Grape berry samples were harvested from field-grown “Touriga Nacional” cultivar grapevines (Vitis vinifera L.) at 12 years old and 10–12 buds per vine were pruned; grapevines were grafted onto 1103P rootstock, conducted in a double Guyot system, from a 21 ha vineyard at “Escola Profissional do Rodo”, in the Baixo Corgo Sub-Region from the Douro Demarcated Region (Denomination of Origin Douro/Porto), located at Peso da Régua, Portugal (41°10'35.80''N; 7°47'48.00''W). The canopy height and thickness were 1.1–1.2 m and 0.5–0.6 m, respectively. The plants were managed without irrigation and grown using standard cultural decisions as applied by commercial farmers.

The climate is typically Mediterranean, with mild rainy winters and long, hot and dry summers (Jones and Alves, 2012). Total annual rainfall is 1112 mm, of which 14.1 % is from June to September. The warmest months are July and August, and the coldest are January and December, with mean daily temperatures of 20.7 °C and 6.1 °C, respectively (data for the period 1991–2021, according to the database of Climate-Data.org).

Mature leaves of 27 plants distributed over three vineyard rows from each of three 0.3 ha plots randomly located throughout the vineyard were manually sprayed (carefully, so the berry clusters were not sprayed) with an aqueous solution containing 2 mM sorbitol and 2 mM mannitol (with 0.05 % Tween-20 as surfactant to improve adhesion). This concentration was chosen because it is not sufficient to promote polyol-induced osmotic stress in leaves and is in the same order of magnitude as the K _m of the polyol transporter VvPLT1/VvPMT5 (also present in leaves) for mannitol (K _m of 5.4) and sorbitol (K _m of 9.5), respectively (Conde et al. 2015). The first application was performed in July 2019, at the late green berry phase and right before véraison (EL-33). A second application was performed during the véraison (EL-35) stage (Coombe, 1995). In each of the three other plots of 0.3 ha, the leaves of 27 plants from three vineyard rows were exogenously treated with an aqueous solution of just 0.05 % Tween-20 serving as the control condition. All grapevine rows were side-by-side (guaranteeing similar edaphoclimatic conditions) on a hill with an N–S orientation.

Leaf gas exchangewas measured with an infrared gas analyser (LC pro+, ADC Bioscientific Ltd., UK) operating in the open mode. Measurements were carried out in the morning (9:30–11:30). The photosynthetic photon flux density (PPFD) incident on the leaves was always higher than 1500 μmol m^–2 s^–1, which is above photosynthesis saturation in the considered species (Flexas et al., 2002). Net photosynthetic rate (A), stomatal conductance (g_s), transpiration rate (E) and intercellular CO₂ concentration/ambient CO₂ ratio (C_i/C_a) were calculated using the equations of von Caemmerer and Farquhar (1981). The A/g_s ratio was used as a proxy for intrinsic water use efficiency, according to Iacono et al. (1998).

Grape berry sampling was performed at two different stages: véraison—EL-35 (after the second application) and mature stage—EL-38. Phenological parameters, such as average berry weight, pH, water content and total soluble sugars (°Brix) were assessed. When sampling véraison berries, a representative mix of pigmented and non-pigmented berries was obtained. This careful procedure (as in Conde et al. (2016)) was followed both in the cluster and for different clusters of the plant, with half-pigmented and half-non-pigmented berries collected from each condition. Berry ripening rate was similar between conditions with no apparent phenological displacement. No difference in véraison pigmented:non-pigmented berries ratio between treatments was evident. For each ripening phase and each experimental condition, three sets of berries (each ranging from 81 to 108 berries in total), one for each of the three experimental plots and considering all precautions described above, were obtained. Each pool (from each plot) represented a biological replicate. Approximately half of each set of berries was ground to a fine powder under liquid nitrogen refrigeration and stored at –80 °C. The berry skin of the remaining grape berries was gently separated from the mesocarp and ground to a fine powder using a mortar and a pestle under liquid nitrogen and stored at –80 °C. When obtaining véraison exocarps, a representative mix of pigmented and non-pigmented berries was obtained from each set, with half of pigmented and half of non-pigmented berries used for exocarp separation.

