Seasonal changes in the physiology and metabolism of grapevine latent buds

Our study reveals distinct physiological and metabolic phases of grapevine bud dormancy, identifying sugar-phosphate dynamics and Tre6P accumulation as putative key signals linking metabolic reactivation to bud burst.

Abstract

Grapevine (Vitis vinifera L.) buds undergo seasonal dormancy to survive unfavourable conditions and synchronize growth with environmental cues. While dormancy transitions have been widely studied in temperate woody perennials, the physiological and metabolic dynamics underlying these transitions in grapevine remain poorly understood. Our study investigates seasonal changes in the physiology and metabolism of perennating buds in V. vinifera cv. Cabernet Sauvignon, focusing on dormancy depth, respiration, sugar metabolism, and cell cycle activity. We identified three distinct phases of bud quiescence: (i) para-dormancy (early summer), characterized by active metabolism and high levels of tricarboxylic acid cycle intermediates, shikimate, and myo-inositol; (ii) endo-dormancy (late summer to autumn), where dormancy depth peaked in late summer and was marked by reduced respiration and bud water content, and a sharp decline in hexose levels with a concomitant increase in raffinose; and (iii) eco-dormancy (winter to spring), featuring increased respiration, sugar mobilization (notably sucrose, glucose, and trehalose), and reactivation of cell division, with a shift of cells into the G₂ phase. Just prior to bud burst, we observed a significant accumulation of sugar-phosphates, providing evidence that supports their role in promoting bud burst also in grapevine. Our findings provide new insights into the biochemical and physiological regulation of bud dormancy and bud burst, contributing to a deeper understanding of dormancy transitions in perennial crops.

Introduction

Grapevine (Vitis vinifera L.) is the most economically important fruit crop worldwide, grown and commercially traded between >190 countries (http://www.fao.org/statistics/en). Grapevine was domesticated as a deciduous perennial (Terral et al., 2010); however, it displays wide plasticity to climate when grown in warmer, lower latitude climates (Camargo et al., 2012; OIV, 2017; Velappan et al., 2022a).

The vitality of the latent bud [syn. latent bud, N+2 (Lavee and May, 1997) or R2 (Torregrosa et al., 2021)] is vital to sustainable production, as it bears the vegetative and reproductive initials that form the next cycle of production following growth arrest and bud burst (Lavee and May, 1997). The bud develops under quiescence over a period of nearly 12 months or four seasons, and yet our knowledge of the eco-physiology, and metabolic control of the bud is largely restricted to winter and spring (De Rosa et al., 2021; Velappan et al., 2022b). Hence, a considerable need exists for a more integrated understanding of the development and vitality of the bud throughout its life cycle. This includes the transitions into and out of dormancy in the strict sense (physiological dormancy, syn. endo-dormancy).

The major seasonal cues of light and temperature affect the activity of axillary buds by influencing energy and hormone functions. In non-dormant grapevine buds subjected to forcing conditions, light was shown to induce a photomorphogenic response and accelerated bud burst rates (Signorelli et al., 2018, 2020). In other woody perennial species such as Rosa spp., light is an obligate requirement for bud growth, where it was shown to activate sugar mobilization and induce the expression of vascular invertase enzymes (Girault et al., 2010; Rabot et al., 2012). Likewise, the sorghum (Sorghum bicolor L.) mutant for phytochrome B (phyB) was shown to be unable to fully develop axillary buds, and this was correlated with lower expression of SUGARS WILL EVENTUALLY BE TRANSPORTED (SWEET) sugar transporters and cell wall invertase, suggesting that light perception is key to promote the correct apoplastic supply of sugars that promotes bud development (Kebrom and Mullet, 2016). Moreover, in Rosa spp., sucrose was shown to augment the effect of light on bud outgrowth (Rabot et al., 2012). This effect was also observed when applying non-metabolizable sucrose analogues but not when applying mannitol (Rabot et al., 2012; Barbier et al., 2015a, b). These observations suggest that, although sucrose could have a metabolic effect, its impact was primarily via sugar signalling.

In the last decade, several fundamental studies have shown the importance of sugars in metabolic and hormone control of quiescence and bud outgrowth in annual herbaceous plants (Considine and Foyer, 2023). For instance, sucrose deprivation rapidly represses axillary bud activity in sorghum (Kebrom and Mullet, 2015), while shoot tip removal promotes bud outgrowth in pea (Pisum sativum L.) through transient increases in sugars, including trehalose 6-phosphate (Tre6P) (Mason et al., 2014; Fichtner et al., 2017). These findings established sugar availability and Tre6P as central regulators of bud activation, integrating with hormonal and Target of Rapamycin (TOR) kinase signalling (Xiong et al., 2013; Pfeiffer et al., 2016). Nevertheless, these mechanisms have been primarily characterized in model annuals, and far less is known about how carbohydrate metabolism and sugar signalling operate within the dormancy cycles of woody perennials.

Recent transcriptomic and metabolomic studies have begun to address this gap. In grapevine, short-day photoperiods reprogrammed gene expression associated with abscisic acid (ABA) signalling, carbohydrate metabolism, and cell cycle repression, marking dormancy induction (Fennell et al., 2015; Noriega et al., 2022; Velappan et al., 2022a, b). Metabolomic analyses in tree peony revealed activation of glycolytic pathways and flavonoid accumulation during dormancy transition (Zhang et al., 2020), while in grapevine buds, sucrose and raffinose family oligosaccharides (RFOs) accumulated during endo-dormancy, consistent with their roles in cold tolerance and sugar signalling (De Rosa et al., 2022). In other temperate species, including pear and sweet cherry, dormancy progression was associated with respiratory suppression, accumulation of protective metabolites, and strong dependence on temperature fluctuations (Baldermann et al., 2018; Horikoshi et al., 2018; Fouché et al., 2023). Early physiological studies further reported reduced respiration and water content during endo-dormancy in kiwifruit and cherry (McPherson, 1997; Götz et al., 2014). Collectively, these findings indicate that dormancy establishment and release involve coordinated regulation of respiration, carbohydrate metabolism, and stress-related metabolites, yet the seasonal dynamics of these processes in grapevine buds remain poorly understood.

Although dormancy regulation in woody perennials shares conserved mechanisms with annual species, the seasonal progression of quiescence and dormancy in grapevine represents a far more complex ecological adaptation. The timing and coordination of metabolic, respiratory, and environmental cues remain poorly resolved. In this study, we investigated how seasonal changes in sugars and organic acids relate to the physiological state, metabolic activity, and environmental variables of V. vinifera cv. Cabernet Sauvignon buds from mid-summer through early spring. Our results provide a benchmark to advance our understanding of the role of metabolic control in the perception and transduction of seasonality in important perennial species.

Materials and methods

Unless otherwise stated, all chemicals were analytical grade from Sigma Aldrich (NSW, Australia; www.sigmaaldrich.com).

