Journal of Experimental Botany, Vol. 57, No. 11, pp. 2577–2587, 2006 doi:10.1093/jxb/erl020 Advance Access publication 25 July, 2006

RESEARCH PAPER Ripening remain hydraulically connected to the shoot

Markus Keller1,*, Jason P. Smith2 and Bhaskar R. Bondada3 1 Irrigated Agriculture Research and Extension Center, Washington State University, 24106 N. Bunn Road, Prosser, WA 99350, USA Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 2 National Wine and Grape Industry Centre, Charles Sturt University, Locked Bag 588, Wagga Wagga, NSW 2678, Australia 3 Washington State University Tri-Cities, 2710 University Drive, Richland, WA 99354, USA

Received 19 December 2005; Accepted 11 April 2006

Abstract with solute accumulation in the apoplast may be responsible for the decline in xylem water influx into Berry diameter was monitored during dry-down and ripening grape berries. Instead, the xylem may serve to rewatering cycles and pressurization of the root sys- recycle excess phloem water back to the shoot. tem of Vitis vinifera (cv. Merlot) and Vitis labruscana (cv. Concord) to test changes in xylem functionality Key words: Apoplastic dye, deficit irrigation, grape berry, during grape ripening. Prior to veraison (onset of phloem, root pressure, vascular connection, water stress, ripening), berries maintained their size under declining xylem. soil moisture until the plants had used 80% of the transpirable soil water, began to shrink thereafter, and recovered rapidly after rewatering. By contrast, berry diameter declined slowly but steadily during post- Introduction veraison water stress and did not recover after rewa- tering; irrigation merely prevented further shrinking. In premium wine production, quality is considered Preconditioning vines with a period of water stress to be far more important than crop yield, and small berry after flowering did not influence the berries’ reaction size is deemed desirable. There is a widespread belief to subsequent changes in transpirable soil water. Pres- in the wine industry that rain or irrigation close to harvest surizing the root system led to concomitant changes may increase berry size and cause a ‘dilution’ of solutes in berry diameter only prior to veraison, although some (, acids, anthocyanins, tannins, etc.) or even cracking post-veraison Concord, but not Merlot, berries cracked (splitting) of berries. In Europe, this belief is often written under root pressurization. The xylem-mobile dye basic into the law, and irrigation is prohibited or strictly fuchsin, infused via the shoot base, moved throughout regulated. Even in the New World, wineries may encourage the berry vasculature before veraison, but became growers to withhold irrigation water at this critical time gradually confined to the brush area during ripening. to avoid any perceived adverse effects. Surprisingly, there When the dye was infused through the stylar end of is little scientific evidence that late-season water uptake attached berries, it readily moved back to the plant by the roots and transport to the fruit is detrimental to grape both before and after veraison. Our work demonstrated quality. In the absence of rain, which may lead to direct that berry–xylem conduits retain their capacity for water uptake through the skin (Considine and Kriedemann, water and solute transport during ripening. It is pro- 1972; Lang and Thorpe, 1989), water influx into oc- posed here that apoplastic phloem unloading coupled curs via both the xylem and the phloem. Conversely, efflux may be due to fruit transpiration and xylem ‘backflow’

* To whom correspondence should be addressed. E-mail: [email protected] Abbreviations: Cs, mesocarp cell solute concentration; FTSW, fraction of total transpired soil water; gs, stomatal conductance; mc, mesocarp cell; Px, xylem pressure; R, universal gas constant; rh, hydraulic resistance; T, absolute temperature; VPD, vapour pressure deficit; D, gradient; p, osmotic pressure; W, water potential.

