development and finally to an early start of bud swelling (2013/14: SB = 63 DOY, 4 March, Figure 1). Between SB and SG air temperature further increased in all seasons and reached on average 7.7°C. In the following 16 days from SG until OC, air temperature was strongly influenced by the annual weather variability. For this reason no homogeneous rise in air temperature was observed.

Table 1: Average timing, duration and air temperature (T in °C) in the phenophases: endodormancy [leaf fall (LF) - end of endodormancy (t₁)], ecodormancy [t₁ - beginning of ontogenetic development (t₁*)], and during ontogenetic development after t₁* (SB: swollen bud, SG: side green, GT: green tip, TC: tight cluster, OC: open cluster), 2011/12 - 2015/16, x: mean air temperature, s: standard deviation.

Variability of ABA content in cherry buds

Figure 1 shows weekly changes in ABA content of ‘Summit’ sweet cherry buds. It can be seen that in all seasons the ABA content increased from the first sampling date (S1: beginning of October) to maximum values in November or beginning of December. This plateau of maximum ABA contents (on average 5.65 μg g^-1 DW^-1, Table 2) were held for 3 (2012/13) to 6 (2015/16) weeks and marks approximately the endodormancy phase. Subsequently, ABA content steadily decreased until open cluster (OC) in April. Only in the season 2011/12 a second peak during ontogenetic development occurred, which cannot be explained and was never repeated in the following seasons. Additionally, 2011/12 was the season with the lowest ABA content at the first sampling date (0.35 μg g^-1 DW^-1) and the highest mean ABA content during endodormancy (10.1 μg g^-1 DW^-1, Table 2).

Table 2: ABA content (µg g^-1 DW^-1) of ‘Summit’ sweet cherry buds during the end of paradormancy (S1-LF), endo- (LF-t₁), ecodormancy (t₁-t₁*) and ontogenetic development (t₁*-OC), 2011/12 - 2015/16, given are means/standard deviations, different letters indicate significant changes (Tukey HSD, n=4, p=0.05), ABA_plateau gives the highest mean ABA content in each season (see dashed line in Fig. 1), S1: first sampling date, OC: open cluster, x: mean, s: standard deviation.

The initial mean ABA content in the other 4 seasons was much higher (3.15 μg g^-1 DW^-1).

Despite this fact, in all seasons the highest mean ABA content (marginal exception 2011/12) was reached during the endodormancy phase (Table 2: LF-t₁). In this period also the absolutely highest ABA content was observed (Figure 1). In the seasons 2012/13 and 2013/14 the ABA content significantly reduced after t₁, derived from the classic approach [1]. However, in the other 3 seasons the experimentally found date for t₁ was not exactly in agreement with a significant reduction in the ABA content (Table 3). In 2011/12, it significantly reduced 4 and in 2015/16, 3 weeks after t₁ and in 2014/15 it already declined 1 week before t₁. Assuming that the timing of t₁ is not an absolutely fixed date, it can be conclude that the significant reduction in ABA content after a period of constantly high values (3-6 weeks) was the first hint for endodormancy release. Our previous classical experiments yielded a mean chilling requirement for ‘Summit’ from 1 September until t₁ of about 40 chill portions (classic approach, Table 3). Using the significant reduction of ABA content after reaching the plateau (p<0.05), the chilling requirement for ‘Summit’ would be about 44 CP (metabolomic approach, Table 3), which nearly confirms our previously derived value for the chilling requirement (C*). The highest differences in C* between the classic and metabolomic approach occurred in the season 2011/12 and 2015/16, where C* reached 55.1 and 52.4 CP, respectively (Table 3).

Table 3: Dates of endodormancy release (t₁) and chilling requirement (C*) of ‘Summit’ sweet cherries, 2011/12 - 2015/16 according to experimental studies with twigs (classic approach) and derived from changes in abscisic acid (ABA) and sucrose (SUC) content (metabolomic approach), x: mean, s: standard deviation, CP: chill portion, DOY: day of year.