Ground berry tissues of whole berry and exocarps were thoroughly mixed with 500 µL of methanol 100 % in an approximately 1:1 powder:methanol ratio and incubated with constant agitation at 4 °C overnight. Samples were centrifuged for 20 min at 18,000 × g, and the supernatants were collected. The supernatants were left in a vacuum drier at 4 °C for 18 h, and the obtained pellet was resuspended in 500 µL Tris-buffered saline 1×. The quantification of ABA was assayed with the Phytodetek Immunoassay Kit for ABA (Agdia Inc., Elkhart, IN, USA) according to the manufacturer’s protocol.

The extraction and quantification of primary metabolites were performed as described by Silva et al. (2017) and Noronha et al. (2021). The liquid extraction of primary solutes was carried out using 100 mg of lyophilised berries samples with 1 mL of ultrapure H₂O and 1 mL of chloroform. This mixture was vortexed for 5 minutes and incubated at 50 °C for 30 minutes with continuous agitation. After incubation, the samples were centrifuged at 15,000× g for 15 min at 4 °C, and the aqueous supernatant (upper layer) was collected. The supernatant was filtered using a PTFE 0.2 µm filter and quantified by HPLC-RI using a Rezex RCM-Monosaccharide Ca²⁺ (8 %) column (Phenomenex, California, United States of America) using a flow rate of 0.2 mL min^-1 at 40 °C and water as the mobile phase. The sorbitol concentration of each sample was determined by comparison of the peak area and retention time with standard curves. Mannitol concentration was also tentatively determined but was not detected.

For absolute quantification and characterisation of piceid and cyanidin-3-glucoside, methanolic extracts of grape berries and berries skins were prepared with freeze-dried samples, using a proportion of 1.5 mL of 80 % (v/v) methanol per 25 mg of sample. The methanolic extracts were filtered using 0.22 µm CLARIFY-PTFE 13 mm Syringe filters and prepared for quantification. The samples were separated using a flow rate of 0.8 mL m^-1 using water:formic acid (99:1) and methanol as mobile phase using an RP18 (25∙0.4 mm, particle size of 5 µm, Merck, Darmstadt, Germany) column connected to an HPLC (Hitachi, Ibaraki, Japan) coupled with a diode array detector (DAD). The concentrations of piceid and cyanidin-3-glucoside in each sample were determined by comparison of the area and retention time with standard curves obtained using 5 points/standard solutions derived from > 95 % pure Piceid (Sigma-Aldrich, Missouri, United States of America) and > 95 % pure Cyanidin-3-glucoside (Sigma-Aldrich, Missouri, United States of America) standards, respectively. For piceid quantification, a wavelength of 320 nm was used, whereas 520 nm was used for cyanidin-3-glucoside quantification.