Plant material and growth conditions

Canes of V. vinifera (L.) cv. Cabernet Sauvignon containing buds from node 3 to 12 acropetally were collected from a vineyard in Margaret River (33°47′S 115°02′E), WA, Australia. Margaret River has a Koppen–Geiger climate type Csb (temperate, dry summer, warm summer; Peel et al., 2007). The vines were maintained on a single wire system, head trained, and cane pruned to approximately eight canes per vine. Different nodes were used for different measurements, as indicated below. A single cane was taken per plant across the year; vines were not repeatedly sampled. The sampling dates were 5 January, 15 February, 4 April, 10 May, 28 June, 9 August, and 20 September, chosen as they reflected the major transitions in dormancy of cv. Merlot in an adjacent vineyard (Velappan et al., 2022b). At the earliest sampling time (January), all basal buds up to node >15 were mature and lignified. These canes were transported to the laboratory and immediately analysed or planted for time-course experiments.

Depth of dormancy

For each sampling date, two replicates of 50 single-node cuttings, from node 3 to 12, were planted in pots containing a soil mix (fine composted pine bark, coco peat, and brown river sand in the ratio of 2.5:1:1.5). As a positive control for viability, the same number of single-node cuttings were immersed in 1.25% (v/v) hydrogen cyanamide (HC) (0.31 M) in water for 30 s before planting. Explants were grown in a controlled-environment chamber under dark–light (12:12 h) conditions at 23 °C, and the irradiance, provided by light-emitting diodes, was between 200 µmol photons m⁻² s⁻¹ and 400 µmol photons m⁻² s⁻¹. Bud burst was recorded at least three times a week, and the depth of dormancy was estimated as the time required to reach 50% of bud burst (BB50). Bud burst was scored when a green leaf tip became visible; that is, at stage EL4 according to the modified Eichorn–Lorenz scale (Coombe, 2004). At the end of the experiment, all remaining buds were dissected and assessed for necrosis, and corrections were made to the cumulative bud break based on the number of healthy buds.

Weather data

The weather data were obtained from Moss Wood vineyard in Margaret River. The average daily mean air temperatures were calculated from hourly air temperatures using the R statistical package (R Development Core Team, 2014) and plotted using the ggplot2 package of R (Wickham, 2009) as scatter dot plots fitted with a quadratic spline with degree of freedom (df)=4 and degree=2. For the chilling model, we used the Daily Positive Utah Chill Unit (DPCU) model, which is a modified version of the Utah model suggested to accurately predict chill accumulation in mild winter regions (Linsley-Noakes et al., 1995). This model records chill as positive chill units (PCU) and considers optimal chilling to occur at temperatures between 2.5 °C and 9.2 °C, and nil chilling at temperatures <1.5 °C or >12.5 °C. Hence, 1 h at 2.4 °C<T≤9.1 °C equals 1 PCU, 1 h at 1.4 °C<T≤2.4 °C or 9.1 °C<T≤12.4 °C equals 0.5 PCU, and 1 h at T≤1.4 °C or T≥12.4 °C equals 0 PCU. The cumulative chilling hour or chill unit was estimated for each day from the hourly mean air temperature, calculated and represented as a line plot using the R package (R Development Core Team, 2014) and the ggplot2 package of R (Wickham, 2009), respectively.

CO₂ release and O₂ consumption

To determine the CO₂ release, four pools of eight buds each were placed in an insect respiration chamber (6400-89; Li-COR, Lincoln, NE, USA; www.licor.com) attached to a Li-6400XT portable gas exchange system. Only buds from nodes 7 to 11 were used for respiration. The measurements were performed in complete darkness, at 23 °C, in CO₂-controlled air (380 µmol CO₂ mol⁻¹ air) with 100 µmol s⁻¹ air flow, at 55–75% relative humidity.

Oxygen consumption was determined in four pools of eight buds each, nodes 5–12. The buds were placed in a 4 ml micro-respiration chamber and the O₂ content was monitored for at least 10 min using an O₂ microsensor (Clark-type, OX-25; Unisense, Aarhus, Denmark). Measurements were performed at 23 °C in complete darkness and the micro-respiration chamber was kept under water to prevent sudden changes in temperature. The volume of the buds was determined using a 25 ml density bottle, where water was used, and the volume was measured by determining the milligrams of water displaced by the bud. This information was necessary to determine for each replicate the final volume of air inside the chamber.

For both CO₂ release and O₂ consumption, the buds were dried in an oven for 72 h at 70 °C to determine the DW, and the respiration rates were referred to the DW of the buds. Respiratory quotient (RQ) values cannot be deduced directly from these data because the conditions of measurements were different for CO₂ and O₂. Therefore, we calculated a relative RQ (RRQ) in which the RQ was calculated for each sampling date, and was divided by the average of the RQ, to normalize the average to 1. Hence the higher values indicate a more anaerobic metabolism, while the lower values indicate a more aerobic metabolism, assuming that carbohydrates were the predominant respiratory substrate.

Bud water content

Three sets of 10 buds each, from node 3 to node 12, were used to determine the moisture content. The buds were dissected from the cane, and the fresh weight (FW) was recorded. Then, they were desiccated during 96 h in an oven at 70 °C and the dry weight (DW) was determined. The moisture content was calculated as follows:

$$\mathrm{Moisture}(\text{\%})=100\times (\mathrm{FW}-\mathrm{DW})/\mathrm{FW}.$$

Internal pO₂ profiles

The internal oxygen was measured using an oxygen microsensor (Clark-type) with a 25 µm tip (OX-25; Unisense A/S, Aarhus, Denmark). The calibration of the sensor and measurements were done using the Sensortrace Suite software (https://unisense.com). For calibration at 0 kPa oxygen, the sensor was flushed with N₂ until stabilization, and for air-saturated O₂ concentration, an aquarium pump was used to flush the sensor with atmospheric air (21% O₂). The buds were removed from the cane, node 5, and stood on a flat surface, and then the electrode was electronically inserted from the top to reach the centre of the meristem (Shaw et al., 2017). The depth of the path was ∼2200 µm, depending on the size of the bud, with steps of 35 µm, and three measurements per step were performed. At least four buds were used for each sampling date.

Apoplastic connection

Eosin Y and Acid fuchsin solutions were prepared by dissolving 0.5 g of the respective dye in 0.01 M potassium phosphate buffer pH 6.0. The solutions were filtered through GF/A filter paper and then through a 0.2 µm membrane filter. Eight single-node cuttings, nodes 3 and 4, for each sampling date were cut 1.5 cm below the base of the bud and immediately transferred to 25 ml plastic tubes containing 2 ml of Eosin Y or Acid fuchsin solutions (four replicates each), in such a way that the cane was immersed but the bud was not submerged. The incubation time in the solution was 4 h. A razor blade was used to cut the cane in the longitudinal plane and the images were taken using a camera coupled to a magnifying glass.

