Maximum stomatal conductance rather than stomatal sensitivity to drought differentiates the PIWI grapevine cultivar Souvignier gris from Muscaris and Donauriesling

Pioneering grapevines , also referred to as “PIWIs” (after the German expression “ pilzwiderstandsfähig ” = fungus resistant), represent an interesting alternative to classic cultivars in several wine regions. Their resistance to some fungal diseases places them among the potential solutions for sustainability, especially under organic production systems. However, little is known about their response to abiotic stressors, such as water stress under drought conditions. Here we studied the response of three PIWI cultivars (Donauriesling, Muscaris, Souvignier gris, and Riesling or Grüner Veltliner Vitis vinifera cultivars as comparison; all plants without fruit) to drought under semi-controlled conditions over two consecutive seasons. While in season 2020, we imposed a single dry-down and re-watering cycle inside a glasshouse, during season 2021, we subjected the vines to three cycles of dehydration and re-watering under field conditions with the aid of a rain shelter. Along the experiments, we monitored leaf gas exchange and water potential during soil dehydration and characterised leaf area development and some key leaf hydraulic traits. Despite different conditions in our two study seasons, we found connective significances in some key parameters. Under well-watered conditions, Souvignier gris had the highest rates of stomatal conductance, whereas Muscaris exhibited the most conservative water use behaviour. When under drought, Souvignier gris showed an apparent less tight stomatal control than the other PIWIs in coordination with a lower (more negative) osmotic potential and turgor loss point. Our results suggest that Muscaris (and Donauriesling showed similar behaviour) might be more suitable for non-irrigated conditions than Souvignier gris. However, further research on berry composition and open field is needed to upscale our results and address the water management of PIWI vineyards.


INTRODUCTION
Fungus-resistant grapevine cultivars, also named "PIWIs" (short form for German "pilzwiderstandsfähige Rebsorten" = fungusresistant grapevine cultivars), were developed by multiple backcrossing of interspecific hybrids (the cross-breeding of Vitis vinifera L. and North American and Asian Vitis species) with Vitis vinifera (Montaigne et al., 2016;Pedneault and Provost, 2016;Töpfer and Trapp, 2022). The advantage of these new genotypes is to interlink the resistance of most American species to fungal diseases (mainly Erysiphe necator, Plasmopara viticola) while retaining the high fruit quality characteristics for the vinification of Vitis vinifera cultivars (Hofmann et al., 2014;Fischer-Colbrie et al., 2015). In addition to the expediency of fungus resistance, there are economic advantages since the seasonal treatments of plant protection can be decreased and subsequently lead to lower annual doses of pesticides and fuel (Pedneault and Provost, 2016;Pertot et al., 2017). As winegrower's decisions towards sustainable management strongly depend on national and regional regulations (Chen et al., 2022), and the European Commission aims to halve the use of pesticides by 2030 according to the "Farm to Fork" strategy (European Commission, 2020), the demand for robust new cultivars with low susceptibility to pathogens is raising. However, while the screening of resistance against pathogens and pests is part of the breeding and selection program (Reynolds, 2015;Merdinoglu et al., 2018), there is limited information on the agronomic performance of PIWIs in different climates (Casanova-Gascón et al., 2019) and on their response to abiotic stresses (e.g., drought).
Extensive world viticultural regions are located in areas characterised by warm and dry summers, often managed in rainfed conditions, with periods of drought and heatwaves (Moriondo et al., 2013). Moreover, droughts are further forecasted to increase due to climate change (Schultz and Stoll, 2009;Cook et al., 2018), and irrigation may not be sustainable in several regions (Costa et al., 2016;Gambetta et al., 2020). Stomatal regulation is considered a proxy to describe how grapevines respond to drought stress (Medrano et al., 2002;Chaves et al., 2010). Since stomata are responsible for CO 2 intake, their regulation is of great importance in terms of photosynthesis and water use efficiency (Schultz and Stoll, 2009;Medrano et al., 2018). There are biochemical (mainly driven by the plant hormone abscisic acid) and hydraulic mechanisms involved in the stomatal reaction to drought stress which are hard to disentangle (Buckley, 2019;Gambetta et al., 2020). Although V. vinifera cultivars may respond differently to drought, some exhibiting a tighter or a looser stomatal control in response to water deficit (Tombesi et al., 2014;Hochberg et al., 2016;Dayer et al., 2020;Levin et al., 2019), it is not yet clear to what extent such differences result from genetical or environmental factors or the combination of both (Van Leeuwen et al., 2019;Gambetta et al., 2020). Furthermore, to the best of our knowledge, no study has assessed the stomatal regulation of grapevine PIWIs nor compared it against V. vinifera cultivars. In this study, we set out to investigate the response to the drought of three PIWI grape cultivars of commercial importance in Europe, focusing on their stomatal regulation dynamics during soil dehydration.
For the first experiment in 2020, a total of 64 plants (n = 16 per cultivars D, M, S, and R) were potted on the 18th April and randomly arranged within 6 rows inside the glasshouse. Plants were allowed to develop one single shoot and were grown without water limitations for 80 days (establishment period) until each vine reached a height of about 1.80 m; all lateral shoots were removed. The plants were watered daily by drip irrigation (two drippers per pot) until dripping was observed from the bottom of the pot (on average between 1.5-2 L per day). In season 2021, a total of 160 plants of the same PIWI genotypes and Grüner Veltliner as V. vinifera cultivar (n = 40 per cultivar D, M, S and G) were potted on the 8th May and arranged within three rows (nort-south orientation) under filed conditions in blocks of 10 plants per cultivar each. The three rows were covered by a rain shelter (transparent plastic roof of 4 m height and opened at the sides as similarly described in Herrera et al., 2015). The vines were grown for 86 days without water limitations by daily irrigating them (one dripper per pot) to field capacity until dripping was observed from the bottom of the pot (on average between 2-2.5 L per day); growth was not restricted, and laterals were not removed. Fertilisation was performed by adding 10 g of EntecVino® in each pot towards the end of May.