ª The Author [2006]. Published by Oxford University Press [on behalf of the Society for Experimental ]. All rights reserved. For Permissions, please e-mail: [email protected] 2578 Keller et al. from the fruit to the vine. Many studies found that water response to late-season water stress, presumably by altering flow via the xylem into grape berries decreases markedly the hydraulic connection of the berries to the shoot, as during ripening (reviewed by Ollat et al., 2002). Whereas reported for (Davies et al., 2000). Berry diameter xylem sap is the main source of water for the berries before was monitored during repeated dry-down/rewatering cycles veraison (colour change, beginning of ripening, and re- both before and after veraison. In addition, water was sumption of cell expansion after a brief lag phase), phloem forced through the xylem using a root pressure chamber to sap becomes the primary water source after veraison. How- determine if soil water uptake would contribute to berry ever, the causes of the observed decrease in xylem water volume changes (Davies et al., 2000). Changes in both influx remain unclear. Physical disruption of the xylem soil moisture and air pressure applied to the root system conduits inside the berry is usually cited as the main reason are useful to manipulate xylem pressure (Px) (Wei et al., for impeded xylem flow (Du¨ring et al., 1987; Findlay 1999). Finally, an apoplastic dye was used to trace xylem et al., 1987), but this has been disputed (Creasy et al., connections between the berries and the rest of the plant. 1993). Fruit transpiration declines during development Two genetically distinct grape cultivars, Merlot (Vitis Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 (Rogiers et al., 2004), which implies that evaporative water vinifera L.) and Concord (Vitis labruscana Bailey), were loss ceases to function as the driving force for water influx. used for the three sets of experiments. Moreover, the cessation of xylem inflow into grape berries may not be complete (Greenspan et al., 1994; Rogiers et al., 2001). Materials and methods Evidence for a breakdown in xylem functionality Plant growth comes mainly from studies of apoplastic dye perfusion through the berry pedicel (Du¨ring et al., 1987; Findlay Three-year-old, own-rooted grapevines V. vinifera cv. Merlot and V. labruscana cv. Concord were grown in white 20.0 l PVC pots et al., 1987). Other experimental findings were sometimes containing a mixture of 50% sandy loam, 25% peat moss, 25% at odds with this dogma, but while doubts have been pumice, and 30 g lÿ1 dolomite, with a volumetric water content of raised about the validity of the interpretation of dye uptake ;34% at field capacity. The vines were grown outside at the Irrigated studies (Lang and Thorpe, 1989; Rogiers et al., 2001; Agriculture Research and Extension Center in Prosser, Washington, Tyerman et al., 2004), the basic view has remained largely USA (468179 N; 1198449 W; elevation 270 m), before being trans- ferred to a whitewashed, air-conditioned greenhouse (photosynthetic unchallenged. Indeed, the development of a xylem discon- photon flux ;860 lmol mÿ2 sÿ1 under clear sky at midday) for the tinuity in the pedicel or inside the fruit (i.e. hydraulic series of experiments described below. Meteorological conditions isolation of the fruit) is now often regarded as a prerequi- in the greenhouse were recorded with a HMP45C temperature and site to prevent loss of solutes via the xylem in fruits that relative humidity sensor (Vaisala, Helsinki, Finland). Before starting employ apoplastic phloem unloading, such as grape the experiments, water deficit was imposed during the post-anthesis cell division phase of Merlot berries in an attempt to enhance the (Patrick, 1997; Sarry et al., 2004), tomato (Ho et al., xylem connection between the berries and the rest of the plant 1987; Patrick and Offler, 1996), or (Drazˇeta et al., (Davies et al., 2000). After watering the pots to field capacity, 2004; Zhang et al., 2004). It has been argued that high they were dried down twice for two consecutive 14 d periods (Merlot DI) before being returned to the same daily (standard) hydraulic resistance (rh) inside such fruits and restricted xylem influx may be required to promote phloem unload- irrigation regime used for the remaining Merlot (Merlot STD) and Concord vines (Concord STD). Plant leaf area was estimated by ing and fruit softening and to protect the fruit from measuring the length of the main vein of each leaf and regressing excessive backflow, for example, during periods of high length against area as determined on surplus plants, using a Li-Cor evaporative demand (Malone and Andrews, 2001; Tyerman 3100 leaf area meter (Lincoln, NE, USA). Upon completion of the et al., 2004). However, it is not clear how such hydraulic experiments, vines were removed from the pots to determine root isolation should come about. Even if fruit tracheids were distribution and dry weight following drying at 60 8C to constant weight. indeed disrupted, the cell wall and intercellular space apoplast of the fruit would still be contiguous with the Dry-down experiments rest of the plant. Moreover, the xylem and phloem are Three separate experiments were conducted to determine the re- interconnected along their entire length and can readily sponse of berry size to dry-down and rewatering cycles applied exchange water and solutes (Esau, 1977; Zwieniecki et al., before and after veraison. At the start of each experiment, the pots 2004). Thus, while physical xylem disruption would were watered to field capacity and then allowed to dry down. The pot surface was sealed to prevent water loss by means other than certainly increase rh within the fruit pericarp, it would not transpiration. During the second (pre-veraison) and third (post- entirely prevent apoplastic exchange of water and solutes veraison) dry-down cycles, daily plant