Cell cycle status

Intact nuclei were isolated from fresh buds using slight modifications to the protocol used by Arumuganathan and Earle (1991) and Hermawaty et al. (2022). Two or three buds (∼50 mg FW), from nodes 6–8, per biological replicate (three biological replicates) were dissected from the canes. The scales and woolly hairs surrounding the bud were carefully removed using forceps prior to chopping them with a razor blade in ice-cold nucleus-isolation buffer (NIB) at pH 7.4 [10 mM MgSO₄, 50 mM KCl, 5 mM HEPES-K⁺, 6.5 mM DTT, 0.5% (v/v) Triton X-100, and 3% (w/v) polyvinylpyrrolidone (PVP)-40] on ice, and incubated on ice for 2 h with gentle swirling every half an hour. Following incubation, the suspension was passed through a 100 µm and a 41 µm nylon mesh filter (Millipore, Germany), respectively, and centrifuged at 450 g for 8 min at 4 °C. The supernatant was carefully discarded. The nuclei were resuspended in 2 ml of NIB and treated with 50 µg ml⁻¹ RNase A (Thermo Scientific, Australia, EN0531) for 6 min at room temperature before being stained with 20 µg ml⁻¹ propidium iodide and stored on ice until analysis. The ploidy of the nuclei was determined by fluorescence-activated cell sorting (FACS) on a FACSCaliber flow cytometer (BD Sciences; Franklin Lakes, NJ, USA; https://www.bdbiosciences.com) at the Centre for Microscopy Characterisation and Analysis (University of Western Australia, Perth, WA). The flow cytometer was equipped with a primary blue 488 nm laser, and data for ∼100 000 nuclei were recorded [i.e. until 20 000 G₂ (4C) nuclei were collected]. The proportion of nuclei with 2C and 4C DNA content was recorded. The data were visualized in real-time using scatter dot plots [forward scatter (FSC) and side scatter (SSC)] and histograms. The results were analysed using the Flowing Software (version 2, https://flowingsoftware.com) by manual gating to eliminate debris from the population of interest on the scatter plots and subsequent generation of histograms from the scatter plot data in which the different populations were gated to obtain the final proportion of G₁, S, and G₂ nuclei computed by the software.

Metabolite extraction and measurements

For each sampling date, two buds from nodes 6 and 8 were excised in the field, immediately collected in microtubes, and placed in liquid nitrogen to be used for metabolite analysis (n=5). The frozen buds were ground to a fine powder at liquid nitrogen temperature. Water-soluble metabolites were extracted with chloroform–methanol as described previously (Lunn et al., 2006). Sugars were separated by hydrophobic interaction HPLC using a Luna Omega Sugar column (3 µm, 100 Å, 150×2.1 mm; Phenomenex, Torrance, CA, USA, www.phenomenex.com) fitted with a SUGAR (4×2.0 mm) security guard cartridge (Phenomenex) according to the manufacturer’s instructions, and quantified by tandem MS on an AB Sciex 5500 Q-Trap triple quadrupole mass spectrometer (Feil and Lunn, 2018). Tre6P, other phosphorylated intermediates, and organic acids were measured by anion-exchange HPLC coupled to tandem MS as described in Lunn et al. (2006) with modifications as described in Figueroa and Lunn (2016).

Statistical analysis

Normality of the data was evaluated using the Shapiro–Wilk test, and homogeneity of variances with Levene’s test. When both assumptions were met, data were analysed by one-way ANOVA followed by Tukey’s HSD post-hoc test to identify pairwise differences at the P<0.05 level. In cases where assumptions were not met (e.g. CO₂ release and cell cycle status), data were log-transformed, which in most instances restored approximate normality and homogeneity, allowing subsequent ANOVA and Tukey’s HSD. When normality and/or homogeneity were not achieved even after transformation (e.g. O₂ consumption), more robust alternatives were applied, including Welch’s ANOVA (for heteroscedastic data) and the non-parametric Kruskal–Wallis test, with appropriate post-hoc tests. For metabolite profiling, only a minority (∼15%, e.g. Tre6P, citrate, malate, glucose, maltose, and raffinose) of metabolites met strict normality and variance homogeneity criteria; however, given that ANOVA is considered robust to moderate deviations and to ensure methodological consistency across the dataset, one-way ANOVA with Tukey’s HSD was applied uniformly.

Results

Depth of dormancy through the year and the effect of weather conditions

We evaluated the depth of dormancy in buds from the southern hemisphere summer to spring (seven time points: January, February, April, May, June, August, and September). HC treatment was used as a positive control for bud viability, as well as a reference for the depth of dormancy. The results showed that the depth of dormancy was greatest in late summer (February), with untreated buds taking 297 d to reach BB50 (Fig. 1A). Overall, from mid-summer to early winter, the depth of dormancy was considerable, taking ∼60 d or more to reach BB50. By June, the untreated buds reached BB50 within 25 d and the difference between the untreated and HC-treated positive control buds was negligible (Fig. 1A). January and February were the hottest months, with temperature maxima up to 39 °C, and the coolest months were July and August with maxima <20 °C and minima as low as 4 °C (Fig. 1B). Rainfall occurred more frequently from May to October, with occasional rainfall >40 mm d^–1 in June and July (Fig. 1B). The mean temperature decreased from January to August, with a more significant decline occurring from April to June (Fig. 1C). During the period of greatest dormancy (BB50), mean temperatures were >20 °C (Fig. 1C). Environmental chilling, determined as PCU, did not accumulate appreciably until May, by which time the depth of dormancy had declined substantially from the peak in late February, and the difference between untreated and HC-treated buds was marginal (Fig. 1B). In this sense, we identified a period of physiological dormancy (commonly referred to as endo-dormancy) between February and March, and two periods of quiescence, one prior to dormancy (January) and the other after dormancy (from April), often called para-dormancy and eco-dormancy by many authors. The initial decrease in BB50 from February to April did not correlate with temperature change or rainfall, although there was a significant decrease in terms of photoperiod during this time. Previous studies have identified a photoperiod threshold of 12.4 h for the para- to endo-dormancy transition in Cabernet Sauvignon (Cragin et al., 2017), a threshold that was reached before our sampling date in April (Fig. 1D). The second decrease in the depth of dormancy (April to July) coincided with the decrease of mean temperatures and an increase in cumulative PCU (Fig. 1C). The average rainfall was low during the period of deepest dormancy (February) and only increased appreciably from April until July, declining thereafter (Fig. 1D). In this sense, we did not observe a clear correlation between the depth of dormancy and rainfall.

Fig. 1.

Seasonal changes in perennating bud dormancy in grapevine. (A) Dynamic changes in the depth of dormancy of grapevine buds, grown as explants in permissive conditions and expressed as the time to 50% bud burst (BB50). Grapevine canes were taken at intervals during the southern hemisphere summer to spring (5 January, 15 February, 4 April, 10 May, 28 June, 9 August, and 20 September). Single-node cuttings were treated with water (C) or 1.25% hydrogen cyanamide (HC) for 30 s, planted on a potting mix, and subjected to forcing conditions in controlled growth chambers. The violin plots are the result of five replicates. Asterisks indicate statistical differences based on one-way ANOVA with Tukey’s post-hoc HSD testing at P<0.05 for the samples connected with a light dashed line. (B) Daily mean, maximum, and lower air temperature represented by a scatterplot, and daily rainfall represented by a dashed line in the Margaret River region of Western Australia in 2016. (C) Relationship between the depth of dormancy, temperature annual trend, and positive chill units (PCU) accumulated. The dotted line represents the depth of dormancy as determined in (A). Temperature is represented as a regression of the daily average temperature. The grey area represents ±95% confidence level intervals for prediction for the adjusted regression. (D) Relationship between the depth of dormancy, rainfall annual trend, and length of photoperiod. The dotted line represents the depth of dormancy as determined in (A). Rainfall is represented as a regression of the daily rainfall. The grey area represents ±95% confidence level intervals for prediction for the adjusted regression. Photoperiod is represented as hours to sunrise/sunset and dawn/dusk. The horizontal dashed line at 12 h is plotted as a reference to see when days become shorter than nights. Meteorological seasons for the southern hemisphere temperate zone are marked in each panel.