Environmental conditions
Air temperature and relative humidity were recorded hourly during study periods using a data logger (UT330B, Uni-trend Technology, Hong Kong) in the glasshouse in season 2020 and the local weather station of IWOB vineyard in 2021 (positioned 50 m apart from the study site); temperature and humidity data were used to calculate the vapour pressure deficit during the central hours of the day (between 10-14 h; VPD max ). The main climate parameters were followed over both years in both experimental set-ups: Records of climate data in season 2020 showed an average temperature of 19 °C during the experiment (between DOY 189 and 212) in the glasshouse, with maximum temperatures reaching 33 °C during the drought period and at the end of the experiment (Supplementary Figure 1A). The daily minimum temperature was relatively stable at about 13 °C. The vapour pressure deficit during the central hours of the day (VPD max ) was, on average, 2.04 kPa during the experiment, reaching peak values of 2.96 kPa, 2.36 kPa, and 2.45 kPa on DOY 192,DOY 197,and DOY 202,respectively. In 2021, the average temperature between DOY 214 and DOY 261 in the semicontrolled field was 17.6 °C, while increasing temperatures in the first days of the experiment. In the duration of the experiment, temperatures peaked at 29.2 °C on DOY 222 and 32.3 °C on DOY 227. The average maximum and minimum temperatures were 24 °C and 12 °C, respectively. The vapour pressure deficit during the central hours of the day (VPD max ) was, on average, 1.32 kPa, with peak values of 2.61 kPa on DOY 227 and 2.52 kPa on DOY 252 (Supplementary Figure 1B).

