Stomatal closure is induced by hydraulic signals and maintained by ABA in drought-stressed grapevine

Water saving under drought stress is assured by stomatal closure driven by active (ABA-mediated) and/or passive (hydraulic-mediated) mechanisms. There is currently no comprehensive model nor any general consensus about the actual contribution and relative importance of each of the above factors in modulating stomatal closure in planta. In the present study, we assessed the contribution of passive (hydraulic) vs active (ABA mediated) mechanisms of stomatal closure in V. vinifera plants facing drought stress. Leaf gas exchange decreased progressively to zero during drought, and embolism-induced loss of hydraulic conductance in petioles peaked to ~50% in correspondence with strong daily limitation of stomatal conductance. Foliar ABA significantly increased only after complete stomatal closure had already occurred. Rewatering plants after complete stomatal closure and after foliar ABA reached maximum values did not induced stomatal re-opening, despite embolism recovery and water potential rise. Our data suggest that in grapevine stomatal conductance is primarily regulated by passive hydraulic mechanisms. Foliar ABA apparently limits leaf gas exchange over long-term, also preventing recovery of stomatal aperture upon rewatering, suggesting the occurrence of a mechanism of long-term down-regulation of transpiration to favor embolism repair and preserve water under conditions of fluctuating water availability and repeated drought events.

observed between the two cultivars in terms of soil water content except on day 16. Pre-dawn water potential was similar in the two cultivars during the first 7 days of the experiment ( Fig. 2A). From day 8 until the end of the experiment, Sangiovese had slightly lower Ψ pd than Montepulciano, and in days 8, 11, 12 and 15 the difference recorded was statistically significant (P < 0.05). Midday stem water potential (Ψ md ) decreased progressively during the experiment and was lower in Sangiovese than in Montepulciano excluding days 4, 6, 7, 9, and 11 when the difference was not significant, or days 10 and 12 when Sangiovese had significantly higher Ψ md than Montepulciano (Fig. 2B).
Stomatal conductance (g s ) measured at midday decreased during the experiment showing a similar trend between cultivars (Fig. 3A). In more detail, the largest g s decrease occurred between days 5 and 8 when stomatal conductance dropped from about 0.130 mol m 2 s −1 to 0.008 mol m 2 s −1 in both cultivars. Net assimilation (A n ) rates followed a pattern similar to g s (Fig. 3B), yet over the first days of the experiment Montepulciano had higher assimilation rates than Sangiovese (days 1 and 2), whereas in the second part of the experiment, when g s was limited and stem water potential decreased (after day 7), Montepulciano and Sangiovese had similar net assimilation values. Foliar ABA remained constant (mean values per cultivar were not significant, P > 0.05, ANOVA) from day 1 until day 8 (Fig. 3C). On day 9, Foliar ABA increased about 3 fold in both cultivars and in Sangiovese it continued to rise until day 11, when foliar ABA peaked to a value almost 6 fold higher as compared to initial values.
Stomatal conductance and net assimilation rates showed a faster decline at higher Ψ md (i.e. less negative) in Montepulciano than in Sangiovese (Fig. 4A). In both cultivars, g s (Sangiovese R 2 = 0.75 P < 0.001; Montepulciano R 2 = 0.83 P < 0.001 ANOVA) and A n (Sangiovese R 2 = 0.90 P < 0.001; Montepulciano R 2 = 0.87 P < 0.001 ANOVA) were significantly correlated to Ψ md . In particular, g s started to decline in Montepulciano at Ψ md < − 0.7 MPa and reached values close to zero at Ψ md < − 1.1 MPa, while, in Sangiovese, g s started to decline at Ψ md < − 1.0 MPa and reached values close to zero at Ψ md < − 1.5 MPa. Similar trends were observed in leaf A n in both cultivars, with Montepulciano displaying a steeper reduction of A n at progressively lower Ψ md than Sangiovese (Fig. 4B).