Moisture content and vascular connections

In order to determine the relationship between moisture content, vascular development, and bud dormancy, the moisture content and apoplastic connectivity were evaluated throughout the year. It is shown in Fig. 2A that the moisture content of the buds was higher in summer and then became lower from late summer to late winter, rising by spring (September). This behaviour is counterposed to the annual trend in rainfall (Fig. 1D). The apoplastic path is vital to water transport in plants. As observed in Fig. 2B, the buds were apoplastically isolated from the cane throughout this entire period, explaining the absence of correlation between rainfall and water content in the buds. It is noteworthy that this limited apoplastic transport is sustained until spring when some buds are already swelling (see September in Fig. 2B).

Fig. 2.

Apoplastic isolation of buds throughout the season. (A) Moisture content. Bars indicate the mean of five biological replicates consisting of two buds each, error bars represent the SD, and different letters show statistical differences based on one-way ANOVA with Tukey’s post-hoc HSD testing at P<0.05, n.d., not determined; (B) Apoplastic connectivity, as visualized by the acrotonic passage of two apoplastic dyes—Acid fuchsin (upper row) and Eosin Y (lower row).

Internal oxygen profiles and metabolic activity

Internal oxygen profiles (pO₂) of the buds were also evaluated, as oxygen levels were suggested to regulate dormancy and bud burst (Meitha et al., 2015; Considine et al., 2017). In January and February, the internal O₂ concentration was, on average, >100 µM for the peripheral 1000 µm and then decreased to 25–50 µM in the inner part of the bud. During April, May, and June, the levels of internal O₂ concentrations were more constant throughout the bud (∼100 µM), with a more pronounced decrease from 1500 µm in June. In September, however, the oxygen levels were greater for the outer sections of the bud than the inner section, at ∼1500–2000µm, where it was more hypoxic (Fig. 3A). The spatial variation within replicates is expected, considering the heterogeneity of the tissue, which comprises layers of bud scales alternated with trichomes (Signorelli et al., 2020).

Fig. 3.

Internal O₂ concentration and respiration of buds during the year. (A) Internal oxygen profiles (pO₂) represented as scatterplots of raw data (n>4), with a regression curve applied and 99% confidence intervals (in grey shading). (B) CO₂ production (n>4). (C) O₂ consumption (n>4, except August n=2). (D) relative respiratory quotient (RRQ). For (A–C), individual observations are plotted as light dots, the line is an adjusted regression of the data, and the grey regression represents ±95% confidence level intervals for prediction for the adjusted regression. For CO₂ production, different letters show statistical differences based on one-way ANOVA on log-transformed values, followed by Tukey’s post-hoc HSD testing at P<0.05. For O₂ consumption, different letters show statistical differences based on Welch’s ANOVA followed by pairwise Welch’s t-tests with Holm–Bonferroni correction for multiple comparisons at P<0.05. For (D), the data are represented as a violin plot, with individual observations plotted as light green dots.

The metabolic activity was followed by measuring the CO₂ release and O₂ consumption (Fig. 3B, C). The highest metabolic activity was found in summer (January) before both CO₂ production and O₂ consumption declined. For both parameters, there was an increase from May to September; however, this was only statistically significant for CO₂ release. The greater increase in CO₂ release relative to O₂ consumption towards the end of the season results in a slight increase of the RRQ (Fig. 3D), which may indicate a change towards metabolism of more oxidized substrates or use of reducing power for alternative processes. Likewise, the low O₂ consumption found during the dormancy peak (February), with a CO₂ release that was moderate, resulted in a high RRQ (Fig. 3D).

Cell cycle

The population of cells in the different phases of the cell cycle was evaluated during the year to determine whether there was any correlation between the quiescent state of the bud and the cell cycle (Fig. 4). In January, ∼54% of cells were in G₁ phase, 37% in S phase, and 8% in G₂ phase. In the transition to late summer (January–February), when the depth of dormancy was greatest, the number of cells in S phase significantly declined. After this period, the fraction of cells in S phase remained unchanged until bud burst, at ∼30%. The fraction of cells in G₁ remained unchanged from January to late June with values of ∼50%, in line with a number of studies showing that the cell cycle arrests in G₁ phase in dormant tissues (Velappan et al., 2017). However, in August and September the fraction of cells in G₁ had declined to ∼38%, coupled with an increase in the fraction of G₂ cells. This change in the distribution of cells at each phase of the cell cycle suggests that dormancy release had been initiated. The G₂/G₁ ratio is a common metric of cell division. Here we observed that August and September saw a 2-fold increase in the G₂/G₁ ratio, indicating a significant shift towards proliferation (Fig. 4B). Notwithstanding, these data have no spatial resolution among the heterogenous tissues that comprise a primordial shoot.

Fig. 4.

Variation of the cell cycle phases through the year. (A) Cell cycle phases per month. (B) G₂/G₁ ratio. Bars indicate the mean of five biological replicates consisting of two buds each, error bars represent the SD, and different letters show statistical differences based on one-way ANOVA on log-transformed values with Tukey’s post-hoc HSD testing at P<0.05, n=3.

Analysis of sugars and metabolic intermediates during the seasonal changes

We investigated seasonal changes in the levels of sugars and respiratory intermediates in the buds. A principal component analysis (PCA) of our metabolite data showed that January and February are clearly separated from the other months, which all clustered together, with a few replicates being away from their group (Fig. 5A). In this sense, the PCA of all metabolite data distinguished the samples by depth of dormancy, as February and January were the months where the deepest dormancy was found (Fig. 1). We performed a hierarchical cluster analysis (HCA) of the metabolite data for all the months and replicates. In general, we observed that most sugars with an energetic function, sugar-phosphates, and tricarboxylic acids (TCAs) peak in January; their levels are low during the period of quiescence until September when many of them increase again (Fig. 5B). The sugars that have been associated with freezing protection and osmotolerance had a lower concentration during the summer (January–February) and spring (September) periods and greater concentrations during the autumn and winter periods (April–August), consistent with their functions and the lower water content (Figs 2A, 5B). As expected, myo-inositol, which is used to produce galactinol and RFOs, has a contrasting pattern.

Fig. 5.

Principal component analysis (PCA) and hierarchical cluster heatmap of metabolites from grapevine buds during the year. (A) PCA using all metabolite data. Samples from different months are represented in different colours and labelled accordingly. Ellipses highlight either replicates from the same month that cluster together and separate from other months (e.g. Jan and Feb), or samples from different months that group closely in the PCA space. The eigenvectors in magenta show the contribution of the main variable (metabolite) driving the principal component. (B) Heatmap: the values were scaled for each metabolite, and the order of the months was fixed. Five replicates of grapevine buds were used to determine the concentration of metabolites.