water use for transpiration was between the fruit and the plant. On the other hand, although determined gravimetrically, and all pots were watered back to field Lang and Thorpe (1989) reported xylem flow from ripe capacity when daily water use had decreased to 10% of the maximum grape berries back to the vine, direct evidence for xylem rate. Daily water use was calculated as a fraction of the total tran- spired soil water (FTSW), which gives an indication of the actual backflow is thus far lacking. plant-available water in the soil (Lebon et al., 2003). This water The present study examined if water stress during the depends on root distribution and, therefore, differs from the early development of grape berries would affect their commonly used (potential) plant-available soil water, which is Grape berry hydraulics 2579 calculated based on soil properties alone and thus ignores the fact that Dye movement experiments roots may not access a portion of the total available soil volume Shoots with four mature leaves and one cluster were cut from because they do not grow there. Although the calculation of FTSW field-grown grapevines in a nearby vineyard before (all berries ignores plant water capacitance, the error due to this omission is green), during (berries ranging from green to blue), or after veraison likely to be small, considering that 3-year-old, pot-grown grapevines (all berries blue). The cut end was immediately placed in a centrifuge were found to contain <250 ml of water in the roots, trunk, and shoots tube containing 30 ml of a 0.1% aqueous solution of the xylem- (Keller and Koblet, 1995; M Keller, unpublished results). mobile dye basic fuchsin (C20H19N3HCl; mol. wt 338 Dqa), and the During each dry-down and rewatering cycle, changes in berry tube was sealed around the shoot to prevent evaporation (‘forward’ diameter were recorded on three berries per vine, using FI-XSM dye infusion). The diameter of 10 marked berries per cluster was linear variable displacement transducers (Phytech, Rehovot, Israel). measured with digital calipers, and the stylar (distal) end of selected Transducers were connected via an AM416 relay multiplexer berries was sliced off with a razor blade to stimulate water (Campbell Scientific, Logan, UT, USA) to a CR10X data logger evaporation from the exposed surface. The pedicel of additional (Campbell Scientific), and their output was calibrated against berry berries on the same cluster was girdled with the blunt end of a razor to starting diameter measured with electronic calipers (0.01 mm re- inhibit phloem influx. Berries were sampled at regular intervals to solution). Testing of the FI-XSM sensors found negligible tem- monitor the pattern of dye movement through their xylem con- Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 perature influence between 10 8C and 50 8C and <2.1% deviation ÿ1 duits (tracheids with spiral thickenings; Pratt, 1971; Findlay et al., from the specified displacement of 200 mV mm . It was also tested 1987). The berries were visually rated for the progression of skin whether the force exerted on the berries by the spring-loaded trans- pigmentation (green, blush, pink, red, purple, blue) before being ducers (0.40 N at the beginning of stroke to 0.75 N at the end of sectioned longitudinally through the centre for light microscopy. The stroke) would lead to fruit deformation. Two sensors were placed extent of dye movement into the berry from the shoot was rated simultaneously at a 908 angle on mature, detached Concord berries: visually and assigned a number from 1 (in brush only) through 6 one sensor with high force (as indicated by higher voltage, greater (continuous throughout the entire vascular network). The berry juice sensor compression) and one with low force. There clearly was was then expressed to determine concentration. a directional decrease in diameter in the direction of the stronger To test whether water can move back from grape berries to the sensor. Over a 21 d period, the high-force sensors indicated a diameter parent vine (‘xylem backflow’), dye was also fed to berries on intact, loss of 17.461.6% (n=4) and the low-force sensors a loss of only well-watered Merlot and Concord vines before and after veraison. 2.061.2%. In no case did the weaker sensor register an increase in The root pressure chamber was used to bring the plant to full diameter, implying that the berries were not pressed into an oval hydration. While the chamber was under pressure, the stylar end of shape by the stronger sensor. It was therefore assumed that the sensor selected berries was removed with a razor blade to expose the output was a reasonable reflection of fruit volume changes. At the peripheral (dorsal) and axial (ventral) vascular bundles, and the cut end of each experiment, the selected berries were removed and the end of each berry was placed in 0.1% basic fuchsin (‘reverse’ dye solute concentration (8Brix) recorded by refractometry. infusion). The pressure, which served to avoid cavitation in the berry vasculature due to the cut, was then released and the dye was Root pressurization experiments allowed to move back through the berry. The experiment was re- peated with a different set of Merlot and Concord vines without When pressure is applied to a plant’s root system, the entire xylem pressurizing the root system, similar to an experiment with im- becomes pressurized until the pressure at the surface reaches Px >0 mature (green) tomato (Malone and Andrews, 2001). After ;3 h, the (atmospheric pressure, see Wei et al., 1999). If the pressure is clusters were removed, and cross-sections of the pedicel (proximal) increased to a point where xylem flow exceeds transpiration, water end of the treated berry, the pedicel, the rachis, other berries on the floods the apoplast and drips from the hydathodes. A custom-built same cluster, and the shoot above and below the cluster were 26.0 l metal