Metabolites of processed grape berries and exocarps were extracted from freeze-dried samples using a ratio of 1.5 mL of 80 % (v/v) methanol per 25 mg of dry weight. Samples were macerated and sonicated for 30 min and left at 4 °C in the dark overnight. The samples were centrifuged at 18,000 × g for 10 min, and the supernatants were collected. Ultra-performance liquid chromatography coupled to mass spectrometry (UPLC-MS)-targeted metabolomic analysis was performed as optimised previously (Billet et al., 2018; Martins et al., 2021), using an ACQUITY UPLC system combined with a photodiode array detector and a Xevo TQD mass spectrometer (Waters, Milford, MA, United States) equipped with an electrospray ionisation source regulated by Masslynx 4.2 software. Analyte separation was achieved by using a Waters Acquity HSS T3 C18 column (150 × 2.1 mm, 1.8 μm) with a flow rate of 0.4 mL/min at 55 °C. The injection volume was 5 µL. The mobile phase consisted of solvent A (0.1 % formic acid in water) and solvent B (0.1 % formic acid in acetonitrile). Chromatographic separation was carried out using an 18-min linear gradient from 5 to 50 % (v/v) solvent B. MS detection was performed in both positive and negative modes. The capillary voltage was 3000 V and sample cone voltages were 30 and 50 V. The cone and desolvation gas flow rates were 60 and 800 L h⁻¹. Identification of analytes was based on retention times, m/z values and UV spectra in comparison with commercial standards. UPLC–MS analyses were achieved using the selected ion monitoring (SIM) mode of the targeted molecular ions. UPLC-MS analyses were carried out using a selected ion monitoring (SIM) mode of the targeted molecular ions. SIM chromatograms were mathematically integrated using the subprogram QuanLynx 4.2. Peak integration was made using the ApexTrack algorithm with a mass window of 0.1 Da and a relative retention time window of 1 min, followed by Savitzky–Golay smoothing (iteration = 1 and width = 1). To evaluate the robustness of measurements and analytical variability, a pool of samples was prepared from all experimental conditions and maturation phases to obtain a quality control sample (QC) and the samples were randomly injected independently from the treatment or developmental phase. Three QC samples were injected to equilibrate the system before each analytical sample set, and one QC sample was injected every eight samples to check for potential analytical drift. Relative quantification was determined for L-proline (m1), L-iso-leucine (m2), L-leucine (m3), L-phenylalanine (m4), L-tyrosine (m5), L-tryptophane (m6), cyanidin-3-O-glucoside (m7), peonidin-3-O-glucoside (m8), delphinidin-3-glucoside (m9), petunidin-3-O-glucoside (m10), cyanidin-3-O-(6-O-acetyl)-glucoside (m11), malvidin-3-O-glucoside (m12), petunidin-3-O-(6-O-acetyl)-glucoside (m13), malvidin-3-O-(6-O-acetyl)-glucoside (m14), petunidin-3-O-(6-p-coumaroyl)-glucoside (m15), malvidin-3-O-(6-p-coumaroyl)-glucoside (m16), malvidin-3,5-O-diglucoside (m17), gallic acid (m18), citric acid (m19), E-resveratrol (m20), E-piceatannol (m21), catechin (m22), epicatechin (m23), coutaric acid (m24) caftaric acid (m25), fertaric acid (m26), piceid (m27), catechin gallate (m28), kampferol-3-O-glucoside (m29), pallidol (m30), E-ε-viniferin (m31), quercetin-3-O-glucoside (m32), quercetin-3-O-glucuronide (m33), myricetin-hexoside 1 (m34), myricetin-glucoside (m35), quercetin derivative (m36), procyanidin B1 (m37), procyanidin B2 (m38), procyanidin B3 (m39), procyanidin B4 (m40), kaempferol-3-O-rutinoside (m41), procyanidin gallate 1 (m42), procyanidin trimer 2 (m43), procyanidin gallate 1 (m44), procyanidin gallate 2 (m45) as in Billet et al. (2018) and Martins et al. (2021).

Total RNA was extracted from approximately 0.2 g of freshly ground samples following the protocol described by Reid et al. (2006). RNA was purified using the GeneJET Plant RNA Purification Kit (Thermo Fisher Scientific, Massachusetts, United States of America) followed by DNAse I treatment (Thermo Fisher Scientific, Massachusetts, United States of America). Using 1 µg of total RNA, cDNA was synthesised utilising the Xpert cDNA Synthesis Kit Mastermix (Grisp, Porto, Portugal). The purity and concentration of RNA were evaluated using NanoDrop.

The expression of several targeted genesin berries and exocarps at different developmental stages from control and polyol-treated vines was analysed by real-time qPCR. Real-time qPCR analyses were performed with Xpert Fast SYBR PCR Kit (Grisp, Porto, Portugal) using 1 µl of cDNA (diluted 1:10 in ultra-pure water) in a final reaction volume of 10 µl per well. Experiments were performed in biological triplicates using as reference genes VviACT1 (actin) and VviGAPDH (glyceraldehyde-3-phosphate dehydrogenase) as these genes were demonstrated to be highly stable and ideal for qPCR data normalisation in grapevine (Reid et al., 2006). Melting curves were performed at the end of each run to confirm the absence of unspecific or primer-dimer amplification. Gene-specific primers used in this work are listed in Supplementary Table 1. The relative expression data were normalised by the average expression of the reference genes (Pfaffl, 2001) and analysed using the CFX Manager Software 2.0 (Bio-Rad, California, United States of America).