Irrigation treatments
In season 2020, on July 7th (day of the year DOY 189; after the 80 days of establishment period), two treatments were randomly established in half of the plants (i.e., eight vines per cultivar): (i) water-stressed "ws", where irrigation drippers were removed from the pots for 13 days; (ii) well-watered controls "ww", where daily irrigation was maintained to field capacity as before. After 13 days without irrigation (DOY 202), all vines of the ws treatment were manually re-watered to saturation, and thereafter daily irrigation was resumed to the same levels of control. Gas exchange after re-watering was monitored for the successive 10 days (until DOY 212) ( Figure 1A).
In season 2021, half of the plants were always maintained under well-watered conditions by daily irrigating the plants just as during the establishment period. The second half of the plants (i.e., 20 vines per cultivar) were exposed to three cycles of drought (water-stressed "ws"; water-stressed plants remained the same throughout all three cycles) by removing the irrigation drippers from the pots for 8 days; after each drought cycle, all vines were re-watered and maintained under "ww" conditions until the next drought cycle. The first ws cycle was imposed on DOY 214 (August 2nd; after 86 days of establishment period) and lasted until DOY 221. The second ws cycle started at DOY 242 until DOY 249, and the third ws cycle occurred from DOY 253 to DOY 260 ( Figure 1B).

Leaf area measurements
In both seasons, leaf area was measured following the protocol by Herrera et al. (2021), starting on DOY 189 (2020) and DOY 214 (2021), where the leaf length of all the leaves of four vines per cultivar was measured. Another measurement was taken of leaves from three vines per cultivar and treatment at the end of the experiments. In 2020, at the end of the experiment (DOY 212), 15 leaves per cultivar were collected, and the leaf length and area were determined by image analysis using the ImageJ software; the data was used to compute a regression between leaf length and single leaf area per each cultivar. The equation was then used to calculate the leaf area per plant at DOY 189 and 212. Since in season 2021 plants were grown without restrictions (laterals and small leaves included), leaf area was determined by using a leaf area meter (LI-3100 Area Meter, LI-COR. Inc. Lincoln. Nebraska USA); we matched both data sets (season 2020 and 2021) to compute an average regression per cultivar (Supplementary Table 1); at the end of the experiment in season 2021, 10 main leaves (fully expanded) per cultivar were collected and measured to compare the average single leaf area of both seasons.

Gas exchange and water potential
Stomatal conductance (g s ), transpiration (E), and leaf net assimilation rate (A N ) were measured regularly using an infrared gas analyser (LC-ProSD, ADC Bio Scientific, UK). The measurements were done between 09:30 and 11:30 h (local time) on three (in 2020) or four (in 2021) plants per cultivar and treatment (using a healthy and sun-exposed leaf from the main shoot). The instrument was set to a constant airflow of 200 μmol s -1 and saturating light conditions (1500 μmol m -2 s -1 ), while temperature, humidity, and CO 2 (400-450 ppm) were set to ambient conditions. Stem water potential (Ψ stem ) was measured at midday (between 12:00-14:00 h, local time) on one mature leaf per vine (same plants as used for gas exchange analysis) with a Scholander pressure chamber (PMS 600, PMS Instrument Company, USA) as described in Levin (2019).

Leaf pressure-volume curves
In season 2020, at the end of the experiment (DOY 213), five leaves per cultivar (only well-watered control plants) were sampled in the morning (08:00 h) for the construction of pressure-volume curves. The bench dry method was used, as described in Dayer et al. (2020), which consists of plotting the reciprocal of leaf water potential (-1 / Ψ) against the relative water content (RWC). The RWC of the leaf was calculated as RWC = (fresh weight -dry weight) / (full turgor weight -dry weight) × 100. From the pressure-volume curves, the leaf water potential at the turgor loss point (Ψ tlp ), the osmotic potential at full turgor (π 100 ), and the modulus of elasticity (ε) were calculated as described in Dayer et al. (2020).