Variability of sugar content in cherry buds

The mean total sugar content of 69.1 mg g^-1 DW^-1 between first sampling and OC (S1-OC) consisted of 11% fructose (7.3 mg g^-1 DW^-1), 22% sucrose (15.5 mg g^-1 DW^-1) and 67% glucose (46.3 mg g^-1 DW^-1, Table 4). During this period, mean total sugar content in the buds reaches a minimum around leaf fall (Figure 2). Starting from LF, values raised during endodormancy (LF-t₁) and partially in the beginning of ecodormancy (t₁-t₁*), in order to reach a maximum during the latter phase (Table 4). Subsequently, total sugar content decreased until SB. After SB, values raised again until the stages TC or OC (Figure 2).

Table 4: Mean sugar content (mg g^-1 DW^-1) of ‘Summit’ sweet cherry buds in the seasons 2011/12 - 2015/16, during para- (S1-LF) endo- (LF-t₁), ecodormancy (t₁-t₁*), ontogenetic development (t₁*-OC) and for the whole period (S1-OC), given are means/standard deviations, S1: first sampling date, OC: open cluster

The seasonal variability of the total sugar content was mainly driven by changes of glucose, because it usually supplies the energy for metabolic processes in the buds. In contrast to this, fructose showed relatively constant values during both dormant phases and started to rise after SB with always significant higher values at the stage GT. Thus, the highest sucrose, glucose and total sugar content was found during ecodormancy (t₁-t₁*) and for fructose during the ontogenetic development (t₁*-OC, Table 4).

Sucrose was the only sugar component which strongly rose during endodormancy phase and stayed nearly constant during ecodormancy phase (Figure 3). With the beginning of ontogenetic development, it decreased nearly until SB and subsequently rose again until TC or OC. In all years the maximum sucrose content was observed during ecodormancy phase, ranging between 11.1 mg g^-1 DW^-1 (2015/16) and 27.4 mg g^-1 DW^-1 (2011/12, Table 5). It stepwise increased from the end of paradormancy (5.9 mg g^-1 DW^-1) to endodormancy (12.4 mg g^-1 DW- 1) until ecodormancy (21.0 mg g^-1 DW^-1).

Table 5: Sucrose content (mg g^-1 DW^-1) of ‘Summit’ sweet cherry buds during end of paradormancy (S1-LF), endo- (LF-t₁), ecodormancy (t₁-t₁*) and ontogenetic development (t₁*-OC), 2011/12 - 2015/16, given are means/standard deviations, different letters indicate significant changes (Tukey HSD, n=4, p=0.05), SUC_max gives the absolutely highest sucrose content during ecodormancy in each season, S1: first sampling date, OC: open cluster, x: mean, s: standard deviation

Interestingly, the high sucrose content during ecodormancy in the seasons 2011/12 (27.4 mg g^-1 DW^-1) and 2014/15 (25.1 mg g^-1 DW^-1) corresponded with high ABA contents in the same seasons (6.73 μg g^-1 DW^-1 and 5.87 μg g^-1 DW^-1).

The stable level of sucrose content during ecodormancy phase (dashed line Figure 3) allowed us to use this metabolite to identify the generally unknown dates of t₁ and t₁*. If one assumes, that the first sampling date with a constantly high sucrose content during ecodormancy marks the timing of t₁, the mean date of endodormancy release for ‘Summit’ would be on 2 December, if C* reaches 43 CP (Table 3). This result was very close to the value which was previously derived from our classic approach. On the other hand, the beginning of ontogenetic development (t₁*) would be the first date when the stable sucrose content during ecodormancy starts to reduce, significantly (Figure 3). This always happened around t₁*, previously derived by the classic approach.