Diurnal course of Ψ stem , g s and foliarABA progressively changed depending on water stress severity (Fig. 6). On day 2, Ψ stem decreased from 4:00 am to 1:00 pm and remained approximately stable until 6:00 pm. During the same day, maximum g s was measured at 9:00 AM while a progressive decrease was observed during the rest of the day. Montepulciano displayed higher g s values than Sangiovese at all daytimes except at 6:00 pm. Foliar ABA was constant over the whole day and was slightly higher in Sangiovese than in Montepulciano, although this difference was statistically significant only at 4:00 am. On day 8, Ψ stem decreased from − 0.3 MPa and − 0.5 MPa at pre-dawn (4:00 am), to − 1.0 MPa and − 1.2 MPa at 6:00 pm in Montepulciano and Sangiovese, respectively. Ψ stem measured at 6:00 pm was slightly more negative than values measured at 1:00 pm. Montepulciano had Ψ stem significantly higher than Sangiovese over the whole day. On day 8, g s was consistently lower than on day 2 and maximum g s measured at 9:00 am was below 0.1 mol m −2 s −1 , while near-complete stomatal closure occurred at 1:00 pm and was maintained also at 6:00 pm. On the same day, foliar ABA was slightly higher than values recorded on day 2, although there was no difference between the measurements carried out at 4:00 am, 9:00 am and 1:00 pm on day 2 and day 8. On day 8, a consistent increase of foliar ABA was measured at 6:00 pm in both cultivars. On day 15, Ψ stem decreased from − 1.36 MPa and − 1.58 MPa at pre-dawn (4:00 am) to − 1.76 MPa and − 1.8 MPa at 6:00 pm in Montepulciano and Sangiovese, respectively. In Montepulciano, Ψ stem was consistently higher than in Sangiovese at any timing except for 6:00 pm. On the same date, g s was steadily close to zero in both cultivars. Foliar ABA further increased in both cultivars on day 15 compared to day 8.  Although the difference was not statistically significant during the day except at midday, Sangiovese leaves had higher ABA than Montepulciano leaves on day 15.
Leaf petiole percent loss of hydraulic conductance (PLC) ranged across 20% until day 5 in both cultivars (Fig. 7). On day 6, in both cultivars, leaf petiole PLC increased to values close to 50% and remained stable until day 11. On day 12, when petiole PLC was measured immediately after rain, petiole PLC dropped to values below 20%. On this date, Montepulciano had significantly higher PLC than Sangiovese. On day 15 and 16, PLC increased again up to values similar to those measured between days 6 and 11.  Foliar ABA, as well as g s (Sangiovese R 2 = 0.67 P < 0.001; Montepulciano R 2 = 0.71 P < 0.001 ANOVA) and A n (Sangiovese R 2 = 0.79 P < 0.001; Montepulciano R 2 = 0.80 P < 0.001 ANOVA), were significantly correlated with Ψ leaf (Sangiovese R 2 = 0.87 P < 0.001; Montepulciano R 2 = 0.62 P < 0.001 ANOVA) (Fig. 8). Foliar ABA increased at water potential values at which g s was below 0.02 mol m −2 s −1 in both cultivars.