Statistics
The statistical evaluation of the results was carried out using the software SPSS Statistics (version 26, IBM Corp.). One plant was considered a biological replicate. All results Values are averages per cultivar + standard error of the mean (n > 15 and n = 10 in 2020 and 2021, respectively) sampled during the experiments (DOY 189 and DOY 214 in 2020 and 2021, respectively). Different letters identify significant differences between cultivars after a Tukey-HSD post-hoc test (p < 0.05).
were tested for normality and homogeneity of variance prior to being subjected to F-test (p < 0.05) in a two-way ANOVA (results of vegetative growth, gas exchange, and stem water potential; main factors were treatment and cultivar and their interaction) or one-way ANOVA (results of leaf size and pressure-volume curves; factor was cultivar). Post-hoc analyses were performed using the Tukey-HSD test (p < 0.05). All figures were created using SigmaPlot v14 (Inpixon, CA, USA) or ggplot2 package (Wickham, 2016) in R.

Vegetative growth
Despite different conditions (experimental set-ups, environmental conditions, rootstocks) within the two experiments, results show connective significances in some vegetative growth parameters (Figure 2, Table 1).
As observed in Figure 2, Muscaris (M) was the cultivar with the largest leaves (average area of single leaves on the main shoot of ca. 150 cm 2 across seasons), followed by Souvignier gris (S) (ca. 130 cm 2 ) and Donauriesling (D) (ca. 95 cm 2 ). Such differences translated into M recording a higher total leaf area per plant than the other cultivars only in 2020, while in 2021, all cultivars showed similar values of total leaf area per plant (Table 1). In detail, in 2020, before the imposition of drought treatments (DOY 189), the plant leaf area recorded was 3680.3, 2536.9, and 2898.6 cm 2 vine -1 in M, S, and D, respectively; after 23 days of experiments (at DOY 212), M, S, and D recorded (average across treatments) 4996.7, 3394.5, and 3896.2 cm 2 vine -1 , respectively. As expected, drought treatments limited growth in all cultivars that resulted in leaf areas of ws vines being, on average, 15 % smaller than ww ones. Similar growth restriction was observed in 2021 after the three cycles of ws treatments (DOY 261). Statistical interactions of treatment × cultivar were not significant (p > 0.05) in both seasons (Table 1).

Water potential and gas exchange
According to our expectations, drought imposition determined a decrease in stem water potential (Ψ stem ), and values recorded  for the V. vinifera cultivars were not significantly different from those from the tested PIWI cultivars in both seasons (Figure 3, Supplementary Figure 2). In 2020, it is notable that the average Ψ stem of Souvignier gris (S) vines recorded significantly higher values as compared to all other cultivars on DOY 189, 196, and 202 after irrigation withholding ( Figure 3A). The Ψ stem of well-watered control vines ranged between -0.32 and -0.57 MPa without significant differences between cultivars ( Figure 3C). Starting from equal Ψ stem values of -0.35 MPa at DOY 189, water-stressed vines of all cultivars showed a decrease in Ψ stem that reached the lowest values at the peak of drought stress on DOY 202, ranging from -1.10 MPa (in S) to -1.35 MPa (in M) (Supplementary Figure 2). During all three cycles in 2021, drought treatment decreased the average Ψ stem of all cultivars ( Figure 3B). Speaking of statistical significances, in the first cycle, the minimum of Ψ stem was at -0.62 and -1.07 MPa for averaged ww and ws vines, respectively ( Figure 3D Re-watering in 2020 at DOY 203 resulted in stomata reopening of ws plants that did not reach the values of ww controls until DOY 208 ( Figure 4C). In 2021, re-watering after the first drought cycle lasted for 20 days, and, even if not monitored, it is assumed that ws vines reached similar g s values as the ww ones. Indeed, there are no significative differences between ww and ws at the beginning of the second cycle at DOY 242 ( Figure 4D); between the second and third cycles of drought, re-watering of two days did not produce a significant effect in g s ( Figure 4D). Supplementary Figures 4 and 5. Both parameters resembled the stomatal conductance (g s ) behaviour previously described. After a two-way ANOVA analysis with cultivars and treatments as main factors, the results showed that the averaged transpiration rates ( Figure 5B).