Environmental drivers for the sucrose content

If one assumes that sucrose is a cryoprotective compound which protects buds to freeze during winter, the maximum sucrose content during ecodormancy is plausible and matches the coldest winter period (Table 1). For this reason the relationship between mean air temperature and sucrose content was further investigated (Figure 4). The strongest linear correlation coefficient (r=-0.90, n=20, p<0.05) was found between the sucrose content during ecodormancy and the average minimum temperature during endodormancy, where the strongest increase of sucrose took place (Figure 3). The correlation coefficient between sucrose content during ecodormancy and mean air temperature (r=-0.86) or maximum temperature during endodormancy (r=-0.78) was lower, which highlights the relevance of sucrose as a cryoprotective compound.

Figure 4: Relationship between mean minimum temperature (Tn) during endodormancy (LF and t₁) and sucrose content during ecodormancy (t₁-t₁*) for all 4 replications (trees) in 5 years (2011/12 - 2015/16).

Role of ABA and sugars during dormancy

The increase of ABA content and the transient plateau/peak is consistent with findings for poplar, which report an upregulation of critical enzymes (NCED3, ABA1, and ABA2) for ABA biosynthesis at 3 to 4 weeks of short days [9]. In grape [15], bud dormancy is also induced by the short day photoperiod and both the ABA content and the expression of ABA biosynthesis genes (VvNCED1 and VvNCED2) increased during the induction and maintenance of endodormancy. ABA inhibits respiration in non-dormant buds and down-regulates the expression of respiration-related genes such as VvaNAD, VvCYTC and VvCOX6, thus providing a molecular mechanism by which ABA induces bud endodormancy. Because the short day photoperiod induces bud endodormancy in grapevines, the up-regulation of the ABA biosynthesis genes VvNCED1 and VvNCED2 by the short day photoperiod is consistent with the putative role of ABA in the induction and maintenance of endodormancy. The elevated ABA content during ecodormancy could be also the reason for preventing growth and development of small bud tissues during early stages of development, as assumed for seed germination related to embryogenesis [16].

Sugars are providers of carbon and energy and fulfill an important role in coordination with hormonal signalling pathways and controlling various plant physiological processes [17]. Distinct glucose, sucrose, and fructose signalling pathways can be discerned. These signalling pathways might be strongly influenced by the activities of sucrose splitting enzymes (vacuolar, cell wall and neutral invertases, sucrose synthase) since they have strong impact on sucrose to hexose ratios, which are an important parameter in plant responses to the environment, especially under stress. Next to growth and stress responses, it can be expected that sugar signalling is of great importance in flowering time control. This major developmental transition directly affects yield and its exact timing is essential for plant fitness. Recent findings suggest that both subcellular compartmentation and route of sucrolysis are important for plant development, growth, and yield. Enzymes of sucrose metabolism (invertase, hexokinase, fructokinase, sucrose synthase, sucrose-6-phosphate synthase) are not produced from single genes, but from paralog families in plant genomes [18]. It is important to note that it is the differential localisation of both the sugars themselves as well as the sugar metabolizing enzymes, that ultimately led to sugar signalling, and the authors conclude that a combination of subcellular complexity and gene duplication/sub-functionalization gave rise to sugar signalling as a regulatory mechanism in plant cells. Since carbon metabolism occurs simultaneously in different organelles, different sensors may be required under specific conditions or circumstances. Different sucrolytic enzymes might be therefore important for channelling carbon into different metabolic routes, and they further postulate that sugar signals are specific for each paralog/ pathway. It was shown by Bonhomme et al. [19] that floral peach buds - that have come out of endodormancy - have a high capacity to synthesize ATP, which is involved in many metabolic pathways. This indicates an important role of soluble carbohydrates in providing carbon and energy required for the biosynthetic and respiratory metabolism.