Discussion
Vitis vinifera is generally considered an anisohydric species, although intra-specific variability of stomatal behaviour under water limitation condition has been observed in different grapevine genotypes 40 . As also reported for other species 41 , progressive dehydration imposed on Sangiovese and Montepulciano vines induced physiological responses that can be conveniently divided in three successive stages (Fig. 3). At stage I, when soil water is still plenty available and not affected by the onset of drought, gas exchange rates are not influenced by water shortage. In stage II, the progressive reduction of soil available water induces partial stomatal closure and gas exchange limitation. In stage III, stomata are fully and steadily closed, gas exchange rates drop to values close to zero and leaf senescence is triggered. Throughout stage I the vine can still extract water from soil, so that no stomatal regulation of transpiration is apparent and diurnal changes of Ψ stem are mainly influenced by the plant hydraulic resistance and the atmospheric evaporative demand. In our experiment, at this stage, PLC due to embolism was negligible (< 20%) and leaf net assimilation was maximized. When soil water content approached critical values (< 0.1 m 3 m −3 ) in stage II, plant water potential decreased, probably enhancing the plant capacity to extract water from the soil. Progressive drop of plant water potential, however, causes xylem pressure to drop to very negative values, thus increasing the likelihood of embolism induction in xylem vessels and consequent hydraulic failure 7 . Under such conditions, stomatal closure plays a key role in preventing hydraulic failure by regulating the rate of water loss and limiting the xylem pressure drop 6,8,16 . In our experiment, stomata started to close 6 days after the beginning of water withdrawal, concurrently with the recorded increase of PLC of leaf petioles 42 that approached values close to 50%. In particular, gas exchange rates were more limited in the afternoon than in the morning, when plant water status was more favourable and PLC was likely lower due to partial nocturnal vessel refilling [42][43][44] . These results are consistent with the coordination between stomatal movements and PLC reported in other species: in walnut Cochard et al. 45 concluded that cavitation avoidance was a physiological function associated with stomatal regulation during water stress; in douglas-fir, g s was linearly correlated with PLC 46 ; in a study on eight tropical species Brodribb et al. 47 concluded that leaf and xylem hydraulic traits are correlated with the response of stomata to Ψ leaf inducing loss of xylem hydraulic conductivity; and finally, in a literature overview on 70 woody species, Klein 48 concluded that stomatal sensitivity to leaf water potential strongly relates to xylem characteristics.
ABA has been frequently suggested to act as a root-to-leaf chemical signal under progressive soil drying 49 , but some experimental observations apparently contrasts with this view. As an example, Soar et al. 50 recorded gradients in xylem and foliar ABA along shoots of V. vinifera, and found that the concentrations of the hormone were higher close to the apex and decreased downwards. In our experiment, foliar ABA remained below 2 ng mg −1 dw up to the occurrence of almost complete stomata closure (8 th day). A significant increase of foliar ABA occurred only after complete stomatal closure had already occurred. On day 7 foliar ABA rose by 63% (Sangiovese) and 74% (Montepulciano) in comparison with the mean foliar ABA recorded in the previous days. However, we note that these data need to be interpreted with caution, as our experimental approach cannot rule out the hypothesis that early stomatal responses are influenced by very localized increases in ABA concentration, eventually below the detection limit of bulk measurements. Moreover, our data do not exclude the possibility of very local changes in ABA uptake rate into guard cells.
In grapevine leaves fed with exogenous ABA, a similar foliar ABA increase resulted in g s reduction of about 20% 33 , i.e. less than the 64% and 61% reduction of g s measured in our experiment in Sangiovese and Montepulciano, respectively. Furthermore, foliar ABA was correlated with Ψ leaf and increased at Ψ leaf values corresponding to g s < 0.02 mol m −2 s −1 , thus suggesting a potential signal inducing ABA synthesis and mediated by water potential. These data support the hypothesis of foliar ABA accumulation when stomata are nearly or completely closed. These results are consistent with those reported in some gymnosperms and other angiosperm species. As an example, in M. glyptostroboides, ABA production was triggered after leaf turgor loss and stomatal closure 25 . In progressively water stressed Zea mays and Sorghum bicolor, stomatal closure preceded any detectable ABA increase 26 , while in flooded tomato plants, bulk leaf ABA increased only after the onset of stomatal closure 29 . Finally, in stress adapted cotton plants (Gossypium hirsutum L.) higher levels of ABA were not correlated with stomatal response to low leaf water potentials 27 and in one Populus spp. hybrid, stomatal aperture was not affected by ABA 28 . In our study, when stomatal conductance was plotted against foliar ABA, a significant non-linear correlation appeared (R 2 = 0.47, P < 0.0001), but when non limiting (> 0.05 mmol m 2 s −1 ) 51 and limiting stomatal conductance values were separately linearly regressed, no significant correlation was observed (Fig. 5). These results obtained during a relatively fast (2 weeks) dry-down experiment apparently contrast with those obtained by Speirs et al. 52 on vines planted in the field subject to different irrigation treatments over the season.