root pressure chamber, based on the design by examined for dye movement by light microscopy. Yong et al. (2000), was used to determine the effect of rapid changes in plant water status on berry size. The chamber was designed with a split lid such that the entire pot (i.e. root system) could be placed inside the chamber and the lid sealed around either side of the stem. Pneumatic pressure was applied to the chamber under manual control Results with nitrogen. The appearance of xylem sap at the surface of a cut petiole above the cluster was used to indicate full hydration at Dry-down experiments that position on the shoot. Prior to veraison, the pressure was in- The average solute concentration at the end of each ex- creased in 0.2 MPa increments up to 1.0 MPa. Berries at the apical, periment was 5.1 8Brix (early pre-veraison Merlot only), middle, and basal part of the fruit cluster were monitored using the transducers described above, with chamber pressure logged auto- 7.2 8Brix (pre-veraison Merlot), 8.5 8Brix (pre-veraison matically at the same time interval. Additional FI-XSM sensors Concord), 21.1 8Brix (post-veraison Merlot), and 20.0 8Brix were used to monitor the trunk diameter and shoot node diameter at (post-veraison Concord). During the second interval (34 d), the point of cluster attachment. The plants were not watered during Merlot berries expanded ;1.6-fold, gaining ;16 lldÿ1 in the 2 d prior to root pressurization, and the leaf above the cluster did volume and ;0.4 8Brix dÿ1 in solutes (;22 mM hexoses not reach full hydration (W 0 MPa) until the chamber pressure leaf dÿ1). Similarly, Concord berries grew ;1.4-fold and reached 1.0 MPa. ÿ1 ÿ1 For the post-veraison measurements, the procedure was simpli- gained ;15 lld and ;0.3 8Brix d . Grape berries of fied to a rapid one-step increase in root pressure to bring the plant both cultivars reacted similarly to changes in soil moisture to full hydration within a period of 6 min. This pressure was held and plant water status. However, the response of pre- and for ;4 h with adjustments made manually to maintain sap flow post-veraison berries differed in two important ways. First, from a cut petiole above the cluster. The diameter of two berries on each of two clusters per plant was recorded using FI-XSM sensors. pre-veraison berries were able to maintain their diameter Berry sugar concentration (8Brix) was subsequently determined by (and thus their volume) for several days into the dry- refractometry. down period, after which they shrank rapidly (Fig. 1A). 2580 Keller et al. Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021

Fig. 1. Change in diameter of pre-veraison (A) and post-veraison (B) Merlot and Concord grape berries during a soil dry-down and rewatering cycle. The Merlot and Concord STD treatments were previously grown with abundant water supply. The Merlot DI treatment had been subject to deficit irrigation for 4 weeks after flowering. Text labels indicate the treatment ranking (highest to lowest) of berry size at the point of Fig. 2. Relationship between grape berry size and the fraction of rewatering (*indicates onset of veraison for a single berry from the Merlot transpired soil water (FTSW) during a soil dry-down period of pre- STD treatment). veraison (A) and post-veraison (B) Merlot and Concord grapevines. The Merlot and Concord STD treatments were previously grown with abundant water supply. The Merlot DI treatment had been subject to deficit irrigation for 4 weeks after flowering. Means 6SE (n=3; except After veraison, the berry diameter appeared to decrease for pre-veraison Merlot STD: n=2) are shown. slightly but gradually from the first day (Fig. 1B). When normalized against FTSW, pre-veraison berries suddenly A second major difference between pre- and post- began to shrink after 80% of the transpirable soil water veraison fruit was the berry response to rewatering at the had been used, regardless of cultivar and irrigation history end of the dry-down period. In the early pre-veraison (Fig. 2A). After veraison, the decrease in diameter as a experiment, berry diameter recovered entirely within 12 h function of FTSW was more gradual, but occurred over of rewatering (data not shown). In the second experiment, the entire range of transpirable soil water (Fig. 2B). Never- where the fruit was closer to veraison at the start of the theless, the volume loss also intensified beyond 80% dry-down period, the recovery after rewatering was less FTSW, although the extent of shrinkage was much less rapid, especially for Merlot STD (Fig. 1). Nevertheless, severe than before veraison. During the earliest dry-down recovery continued throughout the day, regardless of the experiment, conducted well before veraison, there was a diurnal fluctuation in vapour pressure deficit (VPD). One pronounced diurnal pattern of diameter decrease during the of the berries on the Merlot STD vine began to change day, when evaporative demand of the air was high (high colour from green to red (i.e. it underwent veraison) 5 d into VPD), and partial recovery overnight (data not shown). the dry-down period and simultaneously began to increase However, during the second dry-down experiment, the in size, even before the vine was rewatered (Fig. 1A). All berries were close to veraison (in the lag-phase of growth) berries had shrunk at this point, and the size increase was and did not show the same diurnal fluctuations in size due to the berry returning to full turgor as it began to ripen. as water stress became more severe (Fig. 1A). In particular, In the post-veraison experiment, the daily decrease in berry berry diameter failed to recover overnight, but still de- diameter was stopped by rewatering, but the berries did creased during the day. After veraison, the berries were not return to their initial size (Fig. 1B). even less responsive to plant water stress, with only a small Overall, the response of the