Total protein extraction from grape berry and exocarp powder was performed as described by Conde et al. (2016)with some adjustments. Sample powder was mixed with extraction buffer in an approximately 1:1 powder:buffer ratio. Protein extraction buffer contained 50 mM Tris-HCl pH 8.9, 5 mM MgCl₂, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), 5 mM dithiothreitol (DTT) and 0.1 % (v/v) Triton X-100. The homogenates were centrifuged at 18,000 × g for 20 min, and the supernatants were collected. Supernatants were purified using an Amicon Ultracell-3K (Millipore, Massachusetts, United States of America) to remove contaminants, kept on ice and used for all enzymatic assays. Total protein concentrations were assessed using Bradford's method, which used bovine serum albumin as a standard (Bradford, 1976).

The enzymatic activity of UFGT was assessed as described by Lister et al. (1996) and adapted by Conde et al. (2016). The assay mixture contained 300 mM Tris-HCl buffer (pH 8.0), 1 mM UDP-glucose, 200 µL of protein extract and 5 mM of DTT in a final volume of 1 mL. The reaction was initiated by the addition of 1 mM quercetin as substrate, corresponding to a saturating concentration. The reaction mixture was incubated in the dark under gentle agitation for 30 min. Following incubation, dilutions were prepared using 100 µL of each assay mixture and 900 µL of Tris-HCl reaction buffer to evaluate the production of quercetin-3-glucoside at 350 nm (ε = 21,877 M⁻¹ cm⁻¹).

The enzyme activity of PAL was determined as described by Conde et al. (2016). The assay mixture contained 0.2 mL of protein crude extract and 3.6 mM NaCl in 50 mM Tris-HCl (pH 8.9) of reaction buffer in a final volume of 1 mL. Reactions were initiated by adding 25 mM of L-phenylalanine. The reaction mixture was incubated at 41 °C in the dark under gentle shaking for 30 min, and the formation of trans-cinnamic acid was assessed at 290 nm (ε = 17,400M⁻¹ cm⁻¹).

The enzyme activity of MTD and SDH was assessed as described by Conde et al. (2015). The assay mixtures contained 200 µL of protein extract, 300 mM Bis-Tris Propane (pH 9.0), 1 mM NAD⁺, and 200 mM D-mannitol or 200 mM sorbitol for MTD and SDH enzyme activity, respectively. The reduction of NAD⁺ to NADH was evaluated spectrophotometrically at 340 nm. D-mannitol oxidation and sorbitol oxidation were initiated by the addition of the polyol. The MTD and SDH enzymatic activities in the direction of fructose reduction were performed precisely like the mannitol oxidation assay but using 200 mM fructose to initiate the reaction and 1 mM NADH as a co-factor. The reaction mixture was incubated at 37 °C in the dark under gentle agitation.

The extraction of anthocyanins was carried out using 20 mg of lyophilised tissue from each treatment and 90 % methanol (v/v), as described in Conde et al. (2016). Suspensions were vortexed for 10 min and maintained overnight at 4 °C. The samples were centrifuged at 18,000 × g for 20 min, and the supernatants recovered. From each experimental condition, 200 µL of supernatant was mixed with 1.8 mL of 25 mM KCl (pH = 1.0). The absorbance of the samples was measured at 520 and 700 nm. Total anthocyanin concentration was assessed in relation to cyanidin-3-glucoside (C-3-G) equivalents, as expressed in the equation:

[Total anthocyanins ](mg L -1 )=  A 520 -  A 700  ×MW ×DF ×1000 ε ×1  

where MW is the molecular weight of C-3-G (449.2 g mol^-1), DF is the dilution factor, and ε corresponds to the molar extinction coefficient of C-3-G (26,900 M^-1 cm^-1). The concentration was subsequently presented per mg of dry weight (DW) of used exocarp tissue.