Our results suggest reconsidering the role played by ABA in grapevines subjected to drought stress. Apparently, ABA played no critical role in regulating gas exchange during early onset of drought conditions, as stomatal closure was apparently triggered by xylem embolism and consequent reduction of plant hydraulic conductance and leaf turgor. After stomatal closure, ABA increased by 566% and 418% in Sangiovese and Montepulciano, respectively. These levels were well above those measured by Loveys 33 , who measured in ABA-fed grapevine leaves a 50% reduction of stomatal conductance reduction upon ABA increase by 366%. In our experiment, a five-fold rise appeared to be effective at constraining stomatal aperture even after recovery of plant water status. In fact, when a heavy rain event occurred on day 12, water availability in the soil increased to levels not limiting for stomatal conductance, and recovery of Ψ stem and refilling of embolized vessels were also recorded. Nonetheless, stomatal conductance remained close to zero, and this suggests that ABA inhibited stomatal re-opening although water availability was sufficient to fuel the recovery of water potential and xylem conductance. Our results would also suggest that stomatal closure can trigger foliar ABA increase, and such high levels would persist afterwards despite transient recovery of leaf water status. Stomatal behaviour patterns similar to those described here were reported in Pinus radiata and Eucalyptus pauciflora by Brodribb and McAdam 24 and Martorell et al. 39 , respectively. On the basis of these studies and our present results, it might be speculated that this Scientific RepoRts | 5:12449 | DOi: 10.1038/srep12449 behaviour might represent an effective physiological mechanism to prevent water loss in environments where plants are exposed to the risk of periodical summer drought, in that initial drought stress could induce, via ABA accumulation, the a priori down-regulation of leaf transpiration. Furthermore, stomatal closure induced by ABA has been indicated to favour vessel refilling 53 by promoting de-polymerization of starch pools stored in parenchyma cells to produce sucrose or other sugars and generate the necessary osmotic gradients driving refilling 54,55 . The lack of stomatal conductance recovery after re-watering at day 12, coupled with the prompt decrease of PLC, would also suggest that stomatal closure induced by ABA is an important mechanism to favour fast rehydration and the rise of xylem pressure, thus further promoting osmotic-mediated embolism recovery in the case of occasional rewatering 56,57 .
Tallman 58 (2004) hypothesized that diurnal patterns of stomatal movement are linked to ABA fluctuations and that stomatal movements are the results of the combination of ABA catabolism early in the morning and ABA biosynthesis and import from the apoplast around guard cells after midday. In our experiment, foliar ABA was steady during the whole day when water was not limiting. When stomatal closure was observed in the late morning of day 8, foliar ABA rose only in late afternoon. When stomata were steadily closed over the day (day 15) foliar ABA was steadily higher (in comparison to day 2) over the whole day. It has to be noted that foliar ABA levels might not be representative of levels actually active in guard cells. Locally active ABA levels result from the activity of ABA transport and the intensity of ABA metabolism. Therefore, levels of ABA in guard cells, which regulate turgor and are crucial for stomatal movements, could be increased over bulk leaf levels by spatially restricted activity of ABA transporters or ABA cleaving/conjugating/deconjugating enzymes. Our experimental procedures do not allow to detect eventual ABA accumulation patterns in guard cell that could explain stomatal movements during the day, but the analytical method used in the experiment allowed to measure the active form (not conjugated) of ABA. In agreement with previous experiments carried out on almond 59 , our data show that there is no foliar ABA reduction during the day time when stomata are open. These results are partially in contrast with the model proposed by Tallman 58 .