Leaf pressure-volume parameters
Analysis of leaf pressure-volume relations (performed only in 2020) showed that the water potential at the turgor loss point (Ψ tlp ) and the osmotic potential at full turgor (π 100 ) were more negative (by 0.2 MPa) in S as compared with the other tested cultivars ( Table 2). The modulus of elasticity (Ԑ) did not differ among cultivars and was, on average, 10.7 MPa.

TABLE 2.
Pressure-volume curves were used to calculate the water potential at turgor loss (Ψ tlp ), the osmotic potential at full turgor (π 100 ), and the modulus of elasticity (Ԑ). 12.4 ± 0.71 a 10.7 ± 1.33 a 9.08 ± 0.40 a Values are averages ± standard error (n = 5). Within each row, different letters indicate statistical differences between cultivars determined by the Tukey-HSD post-hoc test (p < 0.05).

DISCUSSION
The results obtained in this study allowed the comparison of the water use and response to drought of three commercial PIWI cultivars. We found that Souvignier gris (S) recorded Vertical lines indicate the imposition of drought (red) and re-watering (blue) treatment. The figures represent a two-way ANOVA analysis with cultivars and treatments as main factors; therefore, values are presented as averages ± standard error per cultivar (A, 2020: n = 6; B, 2021: n = 8) or per treatment (C, 2020: n = 12; D, 2021: n = 16) on each date (DOY, day of the year). No interaction between main factors was observed for almost any date of observation. In the figure, different letters denote significant differences (p < 0.05) between the main factor cultivar (A, B) or treatment (C, D); eventual significant interactions are represented with an asterisk (*) between the main factors cultivar and treatment (p < 0.05): No differences among cultivars were recorded within the ws treatment, but within the ww treatment Souvignier gris and Riesling were significantly higher than Donauriesling and Muscaris. D, Donauriesling; M, Muscaris; S, Souvignier gris; R, Riesling/G, Grüner Veltliner; ww, well-watered; ws, water-stressed.
the highest maximum stomatal conductance (g s max ) values among the PIWI cultivars tested (Figure 4), and it was also the one with the lowest (most negative) turgor loss point (Ψ tlp ) and osmotic potential at full turgor (π 100 ) as compared to Muscaris (M) and Donauriesling (D) ( Table 2). Dayer et al. (2020) showed that these leaf traits (g s max ,Ψ tlp , π 100 ) are coordinated within grapevine cultivars in a way that when maximum water use (represented in our case with g s max ) increases, leaf hydraulic traits allow the leaf to operate across a wider range of water potentials (i.e., by exhibiting a more negative Ψ tlp and π 100 ). In our case, this may signify that S would provide increased resilience against drought events. However, under the experimental conditions in our study, S did not reach lower water potentials than M and D despite maintaining higher stomatal conductance (and thus a consequent faster soil water depletion), highlighting the difficulty of inferring plant behaviour from single leaf measurements as other factors also play an important role such as the canopy size and the variability among leaves. For instance, in 2020, M was the cultivar with the lowest g s max but also the one with the largest canopy (and therefore a bigger evaporative surface as compared to S), a combination of factors that could explain a faster depletion of soil water, in turn resulting in a lower midday water potential than S. Furthermore, despite the differences observed among cultivars when plotting g s or Ψ stem against time, our data showed that the water potential ranges for stomatal closure did not differ among cultivar as all of them reached g s < 0.05 mol m -2 s -1 between Ψ stem of -0.8 and -0.9 MPa ( Figure 5); such results are comparable to what is reported in similar studies on potted grapevine cultivars such as Merlot (Hochberg et al., 2016), Grenache and Syrah , or Montepulciano (Tombesi et al., 2014). Therefore, the different drought stress degrees experienced by the PIWI cultivars in our study were more related to the different dynamics of water use by transpiration (and thus the time needed to deplete the soil water reservoir and reach the critical Ψ stem for stomatal closure) than to an intrinsic difference in their stomatal sensitivity to drought.