Merlot STD, Merlot DI, and decrease in diameter during the day and no overnight Concord STD berries to the dry-down and rewatering recovery (Fig. 1B). cycles was very similar. Merlot DI plants used more water Grape berry hydraulics 2581 and took longer than Merlot STD to respond during the applied to the roots (Fig. 3). The subsequent increase in dry-down period. This was not related to differences in berry diameter appeared more buffered from changes in plant leaf area (means, n=3: Merlot STD 0.69 m2, Merlot plant water status than the trunk, and increased gradually DI 0.68 m2, Concord STD 0.76 m2), but instead appeared as the chamber pressure was stepped up. The position of to be due to the DI vines having a deeper and 25% larger the berry on the cluster also appeared to influence its re- root system (Merlot STD 151 g, Merlot DI 188 g, Concord sponse to the increase in root pressure; the diameter of STD 182 g) which enabled greater water extraction from the apical, middle, and basal berries increased by 0.8, 1.0, the pots. Plant water status was not monitored during the and 1.4%, respectively, during the pressurization period. dry-down cycles to avoid increasing defoliation over time. Assuming a spherical berry shape, this was equivalent However, midday stomatal conductance to water vapour to a volume increase of 27, 34, and 82 ll, respectively. (gs, measured using a PLC-6 broadleaf chamber connected When the chamber pressure was released, berry diameter to a CIRAS-2 from PP Systems, Amesbury, MA, USA) of began decreasing almost immediately, but the rate was less the leaf opposite the cluster fell from ;300 to <50 mmol rapid than for the trunk and node at the point of cluster Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 mÿ2 sÿ1 during the course of a typical dry-down cycle. attachment. The volume loss amounted to 26, 29, and 70 ll Similar cycles with potted pre-veraison Concord plants led for the apical, middle, and basal berry, respectively, by to Wleaf –1.6 MPa, which was associated with stomatal dusk, after which there was no further decrease in berry closure and berry shrinkage, and complete recovery over- volume. These observations suggest that the change in berry night upon rewatering (Liu et al., 1978). In addition, whole- diameter is a consequence of changes in cellular water up- plant transpiration decreased progressively with increasing take and turgor, rather than the more direct changes in xylem duration of water stress (data not shown). Transpiration diameter that may have been responsible for the rapid re- is closely coupled with gs which, in turn, is tightly linked sponse of trunk diameter to changes in root pressurization. to leaf water status. For the post-veraison pressure runs, vines were watered prior to being sealed in the chamber. Consequently, the Root pressurization experiments pressure required to bring the plants to Wleaf 0 MPa was lower than before veraison. The response of berries to root Pre-veraison Concord berries responded readily to changes pressurization was somewhat variable: the diameter of in plant water status imposed by changing the pressure some berries did not change, while other berries decreased applied to the root system. In the absence of pressure, berry in diameter by <0.2 mm during the measurement period diameter decreased during the day, while plants were tran- (Fig. 4). This reaction was not related to the solute con- spiring rapidly. Trunk and shoot node diameter reacted centration of these berries, which ranged from 15.5 to almost instantaneously to step-wise increases in root 24.2 8Brix, nor to berry diameter, which varied from 11.8 to pressure, and then dropped rapidly as the pressure was 18.5 mm (data not shown). However, in no case did root returned to zero (Fig. 3). The trunk diameter was particu- pressurization result in a measurable increase in berry larly responsive, indicating rapid transmission of the diameter. For some berries, the rate of decline in diameter chamber pressure to a change in trunk water capacitance. was slowed by bringing the plant to full hydration and The berry response was initially less obvious, with the then accelerated again as the pressure was released. Earlier first increase in diameter seen 1 h after pressure was first work in our laboratory (unpublished) had shown that it was possible to crack mature (22–32 8Brix) Concord berries on intact vines, using the root pressure chamber, and this was verified in the present study (Fig. 4C). There was no measurable increase in berry diameter leading up to the berry cracking, and cracking did not appear to be due to the force exerted by the sensor as several other berries on the vine also cracked. In addition, spontaneous cracking occurred during humid nights following irrigation.

Dye movement experiments Basic fuchsin clearly behaved as a xylem lumen-mobile dye, but not as a general apoplastic tracer. The dye always remained confined to xylem vessels and tracheids, regard- less of whether it was fed via the shoot base (forward infusion) or via the berries (reverse infusion). On shoots Fig. 3. Effect of stepwise increases in pneumatic pressure applied to the with clusters bearing only green (i.e. pre-veraison) berries, root system of pre-veraison Concord grapevines. Changes in diameter of three berries on a single cluster, the node at the attachment point of the dye was consistently found to be continuous through- the cluster, and the trunk were recorded with electronic transducers. out the entire vascular network, including the ovular, 2582 Keller et al. Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021

Fig. 5. Relationship between grape maturity and location inside the berry of the xylem-mobile dye basic fuchsin (1, in brush only; 6, throughout the entire vascular network) fed through the cut shoot base to attached Merlot clusters undergoing veraison (changing colour from green to blue; n=44).