When only sugars were considered for the PCA, we observed that not only January and February were clearly distinct from the other samples, but also September (Fig. 6A). Given that many of these sugars have an energetic function, it is not surprising that these months are separated from the others considering that we observed higher moisture content and metabolic activity (Fig. 2A, B).

Fig. 6.

Principal component analysis (PCA) and hierarchical cluster heatmap of sugars during the year. (A) PCA data using only sugars. Samples from different months are represented in different colours and labelled accordingly. Ellipses highlight either replicates from the same month that cluster together and separate from other months (e.g. January, February, and September), or samples from different months that group closely in the PCA space. The eigenvectors in magenta show the contribution of the main variable (metabolite) driving the principal component. (B) Heatmap: the values were scaled for each metabolite, and the order of the months was fixed.

The hierarchical cluster analysis of these metabolites showed that myo-inositol did not group with any of the other sugars, similar to the separation from all metabolites assayed (Figs 5B, 6B). Raffinose was also separated from the other sugars, and then all the other sugars clustered together, with 1-kestose displaying a more distinct profile (Fig. 6B).

To further understand the profile of the metabolites and their absolute value, we plotted each metabolite per month, grouped as sugars, sugar alcohols, and hexose-phosphates (Fig. 7), and glycolytic and TCA cycle intermediates (Fig. 8). Looking at their concentration, sucrose, tartaric acid, citrate, glucose, and fructose were the most abundant metabolites in the buds (Figs 7, 8). Glucose and fructose followed a very similar pattern, characterized by high values in January and August–September (Fig. 7) when respiration rates were also high (Fig. 3B, C); however, the ratio between them was not always constant, as we observed a significant decrease of the [Glc]/[Fru] ratio towards the bud burst period (Fig. 9). Galactose showed the same profile as glucose and fructose, but at lower concentrations (Fig. 7). Sucrose and its phosphorylated form (Suc6P) gradually accumulated from low levels in January to a maximum in July/August, which preceded the peak of glucose and fructose. The results suggest that the buds accumulate sucrose through winter, and prior to bud burst sucrose is catabolized into glucose and fructose which are further catabolized to provide energy.

Fig. 7.

Concentration of sugar, sugar alcohols, and hexose-phosphates in grapevine buds at different stages of the year. The horizontal dotted bars connecting different months indicate statistical differences based on one-way ANOVA with Tukey’s post-hoc HSD testing at P<0.05, n=5. The grey area represents ±95% confidence level intervals for prediction for the adjusted regression.

Fig. 8.

Concentration of glycolytic and TCA cycle intermediates in grapevine buds at different stages of the year. The horizontal dotted bars connecting different months indicate statistical differences based on one-way ANOVA with Tukey’s post-hoc HSD testing at P<0.05, n=5. The grey area represents ±95% confidence level intervals for prediction for the adjusted regression. The hexose and hexose-phosphate data are the same as those reported in Fig. 7, and are reproduced here to allow direct comparison with downstream glycolytic intermediates.

Fig. 9.

Glucose to fructose ratio in grapevine buds throughout the season. The violin plots show data from five replicates (represented in circles). The horizontal dotted bars in black connecting different months indicate statistical differences based on one-way ANOVA with Tukey’s post-hoc HSD testing, with **P<0.01 and ***P<0.01, n=5.

The accumulation of monosaccharides in spring is key to promote bud burst, as they will act as energy substrates to sustain growth, but the conversion of sucrose to glucose and fructose will also double the osmotic potential, thereby increasing the turgor pressure and driving cell expansion. Moreover, these sugars and some of the phosphorylated intermediates of sugar metabolism may play a role as signalling molecules to promote bud outgrowth. Supporting this idea, we observed a significant increase of all phosphorylated monosaccharides (Gal1P, Glc6P, Glc1P, G1,6P, Fru6P, Fru1,6P, and Man6P) as well as Tre6P in September (Fig. 7). In general, most glycolytic intermediates showed a similar profile (Fig. 8), being statistically higher either in January, September, or both months, which is consistent with the greater respiration rates we observed in these months.

The TCA cycle intermediates displayed a greater concentration in January than the rest of the year, except for citrate which had lower levels during the summer period (Fig. 8). This suggests that the TCA cycle is most active in summer but remains less active during the rest of the year. Malate displayed a different profile, showing the lowest levels in February and April, when the depth of dormancy is greater.

Sorbitol and mannitol showed a similar pattern to trehalose, progressively increasing from April/May (Fig. 7). Another group of molecules showing a similar pattern was galactinol, raffinose, and stachyose (Fig. 7). Galactinol is a precursor of raffinose, which in turn is the precursor of stachyose, so the similar pattern is not surprising; however, it indicates a flow of these sugars all the way through to stachyose or more complex oligosaccharides (e.g. verbascose) that we did not measure. In addition, myo-inositol, a precursor of galactinol, showed an opposite pattern to galactinol (Fig. 7), suggesting a possible use of myo-inositol through the year to produce galactinol until July/August, when galactinol levels are the highest and myo-inositol levels the lowest, and then galactinol is catabolized and its biosynthesis reduced, resulting in an increase of myo-inositol prior to bud burst (Fig. 7). Similarly to myo-inositol, shikimate and tartrate, two abundant metabolites in the buds, progressively declined from their peak abundance in January (Fig. 7). No significant changes in concentrations were observed in melibiose and 1-kestose throughout the experimental period (Fig. 7). UDP-galactose and UDP-glucose shared a common profile, with concentration peaks in January and September, whereas ADP-Glc declined continuously from its peak in January (Fig. 7). ADP-Glc is essential for starch biosynthesis. During late autumn and winter, the absence of leaves does not allow the plant to produce sugars for the build-up of starch reserves, so it is not surprising that the levels of ADP-Glc decrease. On the other hand, UDP-Glc levels significantly increased prior to the bud burst period. Additionally, UDP-Glc is known to be produced from sucrose degradation in sink tissues to build up cell wall synthesis (Wai et al., 2017). UDP-Glc has also been suggested to play a signalling role during development (Janse van Rensburg and Van den Ende, 2018).

Finally, we performed a Pearson correlation analysis between all metabolites and dormancy depth (Supplementary Table S1). The 10 metabolites most strongly correlated with BB50 are summarized in Table 1. All of them showed negative correlations, with fructose exhibiting the strongest inverse association with dormancy depth. Notably, several sugar-phosphates were among the top correlates, indicating that their concentrations are reduced during the peak of dormancy.

Table 1.

Top 10 metabolites most strongly correlated with dormancy depth (BB50)

Metabolite	Pearson coefficient	P-value	Absolute Pearson coefficient
Fructose	−0.527	1.14E-03	0.53
Malate	−0.519	1.40E-03	0.52
Citrate	−0.504	2.01E-03	0.50
Trehalose	−0.493	2.60E-03	0.49
Fru6P	−0.482	3.34E-03	0.48
Glc6P	−0.464	5.03E-03	0.46
Man6P	−0.461	5.28E-03	0.46
Mannitol	−0.46	5.39E-03	0.46
3-PGA	−0.449	6.84E-03	0.45
Suc6P	−0.445	7.40E-03	0.45

Pearson correlation coefficients, P-values for the correlation, and absolute Pearson correlation coefficients are presented. Negative values indicate inverse correlation. For the correlation for the full list of metabolites, see Supplementary Table S1.