A closer look at the cultivar's gas exchange under well-watered conditions provides a useful picture of their water use behaviour. As shown in Supplementary  Figure 6, differences in g s between cultivars existed only at Ψ stem > -0.75 MPa, with S, exhibiting higher stomatal conductance than M and D. In particular, M was the most conservative cultivar among the PIWIs studied in terms of stomatal conductance (i.e., M recorded g s max values that were lower by 30 % than the values measured for S) and transpiration. Our results suggest that M and D might be more suitable for non-irrigated conditions than S as they would deplete the soil water reservoir slower, while S, by maintaining higher transpiration rates longer, could face more intense drought stress situations and, therefore, some irrigation strategy would be needed, consistent with the idea of a "stress distance" introduced by Gambetta et al. (2020).
Interestingly, from the A N ~ g s relationship (Figure 6), it is worth noting that M consistently showed lower A N values than the other PIWIs at g s > 0.15 mol m -2 s -1 , while D and S recorded largely similar A N rates across a wide range of g s values. The A N ~ g s curve ( Figure 6) showed a plateau at maximum assimilation rates of 14 μmol m -2 s -1 for S and D (consistent with values frequently reported in grapevines; Medrano et al., 2002;Bota et al., 2005;Bota et al., 2016), while M reached an assimilation rate plateau between 10 to 12 μmol m -2 s -1 , suggesting a lower intrinsic water use efficiency (WUEi) at the leaf level and under well-watered conditions for M as compared to S and D leaves. However, as discussed before, M presented larger leaves and a bigger canopy than the other cultivars, a fact that could compensate for a lower WUEi. It remains difficult to upscale single-leaf WUEi to whole-plant WUE (Medrano et al., 2015;Tomás et al., 2014), and more studies are needed to further explore this trait, including yield components and total biomass production.   The results of this study provide important hints on the water relations of three PIWI cultivars, but we cannot exclude that under natural settings, the plant response to drought would differ from what was observed in young potted vines. Firstly, the stomatal closure thresholds would be different since, as suggested by Sorek et al. (2021) and Herrera et al. (2022), a seasonal osmotic adjustment in the leaves of grapevines can shift their sensitivity to drought towards lower (more negative) water potentials. In this regard, an open question remains whether different grapevine genotypes can osmotically adjust at different rates, perhaps contributing at least in part to the differences in stomata sensitivity to drought among grapevine cultivars often reported in the literature (Lavoie-Lamoureux et al., 2017). Secondly, under field conditions, a large soil root explored volume (as normally observed in vineyards) would result in a longer time needed to dehydrate the soil . Both factors combined would mean more time for the plants to react and acclimate to water deficit conditions (Hochberg et al., 2017), an aspect still poorly explored in general for grapevines and even more in PIWIs. Finally, we shall not forget that important research questions concerning the impact of drought on the yield and fruit composition of PIWIs are to be approached to have a complete picture and be able to provide viticulturists and oenologists with the necessary information to manage PIWIs under dryer climatic conditions.

CONCLUSION
Taken together, the results of this study showed more conservative water use (i.e., lower maximum stomatal conductance values) of Donauriesling (D) and Muscaris (M) as compared to Souvignier gris (S) under no water limitations. Under mild-drought conditions, S maintained higher rates of gas exchange compared to D and M, but such differences vanished with the intensification of drought. While several important points are still to be studied (grape berry composition), we stress that our experiment produced similar results under greenhouse and semi-controlled field conditions and, therefore, under highly heterogeneous environments. Therefore, the study outputs might be generalised towards a hypothesis-driven validation at the field level to fully address best water management practices for PIWI vineyards.