blue berries (>15 8Brix), it was confined to the brush re- gion. This pattern was independent of initial berry size or volume change during dye infusion (r=ÿ0.03, P=0.92, n=44) and was unaffected by the duration (up to 3 d) of dye uptake by the shoot. Neither removing the stylar end of the berry nor pedicel girdling had any effect on the extent of dye penetration into the berry. Feeding basic fuchsin to additional shoots or excised clusters of V. vinifera cvs Chardonnay, Cabernet franc, Cabernet Sauvignon, Muscat, and Sangiovese, and V. labruscana cv. Niagara growing in the same vineyard always gave the same results; the dye moved into the berries and throughout their vascu- lar network prior to veraison, but only into the brush of post-veraison berries. Regardless of cultivar and develop- Fig. 4. Change in diameter of post-veraison grape berries and associated mental status, the dye always moved through the pedicel changes in stem water potential (Ystem) in response to rapid changes in the pneumatic pressure applied to the root system. The Merlot and Concord and the vascular traces in the receptacle. STD treatments were previously grown with abundant water supply. The In a reversal of this standard procedure, basic fuchsin Merlot DI treatment had been subject to deficit irrigation for 4 weeks after flowering. Means 6SE (n=4) are shown, except for Concord, where one was also fed to the cut stylar end of berries on vines subject berry cracked during the measurement period and is shown individually. to root pressurization. Prior to veraison, root pressuriza- tion resulted in xylem sap exudation from the severed xylem ends in the berries upon excision of their stylar ventral (axial), and dorsal (peripheral) vascular bundles, end. However, after veraison, such sap exudation could and the dorsal network, giving the berries a characteristic never be observed. When the pressure on the root system ‘chicken-wire’ appearance. In clusters undergoing verai- was released, the dye rapidly moved back through the berry son, dye movement into the berries did not stop suddenly tracheids and out through the pedicel of both pre- and at a certain developmental stage. Instead, the extent of post-veraison Merlot and Concord berries (Fig. 6). Dye dye penetration through a berry’s vascular system de- movement through berries was identical when roots were creased gradually as the berry changed colour from green not pressurized, indicating that root pressurization to pre- to blue and began to accumulate sugar (Fig. 5). As soon vent xylem cavitation prior to dye infusion was unneces- as berries began to blush, the dye became discontinu- sary. In both cases, movement of dye could be traced to ous in the stylar end, while still moving into the axial and adjacent and nearby transpiring leaves, but also down- peripheral vascular bundles. With increasing pigmentation ward (basipetally, i.e. against the transpiration stream) in and sugar concentration, the dye was found to penetrate the vessels on the cluster side of the shoot, beyond the progressively less far into the berry until, in purple and most basal leaf, towards the trunk of the vine (Fig. 7). Grape berry hydraulics 2583 Moreover, although dye also always moved to other berries on the same cluster, it moved into those berries (i.e. beyond the brush) only prior to veraison but not afterwards.

Discussion The response of grape berry diameter to changes in soil moisture was similar for two genetically distinct cultivars and independent of the previous irrigation history. The extent of berry shrinkage during soil drying was much more severe before than after veraison, which is consistent with earlier research showing that developing berries become Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 progressively less sensitive to plant water status (Creasy and Lombard, 1993; Greenspan et al., 1994, 1996). More- over, pre-veraison berries rapidly rehydrated and resumed growth upon rewatering, whereas no such response was found in post-veraison berries. Application of irrigation water after veraison merely prevented further shrinkage. In real time, the Merlot DI vines responded more slowly Fig. 6. Movement of the xylem-mobile dye basic fuchsin infused than the STD vines to the dry-down episodes before through the stylar end of attached post-veraison Merlot (A, C) and veraison, but they were able to exploit more soil water Concord (B, D) grape berries. Cross-section of the receptacle/pedicel (A, B) and of the pedicel end of the berry (C, D). Note: the red coloration due to their larger and deeper root system. Thus, plotting outside of the xylem (C, D) is due to anthocyanins rather than basic berry diameter against FTSW showed that pre-veraison fuchsin. berries did not begin to shrink until the vines had transpired 80% of the available soil water. This is remarkably sim- ilar to the rapid drop at FTSW of <20% in grapevine Wleaf and gs (Lebon et al., 2003). Our finding that a shrinking berry undergoing veraison during a dry-down episode sud- denly began to re-expand (before rewatering and while the pre-veraison berries on the same cluster continued to shrink) indicates a sudden increase in phloem influx at veraison. This increase must have exceeded the total water loss from the berry, suggesting that phloem water may be sufficient to sustain pericarp cell expansion in ripening berries. Changes in pneumatic pressure applied to the root sys- tem were readily transmitted to the trunk, as shown by the stepwise and almost immediate changes in trunk diameter. Since pressure applied to the roots is rapidly transmitted to Px (Wei et al., 1999), fluctuations