Quantification of flavonoids was carried out as described by Matejić et al. (2013) with some modifications. Assay mixtures contained 20 µL of standard quercetin solutions or methanolic extracts from the experimental conditions, 30 % ethanol, 0.2 % (w/v) of aluminium chloride and 20 mM of potassium acetate. After incubation at room temperature for 40 min in the dark, the absorbance of the reaction mixtures was assessed at 415 nm. Total flavonoid concentration in methanolic extracts was quantified using a quercetin hydrate calibration curve (5–100 µg mL^-1) and expressed as quercetin equivalents per gram of DW.

The concentration of total phenolics was quantified by the Folin–Ciocalteu method in exocarps from all experimental conditions. Total phenolics were extracted using a ratio of 1.5 mL of methanol absolute per 20 mg of lyophilised tissue from each experimental condition. The homogenates were vigorously agitated for 30 min and maintained overnight in the dark at 4 °C. Subsequently, homogenates were centrifuged at 18,000 × g for 20 min. 20 µL of each supernatant was added to 1.58 mL of deionised water and 100 µL of Folin reagent, vigorously agitated and incubated for 5 min in the dark before adding 300 µL of 2 M sodium carbonate. After 2 h of incubation in the dark, the absorbance was evaluated at 765 nm. Total phenolic concentrations were quantified using gallic acid equivalents (GAE) per gram of DW.

The results were statistically analysed using Student’s t-test resorting to GraphPad Prism 9 software (GraphPad Software Inc., USA). The results are shown as mean ± standard deviation of the mean (SD), and statistical comparisons were calculated with a 95 % confidence interval. Parametric tests were employed since all data sets have Gaussian distributions. For each test result, a P-value indicates the significance of each tested sample. This significance is indicated in the figures with P < 0.05 (*), P < 0.01 (**), P < 0.001(***) or P < 0.0001(****). All results were obtained from the analysis of three biological replicates.

As observed in Table 1, control and polyol-treated berries exhibited similar pH, AWC and °Brix during the véraison (EL-35) and mature stages (EL-38) of development. However, the ABW was higher in polyol-treated berries compared to control berries. Polyol-treated berries were approximately 40 % and 20 % heavier than control berries during véraison and mature stages, respectively.

Table

titre du tableau
		pH		ABW (g)		AWC (%)		⁰Brix
EL-35	Control	2.77 ± 0.19	ns	1.26 ± 0.19	*	81.6 ± 0.2	ns	13.4 ± 0.4	ns
	Polyols	2.77 ± 0.17		1.76 ± 0.23		82 ± 0.3		12.6 ± 0.3
EL-38	Control	3.76 ± 0.14	ns	1.57 ± 0.12	*	72.9 ± 0.3	ns	22.3 ± 0.2	ns
	Polyols	3.71 ± 0.17		1.87 ± 0.14		74 ± 0.3		21.9 ± 0.3

The treatment did not promote any statistically significant change in net photosynthetic rate (A), stomatal conductance (gs), transpiration rate (E), intercellular CO2 concentration/ambient CO2 ratio (Ci/Ca) and A/gs ratio (intrinsic water use efficiency) as described in Table 2.

Table

titre du tableau
	Transpiration rate (E) (mmol.m-2.s-1)Mean SD	Stomatal conductance (gs ) (mmol.m-2.s-1)Mean SD	Net photosynthetic rate (µmol.m-2.s-1)Mean SD	Intrinsic water use efficiency (A/gs ) (µmol.mol-1)Mean SD	Ratio of intercellular and atmospheric CO2 concentration (Ci/Ca) Mean SD
July 16th	Control	5.39	1.10	ns	188.26	48.81	ns	11.59	2.14	ns	62.96	7.36	ns	0.64	0.03	ns
	Polyols	5.71	1.38		205.75	52.82		11.71	2.26		58.45	9.04		0.66	0.04
July 30th	Control	4.02	1.39	ns	146.46	65.52	ns	10.45	4.02	ns	75.25	12.33	ns	0.63	0.04	ns
	Polyols	4.02	1.49		170.97	89.76		11.17	4.65		71.50	13.79		0.65	0.04