Sucrose, as well as glucose and fructose accumulations during winter in plants have often been reported and appear to be related to their properties of cryoprotection and osmotic regulation in the control of bud development [20-22]. Our study clearly confirmed the cryoprotective role of sucrose during ecodormancy in sweet cherry buds. Additionally, it was shown that the annual sucrose content in cherry buds strongly depends on minimum temperatures during endodormancy, where the sucrose content strongly increased. Spring brings warmer and longer days, and the carbohydrate content generally decreases together with bud burst. This supports remobilization of stored resources for the plant to support leaf and stem production, independently of an external supply of nutrients during the active period of growth at flushing time [20]. However, reports that have been described the relationship between bud growth and carbon metabolism in fruit trees are scarce [11]. In addition to the modification in sugar levels, the onset of bud development is linked with variations in almost all metabolic pathways. ABA, auxin and cytokines have been described as major plant growth regulators, involved. Autumn accumulation of soluble sugars (mainly sucrose) is considered a determinant factor for cold tolerance or winter survival and low(er) concentration(s) of accumulated carbohydrates are associated with poor winter survival and reduced spring regrowth. Ito et al. [11] postulate that higher activities of sugar catabolizing enzymes should enhance the capacity of buds to attract assimilates, thereby stimulates bud growth.

The increase of the total sugar content, as well as glucose and sucrose indicate that cherry buds importing/synthesise these carbohydrates until LF/t₁. Accumulation of sugars in buds of other species during autumn and early winter followed similar patterns [22]. The high and stable sucrose content was likely a perfect indicator for the description of the ecodormancy phase, which allows to derive the timing of t₁ and t₁*.

The decline in total, glucose and sucrose content after t₁* mark the resumption of bud metabolic activity. This pre-growth activation has also been observed in other species and was correlated with increased sugar metabolism [23]. The reactivation of sugar metabolism in buds precedes the start of flower morphogenesis, which has high energy requirements. The drop of sucrose and glucose content approximately at t₁*, but comparatively the significant increase in fructose content at GT, indicates a rapid onward metabolism of hexoses. Maurel et al. [23] demonstrated in Prunus persica that hexoses, or at least glucose, may play a key role in triggering the onset of bud break in peach trees. This conclusion is consistent with many reports describing the ability of soluble carbohydrates (glucose and sucrose) to regulate a variety of physiological processes, including induction of flowering. Nevertheless, we agree with Xin and Browse [10] that further investigations are necessary to fill the gap of knowledge of the precise role and the cross-talk of sucrose and hexoses during cold acclimatisation, especially in fruit trees during winter rest and plant development in spring. For deeper and comprehensive understanding the complex of metabolism and to identify the most relevant components during the different phases of dormancy integrative, multidisciplinary approaches using tools from transcriptomics in conjunction with metabolomics and biochemical analysis, including different compartments, from plant tissue to plant organ, are necessary.

In this study the annual variability of ABA and sucrose, glucose and fructose in 5 seasons was presented. Already these metabolites allowed a deeper understanding of dormancy regulation in sweet cherry buds. Following highlights can be summarised:

During endodormancy, ABA content in cherry buds reaches a maximum (plateau), which was maintained for 3-6 weeks. The significant reduction of the ABA content after this period was nearly in agreement with the date of endodormancy release of ‘Summit’ trees.

Minimum temperature during endodormancy regulated the increase of sucrose up to a maximum value, which nearly stayed constant during ecodormancy. This highlights the cryoprotective role of sucrose during the period with the lowest temperatures. The constant level of sucrose during ecodormancy allowed to derive the timing of t₁ and t₁* for cherry trees.

These results show, the use of selected metabolites allow identifying the unknown phenological phases during winter rest of trees, which are important to improve classical phenological model approaches.

This work was supported by the Deutsche Forschungsgemeinschaft (DFG) in the project “Progress in Phenological Modelling on the Basis of Metabolomic Approaches” by the grants (CH 228/5-1, RA 1008/9-1). We are very thankful to Mrs. Susanne Moryson and Mr. Stefan Heider who supported our experimental works at Humboldt-University.