in trunk diameter may be a good indicator of changes in Px. The fact that changes in grape berry diameter before veraison occurred much more gradually compared with the trunk suggests that they were a direct result of water influx and efflux. By contrast, when root pressure was applied after veraison, it was at best able to offset the decrease in berry diameter during the day. This is consistent with the effect of rewatering following a post-veraison dry-down episode and suggests that water may continue to flow through the berry xylem Fig. 7. Movement of the xylem-mobile dye basic fuchsin infused after veraison. However, the general direction of water flow through the stylar end of attached Concord grape berries. The photos on may be from the fruit back to the leaves, especially during the right labelled 1–4 show cross-sections corresponding to the positions periods of high evaporative demand and low P . Green- 1–4 on the left. The red dye is clearly visible in the peduncle (1 and 3), x a non-treated berry adjacent to the one used for dye infusion (2), and the span et al. (1994, 1996) claimed that xylem backflow shoot below the treated clusters (4). from ripening grape berries was likely to be insignificant, 2584 Keller et al. although Lang and Thorpe (1989) had reported consistent sugar from the mesocarp apoplast (Sarry et al., 2004) and substantial backflow. occurs at a faster rate than water influx into mesocarp When the cut end of the peduncle of whole clusters cells (as indicated by increasing hexose concentration), an was immersed in basic fuchsin, dye always ended up at excess of phloem water influx during ripening should be each berry within ;30 min, covering a distance of 50– expected (which is supported by the present results). 100 mm from the point of infusion. However, dye pen- Dissipation of surplus phloem water by berry transpiration etration into the berries declined once they started to may not be a sufficiently robust strategy, because transpi- accumulate hexose sugars, and was confined to the brush ration is highly dependent on VPD and declines during (proximal) area after they had reached >15 8Brix. This development (Rogiers et al., 2004). Indeed, grape berries agrees well with the increase in ‘pedicel equilibrium seem to be ‘designed’ to minimize evaporative water loss; pressure’ during ripening (Tyerman et al., 2004) and with compared with leaves, they have very few stomata that all studies using the standard forward dye infusion (re- become non-functional by veraison (Blanke et al., 1999; viewed by Ollat et al., 2002), but contrasts sharply with Rogiers et al., 2004), and the amount of (epi-)cuticular Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 the dye movement out of the berries during reverse infu- wax is ;10-fold greater (Radler, 1965; Possingham et al., sion in both cultivars. Intriguingly, the decline in dye pene- 1967). Therefore, excess phloem water (that has moved tration coincided with the abrupt reversal of the shrinking to the apoplast osmotically) may be recycled back to the trend of berries that began ripening while the plant suffered leaves in the xylem, at least partly without entering the increasingly severe water stress. Taken together, these mesocarp symplast. Recirculation of excess phloem water results suggest that (slow) transpiration from the pedicel was originally proposed by Mu¨nch (1930) and has been and receptacle, in addition to water transfer to the transport suggested for various fruits, including grape (Lang phloem (Thompson and Holbrook, 2003), may provide and Thorpe, 1989), apple (Lang, 1990), and tomato (Van the driving force for dye (i.e. water) movement towards Ieperen et al., 2003). Lang and Thorpe (1989) assumed the the berries, while dye movement into the berries may be apoplastic pressure in the fruit to be higher than in the inhibited by positive Px inside the ripening berries. Thus, plant, and this was recently confirmed by direct measure- although the berry tracheids apparently remain intact after ments of Px in grape berry pedicels (Tyerman et al., 2004). veraison, grapevines seem to discontinue (or at least Thus phloem influx may cause (positive) ‘back pressure’ substantially diminish) the use of this pathway for water leading to xylem efflux due to the limited extensibility influx, beginning at the stylar (distal) end as the berries of the skin (Matthews et al., 1987), especially if phloem begin to ripen. Our data demonstrate that ripening berries water influx exceeds berry growth and transpiration. This of the genus Vitis remain hydraulically connected to the pressure appears to dissipate in the pedicel, so that xylem- shoot, confirming recent results by Bondada et al. (2005). mobile dyes can move to (but not into) the berry. Indeed, Both physical disruption and air embolisms (cavitation) some of the water exiting the berry may be reabsorbed in can therefore be excluded as possible culprits for the the pedicel and beyond by the incoming phloem to maintain alleged xylem discontinuity and hydraulic isolation of the water potential equilibrium between the transport phloem berries (see Introduction), and also rule out the existence and the apoplast (Thompson and Holbrook, 2003). of axial, apoplastic, solute-reflecting barriers to the move- The key to the formation of berry–xylem back-pressure ment of dye (but not water) through the berry vasculature. could lie in the accumulation of solutes in the berry Though such semipermeable apoplastic ‘membranes’ have apoplast. In roots, release of sugar into the apoplast leads been conjectured (Bradford, 1994), the idea is at odds to positive hydrostatic