Discussion

The physiological and biochemical data delineate three periods of quiescence in grapevine buds

Here we analysed physiological features of V. vinifera cv. Cabernet Sauvignon buds over a 7 month period of quiescence, from the onset of dormancy to immediately before bud burst. Late summer (January to February) was characterized by the establishment of the peak of dormancy. At this stage, the basal buds are lignified but the internal process of dormancy was only beginning to become established, as evident in the increase in time to bud burst over this period. The literature commonly refers to this period as the transition between para-dormancy and endo-dormancy. This peak of dormancy was characterized by an apoplastic isolation of the bud from the cane, a reduction of the moisture content of the bud, a decrease in the proportion of cells in the S phase of the cell cycle, and a reduction of the respiration rate (CO₂ release and O₂ consumption). High levels of myo-inositol, shikimate, tartrate, ADP-Glc, and some TCA cycle intermediates were signature features of this first period of quiescence (Figs 5, 7, 8). Conversely, the levels of citrate were the lowest during this initial period of quiescence (Fig. 8).

In the second major period of quiescence, the depth of dormancy declined in a biphasic manner; it rapidly declined between the end of summer and mid-autumn, followed by a more moderate decline towards winter. If a BB50 threshold of >30 d is used to define endo-dormancy, our results indicate that buds entered the endo-dormant state during late summer and ceased to be endo-dormant when BB50 fell below this threshold in mid-autumn, consistent with the expected chilling-mediated dormancy release. Importantly, our interpretation does not imply that photoperiod drives the decrease in BB50. Rather, short-day photoperiods (already noticeably shorter than in summer, Fig. 1D) may have contributed to the onset or strengthening of endo-dormancy, as previously reported for grapevine (Cragin et al., 2017; threshold of ∼12.4 h daylength for the para-dormancy–endo-dormancy transition) and other woody species such as hybrid aspen, where short days increase ABA levels and expression of ABA receptors, leading to the induction of dormancy, the closure of plasmodesmata, and prevention of growth-promoting signalling from reaching the meristem (Tylewicz et al., 2018). In grapevine buds, ABA treatments were able to induce the expression of starch biosynthetic genes (Rubio et al., 2019), suggesting that ABA may induce starch accumulation. Here, respiration, internal pO₂, and moisture content of buds declined progressively from summer to a point of inflection in early winter. These patterns are consistent with an earlier study in cv. Merlot in the same vineyard (Velappan et al., 2022b). The subsequent decline in BB50, however, occurred when chilling began to increase and rainfall became more frequent, supporting the classic view that cold exposure reduces ABA and promotes dormancy release. During this second period of quiescence, most metabolites remained unvaried; however, high levels of raffinose were a clear metabolic signature of this period, as were a steady increase in stachyose, sucrose, Suc6P, galactinol, and trehalose (Fig. 7). In Thompson seedless grapevines grown in Chile, sucrose accumulations happen to be a bit earlier (late autumn), but in general a similar trend is found, and its accumulation was linked to the repression of invertase genes and increase of sucrose transporter genes induced by short-day photoperiods (Noriega et al., 2022).

The apoplastic pore size subtending the buds, which was determined to exclude molecules >2.1 nm (Signorelli et al., 2020), was restricted throughout the study period, even when buds had begun to swell in spring. This is consistent with the observations from Xie et al. (2018), who showed that Chardonnay buds only establish a vascular connection between the bud and cane after bud swelling. The most pronounced shift in cell division was the increased proportion of G₂ cells in late winter, relatively coincident with bud moisture content. This final transition accompanied a considerable increase in the ratio of G₂/G₁ cells, which is a common proxy for the mitotic index (Velappan et al., 2022a). We observed peaks in metabolic activity in January and September. However, the G₂/G₁ ratio suggests that cells in January are probably still growing but not dividing (low G₂/G₁ ratio, Fig. 4), while in September cell division has been initiated (high G₂/G₁ ratio, Fig. 4). In addition, during this third phase, the depth of dormancy was reduced to the minimum, the moisture of the bud increased, and so did the respiration rate.

The internal pO₂ fluctuated during these three periods, being reasonably uniform through the bud profile in the second period (February to June), whereas in the first and latest periods, summer (January) and early spring (September), respectively, the level of oxygen declined to <20 µM in the core of the bud (Fig. 2C). This could be attributed to a higher metabolic activity of the bud in those two months, which we confirmed by respiration measurements (Fig. 3B, C), which increases O₂ demand and leads to a decrease in internal O₂. Based on our previous work suggesting that internal tissues are not a significant barrier to O₂ diffusion (Signorelli et al., 2020), we consider respiratory demand the primary driver of this hypoxia. On the other hand, the greater oxygenation of the bud in the outer area in September could be linked to the fact that the outer scales and trichomes become looser prior to bud burst, allowing ready influx of O₂. In terms of metabolites, towards the end of this phase peak levels of trehalose, 3-phosphoglycerate (3-PGA), UDP-Gal, UDP-Glc, and many sugar phosphates such as Tre6P, Glc6P, Glc1P, G1,6BP, Fru6P, and Man6P were found. All in all, these data broaden our understanding of bud quiescence or the so-called endo-, para-, and eco-dormancy periods from a physiological and biochemical view.

The dynamics of metabolites in grapevine buds through the season

In annual herbaceous plants, such as Arabidopsis and sorghum, the role of sugars and some sugar-phosphates in dormancy and bud outgrowth control is well established (Barbier et al., 2015a, b; Fichtner et al., 2017). In general, low sugar content has been demonstrated to inhibit bud outgrowth (Mason et al., 2014; Kebrom and Mullet, 2015). However, little is known about the endogenous levels of sugars within grapevine buds and their dynamics across the season. Despite the buds being quiescent, the concentrations of sugars, sugar-phosphates, TCA cycle intermediates, and other metabolites varied significantly during that period. Based on our PCAs, we could separate our samples into up to four subgroups (Fig. 6), January, February, April–August, and September, so we defined these groups as four different metabolic phases to organize the discussion of the metabolite data.

Several monosaccharides, sugar-phosphates, and most TCA cycle intermediates decrease during the onset of the quiescent period in grapevine buds

The first metabolic phase, in early summer, was characterized by the highest concentrations for ADP-Glc, tartrate, most TCA cycle intermediates, pyruvate, myo-inositol, and shikimate (Figs 7, 8). Maltose, glucose, fructose, and galactose, as well as their phosphorylated forms, also had high values during this phase (Fig. 7). Physiologically, this phase represented the onset of dormancy (Fig. 10). The decrease of hexoses, with high levels of pyruvate and TCA cycle intermediates, together with the higher respiration rate during January points to metabolism of the buds being highly active in early summer. It is worth noting that the second decrease in sucrose and glucose (August to September) also correlates with higher levels of respiration, with sharper decreases in glucose concentration when the respiration rate is higher (i.e. January–February versus August–September, Figs 3B, C, 7). This observation suggests that following glucose dynamics can be indicative of higher respiration in the buds.