According to Tardieu and Simonneau 60 and Soar et al. 35 , isohydric vs anisohydric behaviour (different stomatal closure dynamics in response to water potential drop) might arise from species-specific differences in stomatal sensitivity to ABA 49,61 . Sangiovese and Montepulciano are considered anisohydric and near-isohydric cultivars, respectively 36,62,63 . In our experiment, near-isohydric Montepulciano displayed lower levels of foliar ABA during all stages of the drought stress treatment in comparison with the anisohydric Sangiovese. Furthermore, the increase of Montepulciano foliar ABA recorded in stage III was less marked than in Sangionvese. Recent experiments pointed out that xylem vulnerability to cavitation might play a key role in defining stomata behaviour in V. vinifera 36 and in a wide range of other species 48 . In particular, Klein 48 concluded that isohydric vs anisohydric behavior do not represent two distinct categories but rather two extremes in a continuum of hydraulic strategies that dictates stomatal behavior due to the strong relationship between xylem characteristics and stomatal sensitivity to leaf water potential. Our results support this hypothesis, even though they do not allow to rule out a different sensitivity of stomata to similar levels of foliar ABA in these two cultivars.
Early studies on Avena sativa leaves pointed out that leaf senescence was triggered by stomatal closure 64 , and that stomatal closure itself triggers ABA accumulation, so that ABA might be the proximal cause of leaf senescence 65 . Experiments on rice, indicated that ABA is involved in senescence regulation 66 . Indeed, exogenously applied ABA induces the expression of several SAGs (senescence-associated genes) in Arabidopsis thaliana 67 . In a recent study on A. thaliana, Zhang and Gan 68 concluded that there is a unique regulatory chain controlling stomatal movement and water loss during leaf senescence. Our results are consistent with this view, also considering that Sangiovese (which showed higher foliar ABA in comparison with Montepulciano) started leaf senescence and consequent leaf abscission earlier than Montepulciano (in the last experiment day it was not possible to perform measurements on Sangiovese because of the complete abscission of primary leaves).
In conclusion, in V. vinifera passive hydraulic control of stomatal closure appears to be dominant over eventual chemical signalling at the early phases of drought stress. The increase of foliar ABA concentration after stomatal closure was apparently addressed at inhibiting stomatal opening under transient rehydration, thus favouring embolism recovery. This mechanism could be effective in reducing a priori water loss in environments where periodical severe drought frequently occurs during the vegetative season. However, grapevine might also have evolved other, more sophisticated mechanisms to tightly control local ABA activity. More studies are needed to verify this alternative hypothesis. . Pots (60 liters volume) were filled with loam soil. At the end of February, each vine was pruned to retain four spurs with two buds each. All shoots were oriented upright using suitable stakes. A total number of twenty vines was used in the experiment. Ten vines per cultivar were used and initially maintained at field capacity until 7 th July. Water was supplied every day at 8:00 pm. On Scientific RepoRts | 5:12449 | DOi: 10.1038/srep12449 8 th July, drought was imposed on all vines by completely suspending irrigation and covering pot surface by plastic film. The drought treatment was continued until complete leaf abscission in both cultivars.

Materials and Methods
Daily measurements of soil water content were carried out at 4:00 am by a Diviner 2000 capacitance probe (Sentek Environ Tech., Sentek Environment Technologies, Stepney, South Australia), using access tubes located in three pots per cultivar. Measurements were performed at 100 mm, 200 mm and 300 mm depths from the soil surface of the pot (400 mm high). The total soil water content (Θ w ) in the pot was expressed as the arithmetic mean of the measurements performed at different depths.
Gas exchange and water potential. Stomatal conductance (g s ) and net assimilation (A n ) measurements were carried out on adult primary leaves grown between the 4 th and the 10 th node from the shoot base. Measurements were carried out between 4:00 am and 5:00 am, 8:00 am and 9:00 am, 12:00 am and 1:00 pm, 4:00 pm and 5:00 pm from 8 th July until 23 th July on one representative leaf sampled from 5 vines per cultivar using an open gas exchange system (ADC-System, LCA-3, Hoddesdon, UK) equipped with a Parkinson leaf chamber (11.2 cm 2 ). Measurements during the daytime (8:00 am to 5 pm) were performed under saturating light conditions (PPFD > 1200 μ mol photons m −2 s −1 ). Stem water potential (Ψ stem ) was measured over the same days and daytimes using a pressure chamber (Soilmoisture Corp, Santa Barbara, CA, USA). Ψ stem was measured on each vine on one mature leaf that had been wrapped in plastic film and aluminum foil 2 h prior to the measurements 69 . Water potential values measured at 4:00 am are reported as pre-dawn water potential (Ψ pd ).