The concentration of ABA in the véraison stage (EL-35) was 226.97 pmoles g^-1 FW in control berries compared to 792.33 pmoles g^-1 FW of polyol-treated berries, which corresponded approximately to a 3.5-fold increase in polyol-treated berries. Similarly, mature (EL-38) treated berries exhibited a 6-fold increased ABA content compared to control berries, as shown in Figure 1A. Contrarily, in the berry exocarp, the application of a polyol solution did not produce any statistically significant alteration in the ABA content (Figure 1B) despite a visible but non-statistically difference at véraison.

Table

Polyols foliar application enhanced the responsiveness of the mechanisms involved in the phenylpropanoid pathway in the exocarps through the elicitation/activation of key intervenient enzymes, consequently allowing greater availability of precursor molecules for the biosynthesis of stilbenes and anthocyanins. This was the case of PAL, which catalyses the first step of the phenylpropanoid pathway, as denoted by a strong upregulation of VvPAL1 occurring in mature polyol-treated berry skins together with a significant increase of the biochemical activity of PAL measured in protein extracts and of numerous polyphenols quantified by UPLC-MS, particularly stilbenoids and anthocyanins.

The regulation of stilbenoid biosynthesis is modulated by biotic and abiotic factors (Rienth et al., 2021; Savoi et al., 2016). In this work, the foliar application of polyols stimulated the production of stilbene compounds in exocarps, especially at the mature stage (EL-38). In particular, an increased concentration of total stilbenoid DP1 and DP2 in treated exocarps was observed, in accordance with the significant increase of VviSTS1 transcripts. These results agree with the apparent increase of precursor phenylpropanoids synthesised in the phenylpropanoid pathway presented in Figure 3 but also with the increased content of ABA presented in Figure 1 (Miliordos et al., 2022). These data suggest that the application of polyols constitutes a strategy to produce an increased concentration of stilbenes in exocarps. The increased concentration of stilbenoid compounds in exocarps constitutes an important attribute, considering these molecules are particularly active in biotic stress responses by enhancing tolerance against grapevine powdery mildew, for instance. In this regard, these results pave the way for new studies to evaluate the protective effect of exogenous polyols against vine pathogen-induced diseases. Moreover, the nutraceutical and pharmacological properties of stilbenes are recognised and described in several studies (Reinisalo et al., 2015; Sirerol et al., 2016), reinforcing the importance of the present findings.

ABA has been described as an activator of anthocyanin biosynthesis genes and anthocyanin-related transcription factors (Badim et al., 2022; Gagné et al., 2011; Koyama et al., 2018; Sun et al., 2019), so it is possible that the observed changes in anthocyanin levels were mediated by ABA signalling upon sensing of exogenously applied polyols in leaves. The observed increased expression of VviUFGT1 paralleled with the enhanced enzyme activity of UFGT in exocarps, where the biosynthesis of anthocyanins predominantly occurs (Gouot et al., 2019; Walker et al., 2006). The expression of genes such as F'3'H, F'35'H, LDOX, DFR and UFGT, involved in the flavonoid pathway and anthocyanin biosynthesis, are normally upregulated in water-deficit, thus, ABA-triggering conditions (Castellarin et al., 2007a; Castellarin et al., 2007b; Deluc et al., 2009; Ferrandino and Lovisolo, 2014). Interestingly, while the quantities of flavan-3-ols were more than 3-fold lower in mature exocarps than in véraison ones (Figure 5), with non-detected levels in the specific case of catechin, the transcripts of VviLAR2 were similar, and those of VviANR and VviLAR1 were even higher, despite their downregulation by polyol application. This observation suggests that the synthesis of these compounds is strongly regulated at a post-transcriptional level.