Fig. 10.

Schematic representation of metabolite profiles during the season and the correlation of their peak concentrations with the seasons and the quiescent state. The different metabolic phases were defined based on the PCA for the metabolomic data. 3-PGA, 3-phosphoglycerate; Fru, fructose; Gal, galactose; Gal-ol, galactinol; Glc, glucose; MI, myo-inositol; Pyr, pyruvate; Shk, shikimate; Suc, sucrose; Sug-P, sugar-phosphates; TCA, tricarboxylic acid cycle intermediates; Tre, trehalose.

Tartrate, which is known to reach high levels in grape berries, also reached high concentrations in the bud during summer. Radiolabelled tartrate is recovered as CO₂ released by berries, indicating that tartrate is catabolized in grapevines (Hrazdina et al., 1984). Thus, tartrate is probably being catabolized as a source of carbon and energy to support metabolic activity in the developing bud. Shikimate is a precursor for phenylpropanoids (e.g. flavonoids, isoflavonoids, coumarins, aurones, stilbenes, catechin, and lignin), many of which have protective functions against abiotic stress, deterrence of insect pests, and pathogen defence. The very high initial shikimate content and decrease over the seasons (Fig. 7) suggests a substantial investment in defence compounds to protect the bud throughout the rest period. Similarly, chloroplasts within the leaf primordium are likely to be accumulating starch as a reserve for the rest of the season, as suggested by the high levels of ADP-Glc (Fig. 8). Starch is known to have high concentrations in dormant buds, which rapidly decrease during dormancy release and the bud burst period (Rubio et al., 2016; Meitha et al., 2018). Likewise, myo-inositol is needed for the biosynthesis of complex sugars, such as raffinose, which were accumulated in the following months. Myo-inositol and its derivatives are also considered crucial compounds for development and signalling in plants (Valluru and Van den Ende, 2011; Roitman and Eshel, 2024); therefore, a signalling role in the transition between para-dormancy and endo-dormancy or during bud burst should not be ruled out.

The peak of dormancy correlated with the lowest levels of sugars and sugar-phosphates in grapevine buds

The second metabolic phase (February or summer) coincided with the peak of dormancy (Fig. 1) and represented a period in which almost all metabolites remained unvaried, except for citrate and raffinose which were the only two metabolites that were increasing (Figs 7, 8, 10). Thus, both the metabolite data and the physiological dormancy suggested the deepest rest during this month. The levels of many sugars, such as glucose, fructose, trehalose, and sucrose, were at the minimum during this second phase. In concordance, gene ontologies representative of the transcriptional transition from summer to autumn indicated a strong signature of response to metabolic stimuli, including starvation and nutrient levels (Velappan et al., 2022b).

Raffinose and citrate accumulate at the begin of the period of cold, while several sugars and sugar alcohols accumulate towards the end of this period in grapevine buds

The third metabolic phase was the longest one, from late summer to late winter, and was characterized by the peak of raffinose and citrate in autumn (May) and the subsequent accumulation of many metabolites whose concentrations peaked in late winter (August) (Fig. 10). The build-up of raffinose coincided with the period of physiological dormancy or endo-dormancy (Fig. 10), reaching its peak in late autumn. After that period, when the daylength is the shortest and rainfall increases, raffinose levels start to decrease again (Figs 1D, 7). Raffinose accumulation has been documented as a response to cold (Tarancón et al., 2017) and it has been linked to freezing tolerance in grapevine (De Rosa et al., 2022). In seeds, raffinose has been linked to seed longevity (Salvi et al., 2022). Therefore, raffinose could be accumulating in cv. Cabernet Sauvignon buds to protect the buds during the winter period. In cv. Cabernet Sauvignon leaves, a few hours of cold exposure produce a rapid decrease in raffinose concentration (Jiang et al., 2016). We observed a continuous decrease from winter, suggesting that also in the cv. Cabernet Sauvignon buds raffinose is catabolized during the cooler period. In canes of the grapevine cv. Vinhão, a gene involved in raffinose biosynthesis (VviRafS5) was shown to be induced upon cold exposure (Noronha et al., 2022). This could mean either that this cultivar responds to cold by accumulating raffinose or that raffinose is produced in the cane to feed the bud during the dormancy period where it is catabolized. In future research, it would be good to know in one cultivar the levels of raffinose in the cane and bud at the same time and the response of the transcripts involved in its metabolism. The response of raffinose metabolism was shown to be genotype dependent, as in ‘Beta’ (Vitis riparia×V. labrusca) cold-induced raffinose levels (Jiang et al., 2016). In a transcriptomic study, we observed that cv. Merlot buds displayed a strong decline in the expression of homologues of RAFFINOSE SYNTHASE in the transition from summer to autumn (Velappan et al., 2022b), suggesting either that raffinose biosynthesis is differentially regulated in buds of different cultivars or that post-translational regulation may influence synthesis of RFOs. As recently emphasized by North and Kovaleski (2024), integrating differential thermal analysis (DTA) of buds is crucial to disentangle how carbohydrate metabolism contributes to cold acclimation and hardiness. Thus, some of the sugar dynamics reported here may, at least in part, reflect underlying changes in bud hardiness, an aspect that will be important to address in future studies.

In winter (June–August), we found a higher content of sucrose, Suc6P, galactinol, stachyose, mannitol, and trehalose (Fig. 7), and this was immediately followed by an increase of glucose, galactose, fructose, and malate in late winter (Figs 7, 8, 10). Soluble sugars such as glucose, fructose, and sucrose are considered to be induced during the period of quiescence thanks to the activation of amylase activity by chilling which degrades starch reserves (Monteiro et al., 2022). Invertase and SWEET sugar transporters have been shown to be activated to increase sugar content in buds and promote bud outgrowth (Girault et al., 2010; Rabot et al., 2012; Kebrom and Mullet, 2016), and the induction of these enzymes was linked to light perception (Kebrom and Mullet, 2016). We observed that the accumulation of these sugars coincided with the increase in daylength (Figs 1D, 7). Therefore, it is possible that also in grapevine buds photoperiod contributes to regulating the expression of enzymes important to augment sugar content within the bud.

The concomitant accumulation of molecules such as stachyose and sucrose, accompanied by the decrease of raffinose, suggest that during this phase increasing sucrose provides greater substrate availability for raffinose synthesis, although raffinose itself may not accumulate because it is mostly converted into stachyose or RFOs with a higher degree of polymerization. Sorbitol, mannitol, and trehalose showed a similar pattern, progressively rising from April/May (Fig. 6). These sugars are known to be compatible osmolytes with kosmotropic activity, and trehalose has a well-known role as an anti-freezing molecule (Garg et al., 2002), so it is possible that these molecules are accumulated through the year to protect the bud from freezing in winter. In addition, the buds showed the lowest water content in autumn, so their accumulation can be also a response to osmotic stress.