Foliar ABA determination. Foliar ABA was determined on three primary leaves (the same used for gas exchange measurements) per cultivar sampled from different vines between 12 a.m. and 1 p.m. from 8 th July (day 1) to 23 rd July (day 16). On experimental days 2, 8 and 15 (9 th , 15 th and 22 nd July, respectively), [ABA] was determined on leaves sampled concurrently with daytimes when gas exchange measurements were performed. ABA was extracted following the procedure described by Villarò et al. 70 with some modifications. Leaves used for gas exchange measurements were immediately placed in liquid nitrogen and then stored in a freezer at − 80 °C. Then, the material was weighted (fresh weight) and lyophilised (LIO5P, 5Pascal, Trezzano, Italy). Lyophilised material was weighted (dry weight) and grinded (MF10, IKAlabortechnik, Staufen, Germany). Leaf material (0.1 g) was extracted with 10 ml of methanol/water (1:1 v/v, pH = 3 with formic acid) for 30 min using a ultrasonic bath. After centrifugation, the supernatant was filtered through a paper filter and the same procedure was repeated for the remaining pellet. The collected filtrates were extracted twice with dichloromethane (15 ml) and the organic phase evaporated under vacuum. The residue was dissolved to a 1 ml with acetone and water/ acetonitrile (50:50 v/v, 0.1% formic acid) for the HPLC analysis. Analytical standards of (± ) Abscisic acid (purity ≥ 98.5%) was purchased from Sigma-Aldrich, PA-grade methanol, acetone, dichloromethane and formic acid, and HPLC-grade acetonitrile and water were purchased from VWR Chemicals. Analyses were performed on a Perkin-Elmer PE 200 system (Autosampler, Binary Pump and UV-VIS detector) equipped with an IB-Sil C8-HC (5 mm × 250 mm × 4.6 mm Phenomenex) column and IB-Sil C8 (5 mm × 30 mm × 4.6 mm Phenomenex precolumn) at a flow rate of 0.8 mL min −1 ; the injection volume was 20 μ L and the detection was made at 270 nm. The mobile phase of acetonitrile/water (30:70 v/v, 0.1% formic acid) was previously filtered and degassed. The compound was identified by comparing the retention times with those of authentic reference compound. The peaks were quantified by an external standard method, using the measurements of the peak areas and a calibration curve. Stock solutions of ABA standards were prepared by diluting a solution (10 mg mL −1 in acetonitrile) to obtain a range of concentrations from 0.01 to 10 mg mL −1 . The limit of detection (LOD) was 0.005 mg L −1 .
Percentage loss of hydraulic conductance. Percentage loss of xylem hydraulic conductance (PLC) was measured on five petioles per cultivar harvested between 12 a.m. and 1 p.m. on each day. Petioles were cut under water from leaves inserted nearby those used for gas exchange measurements. Hydraulic conductance of petioles was measured by connecting one sample end to plastic tubing filled with a filtered (0.2 μ m) 20 mM KCl solution and connected to a pressure head maintained at a pressure (P) of 6 kPa. Flow (F) was measured by collecting the fluid from the distal end in pre-weighted sponge pieces fitted in plastic tubes. Flow readings were taken over 1 min time intervals. After approximately 30 minutes, once flow was found to be steady, sample hydraulic conductance (K) was calculated as F/P and the samples were flushed at P = 0.2 MPa for 30 min, to remove eventual embolism. After flushing samples, maximum hydraulic conductance (K max ) was measured as above.
Percentage loss of hydraulic conductance (PLC) value was calculated as: Differences between the two genotypes were assessed using the Student's t-test (P < 0.05). The significance of regressions was tested using Pearson Product Moment Correlation.