The targeted metabolomics analysis performed in our previous study (Conde et al., 2015) showed that Tempranillo cv. grapevines under drought stress accumulate a higher-than-normal amount of polyols in the grape berry, including mannitol, sorbitol, galactinol and dulcitol. At the molecular level, the accumulation of polyols was found to be mediated by the stimulation of mannitol transport at the plasma membrane via VviPLT1 (VviPMT5) coordinated with repression of the intracellular oxidation of mannitol/sorbitol into fructose, mediated by mannitol/sorbitol dehydrogenases, while the reverse reaction of the enzymes (reduction of fructose) was stimulated. Thus, in stressed vines, the total amount of polyols accumulated in mature berries was approximately equivalent to a normal sucrose concentration in a mature berry, but this value could be even higher because many other (not quantified) polyols could account for this intracellular polyol pool under drought stress. Thus, in the present study, the observed activation of polyol intracellular oxidation (up-regulation of mannitol and sorbitol dehydrogenases; strong decrease in sorbitol content) in berries from mannitol-treated Touriga Nacional cv. plants follow an opposite trend and deserve further attention, particularly into the possibility of a protective role of polyols that may be, in fact, cultivar-dependent. It is also tempting to speculate that exogenous mannitol and sorbitol drive the metabolism of the intracellular pool of polyols to energy-yielding reactions or to fuel the biosynthesis of secondary metabolites, including stilbenoids and anthocyanins. Supporting this theory appears to be the fact that berries from polyol-treated plants were approximately 40 % heavier than control ones despite the absence of differences in their water content percentages.

The mechanisms underlying the apparent biostimulant effect of exogenous mannitol and sorbitol still need further research, but they share some similarities with the ones reported before for exogenously applied kaolin (Conde et al., 2016). The application of this partially light-reflecting clay induced a set of molecular and biochemical changes that together represented a significant stimulation of phenylpropanoid and flavonoid pathways in grapevine, including an increase in the concentration of anthocyanins in mature berries of plants that were also more protected against water deficit situations (Conde et al., 2016). Seaweed (Ascophyllum nodosum) extracts are also biostimulants of grapevine secondary metabolism and berry quality, making plants that were treated at the foliar level more able to cope with abiotic and biotic stresses (Frioni et al., 2019; Salvi et al., 2019). The berries from these plants (Sangiovese cv.) presented higher phenolic and anthocyanin concentrations at harvest, with several molecular mechanisms underlying the synthesis of these compounds upregulated. Chitosan application in grapevines also demonstrated similar effects in phenolic and anthocyanin concentrations and respective biosynthetic pathways, being, therefore, another example of a compound with biostimulant properties in this plant species (Singh et al., 2020). Considering the results, besides having a biostimulant effect, exogenous application of polyols might even be an interesting strategy to improve freezing tolerance in grapes, as it led to a significant increase in ABA concentration. This phytohormone has a pivotal role in bud dormancy and suffers an abrupt decrease in its concentration during bud break, which increasingly occurs earlier in the vegetative year as sudden climatic changes from freezing to warm conditions are more common (Zheng et al., 2015).

Mannitol has a recognised role in plant–pathogen interaction, and its concentration increases significantly during plant infection by biotrophic or necrotrophic fungi, together with an increased expression of genes involved in the mannitol–synthetic pathway (Calmes et al., 2013; Chaturvedi et al., 1996; Dulermo et al., 2009; Jobic et al., 2007; Voegele et al., 2005). For instance, stilbenes are key defensive metabolites involved in the resistance of grapevine to major fungal pathogens that include Botrytis cinerea (grey mould), Plasmopara viticola (downy mildew) and Erysiphe necator (powdery mildew). Grapevine varieties that are more resistant to biotic stress respond very rapidly to infection by producing high concentrations of stilbenes, including δ-viniferin and pterostilbene, at the sites of infection (Viret et al., 2018). Thus, exogenous mannitol may mimic the presence of mannitol-producing plant pathogens, triggering and eliciting a response that involves the synthesis of stilbenoids and other biotic stress-protecting secondary metabolites. Reinforcing this hypothesis, stilbene synthesis, including piceid concentration and VvSTS1 expression, was highly increased in berry exocarps from polyol-treated plants, displaying some of the most significant responses observed in this study.