Sucrose can promote growth both as a source of energy and as a signalling molecule (Barbier et al., 2015a, b). Since we did not observe an activation of metabolic activity (Fig. 4) and the products of its catabolism (glucose and fructose) until the end of this third phase, we conclude that sucrose is accumulated but not catabolized until the end of this phase when energy is required for growth resumption.

The accumulation of high levels of glucose, sucrose, and Suc6P prior to bud burst may be consistent with their functions in signalling, as indicated in herbaceous species; for example, in activating genes essential for growth resumption via the bZIP11 transcription factor (Ma et al., 2011) and regulation of the SUCROSE-NON-FERMENTING1-RELATED PROTEIN KINASE 1 (SnRK1; Tarancón et al., 2017). In Arabidopsis root hairs, glucose was shown to induce TOR activity which can phosphorylate E2Fa to promote root cell proliferation (Xiong et al., 2013). Thus, the increase of glucose in the weeks prior to bud burst could be playing a similar role in grapevine buds. Moreover, the increase of Tre6P prior to the bud burst period would inhibit SnRK1 (Baena-González and Lunn, 2020), reducing the inactivation of TOR through phosphorylation mediated by SnRK1, and further promoting the induction of metabolic activity by TOR. In this sense, the concomitant increase of Tre6P and decrease of sucrose that we observe prior to bud burst aligns with the current models suggesting that Tre6P can result in reduced levels of sucrose, by either altering sucrose allocation or reducing starch degradation (Göbel and Fichtner, 2023), and play a role in promoting bud outgrowth, models that mainly originated from model plants such as Arabidopsis or pea.

Sugar-phosphates accumulate prior to bud burst in grapevine buds

The fourth and last metabolic phase was defined by the highest concentration of fructose, trehalose, most of the measured sugar-phosphates, 3-PGA, UDP-Gal, and UDP-Glc (Figs 7, 8). At this stage, it is likely that sucrose and glucose are being respired to fuel the bud outgrowth through glycolysis, as we observed an increase of most glycolytic intermediates studied here [Glc6P, Fru6P, fructose-1,6-bisphosphate (FBP), 3-PGA, and phosphoenolpyruvate (PEP)]. This rise in glycolytic intermediates accompanies a rise in respiratory CO₂ production (Fig. 4), which indicates that there is increased flux through the TCA cycle and, by inference, increased glycolytic flux to feed carbon from sugars into the TCA cycle. The lower [Glc]/[Fru] ratio observed towards the bud burst period may suggest a preferential use of glucose, either to fuel metabolic activity or to support cell wall biosynthesis, as indicated by the concurrent increase in phosphorylated glucose (Fig. 8) and UDP-glucose levels (Fig. 7).

In an explant experiment, we showed that light down-regulates the expression of one of the TREHALOSE-6-PHOSPHATE PHOSPHATASE (TPP) genes (VIT_01s0011g05960, Vitvi01g00509 in v4.3) in explant grapevine buds (Signorelli et al., 2018). TPP converts Tre6P into trehalose. Here, we observed that the accumulation of trehalose levels stops with the increase of daylength (from August to September) while an increase in UDP-Glc occurs, suggesting that the down-regulation of TPP mediated by light also occurs in the field. The decrease in trehalose accumulation observed from August to September might also implicate the production of Tre6P, which we observed to reach the greatest levels in September. Tre6P is considered to play a signalling role that promotes growth of sink organs when sucrose supplies are abundant (Figueroa and Lunn, 2016), in part through the inhibition of SnRK1 to promote growth (Schluepmann et al., 2003; Zhang et al., 2009; Delatte et al., 2011). In another transcriptome study of cv. Merlot, we did not observe major changes in genes related to trehalose metabolism until the transition from winter to spring (Velappan et al., 2022b).

Possible metabolic markers to differentiate quiescent states in grapevine buds

We looked for metabolic signatures for the three different period of quiescence (i.e. quiescence prior to dormancy, or para-dormancy; physiological dormancy, or endo-dormancy; and quiescence after dormancy, or eco-dormancy) in grapevine buds of cv. Cabernet Sauvignon. We observed that levels of shikimate, tartaric acid, fumarate, succinate, aconitate, trans-aconitate, and 2-oxoglutarate were high only during the quiescence period prior to dormancy, whereas sucrose, trehalose, galactinol, stachyose, and mannitol were exclusively high during the period of eco-dormancy. Low levels of hexoses (glucose, fructose, and galactose) and an increase in raffinose concentrations were features of the endo-dormancy period. If this is extrapolated to other grapevine cultivars, the sugar profiles of buds could be used as a proxy to differentiate the period of dormancy (endo-dormancy) and quiescent states (para- and eco-dormancy). Such readily measured markers for the different dormancy states could be of great practical value to researchers studying woody perennials.

Conclusion

Using physiological and biochemical markers, we were able to differentiate distinct phases during the annual cycle of grapevine bud growth, dormancy, and bud burst, something not possible using a phenological approach. At the cellular level, we observed that the decrease of dormancy produced in late winter correlates with an increase in the proportion of cells in the G₂ phase of the cell cycle and a decrease of cells in the G₁ phase. At the molecular level, we observed that (i) the para-dormancy period is accompanied by greater levels of ADP-Glc, shikimate, tartrate, myo-inositol, pyruvate, and many TCA cycle intermediates; (ii) the period of endo-dormancy is characterized by the lowest levels of hexoses, their phosphorylated forms, and glycolytic intermediates; and (iii) the eco-dormancy period is characterized by greater levels of disaccharides such as sucrose (and Suc6P) and trehalose, along with monosaccharides, such as glucose and galactose, and precursors and members of the RFOs (e.g. galactinol and stachyose), plus the dicarboxylic acid malate. Thus, we conclude that the contents of different sugars in the grapevine buds have the potential to be used as markers to differentiate dormancy and quiescent states. It is worth noting that this study was limited to bulk tissue analysis. Microdissection of the shoot apical meristem (SAM) might enable greater resolution for metabolite profiling and marker-based approaches. Considering the whole process of quiescence, we observed that fructose is the metabolite inversely correlated the most with dormancy depth. Finally, our data provide suggestive evidence that the role of Tre6P in promoting bud outgrowth can be a conserved mechanism in woody perennial plants.

Supplementary data

The following supplementary data are available at JXB online.

Table S1. The correlation between metabolite content and the dormancy depth (BB50).

Author contributions

SS and MJC: conceptualization, project administration, supervision, and writing—original draft; SS, DH, YV, CC, and JAC: formal analysis; SS, JEL, and MJC: funding acquisition: SS, YV, DH, RF, and JAC: investigation; SS, YV, RF, JAC, JEL, and MJC: methodology; MJC: resources: SS and CC: visualization. All authors contributed to writing—review and editing.

Funding

SS is supported by the CSIC I+D groups program (Uruguay), group name: ‘Food and Plant Biology’ and group number ‘883431’. MC acknowledges funding support from the Australian Research Council (DP150103211; FT180100409). RF and JEL were financially supported by the Max Planck Society.

Data availability

The metabolite data that support the findings of this study are openly available in the Dryad Digital Depository at doi: 10.5061/dryad.ghx3ffc2w (Signorelli et al., 2026).