Px due to osmotic water influx with our observation that reverse-infused dye readily from the soil, which forces water up the plant (Scholander moved out of the berries. et al., 1955; Sperry et al., 1987). A similar mechanism, in Grape berries are thought to switch from symplastic to which both apoplastic sugar and water are derived from apoplastic phloem unloading at veraison (Xia and Zhang, phloem influx, could not only greatly reduce xylem influx 2000), a change that they share with developing tomato into post-veraison berries, but even reverse the flow dir- fruits (Ruan and Patrick, 1995). The deposition of solutes ection. The increasing pmc (=RTCs) during ripening may in the fruit apoplast raises the apoplast’s osmotic pressure not translate into an increasing Pmc (Thomas et al., 2004) (papoplast), leading to water efflux from the phloem (van Bel, due to the simultaneous accumulation of solutes in the 2003). The resulting decrease in phloem pressure (Pp), in apoplast (Pmcpmc–papoplast) and because pmcpapoplast turn, enhances phloem influx (Patrick, 1997). However, at papoplast >1 MPa (Li and Delrot, 1987). Indeed, px berries require very little water compared with leaves, even (=papoplast, unless the two ‘compartments’ do not equili- during rapid expansion. The demand for growth water brate inside the berry) of grape berry xylem exudate was ÿ1 per Merlot berry in this study was <50 lld (over a 30 d found to increase in step with pmc during ripening (Lang post-veraison period), and berry transpiration has been and Du¨ring, 1991). At the same time, the high papoplast ÿ1 estimated at <100 lld (Greenspan et al., 1996; Rogiers would also raise Px (=Wmc+px). The proposed mechanism et al., 2004). Assuming that active removal of unloaded for the generation of hydrostatic xylem back-pressure in Grape berry hydraulics 2585 grape berries would not only account for the observed hexose transporters (Fillion et al., 1999), and aquaporins pattern of dye movement after veraison but would also (Picaud et al., 2003), is incompatible with the idea of com- explain our inability to increase berry diameter (or tomato partmentation breakdown. Membrane disintegration should fruit diameter in the study of Davies et al., 2000) by result in rapid downregulation of membrane-associated root pressurization. The inverse relationship between dye proteins. Thus, solute permeability of the plasma membrane penetration into berries and sugar concentration (rather and tonoplast may differ for influx and efflux, since solutes than just the amount of sugar per berry) and the lack of are actively ‘pumped’ into the vacuoles and retained correlation with berry size in this study are consistent there, while water can diffuse in and out down the DW. with increases in papoplast and Px during ripening. Nevertheless, vacuolar water is likely to be relatively well Estimations of net water flow into grape berries have protected from being pulled back into the berry tracheids usually been based on the assumption that pedicel girdling and shoot xylem vessels. The rising solute concentration eliminates phloem flow without affecting xylem flow inside the vacuoles decreases Wmc (=Pmc–RTCs) and makes (Lang and Thorpe, 1989; Greenspan et al., 1994, 1996). it more and more difficult for the leaves to ‘extract’ water Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 However, interruption of phloem influx would render the from the ripening berries. Moreover, since water moving direction of xylem flow independent of phloem flow and out of the pericarp has to cross both the tonoplast and could effectively remove the main driving force for xylem plasma membrane, closure of aquaporins upon loss of efflux, which would result in systematic underestimations turgor (Chrispeels et al., 1999) would greatly increase rh of both phloem influx and xylem efflux. Alternatively, across the pericarp cell membranes. Nevertheless, pro- partial occlusion of xylem conduits in girdled pedicels longed plant water stress coupled with declining phloem could increase rh (Zwieniecki et al., 2004). However, there influx and continued (albeit very slow) cuticular berry is no need to postulate hydraulic isolation of ripening transpiration (Rogiers et al., 2004) may lead to slow water fruit to account for any of our observations. Water recycling loss from the berries. Therefore, berries may shrink if via the xylem would make fruits less responsive to plant phloem influx becomes too slow to balance water efflux. water status (Van Ieperen et al., 2003) and could decrease their vulnerability to cracking by serving as an ‘overflow’ mechanism. Cracking probably ensues when Pmc exceeds Acknowledgements the cell wall resistance of skin cells. Thus, cracking due to root pressurization indicates that under conditions The authors acknowledge the financial support from the USDA favouring high P (e.g. warm rainy or humid nights), water Northwest Center for Small Fruits Research. The valuable input by x Lynn Mills and John Ferguson at various stages during this study backflow from the berry slows or ceases, so that phloem is greatly appreciated. influx >xylem efflux+berry transpiration. The cracking of post-veraison Concord (but not Merlot) berries under root pressurization without any perceptible increase in References diameter confirms that fruit expansion is limited by the elastic properties of the skin cell walls (Considine and Ageorges A, Issaly N, Picaud S, Delrot S, Romieu C. 2000. Kriedemann, 1972; Matthews et al., 1987). Identification and functional expression in yeast of a grape berry One implication of our results is that, in addition to sucrose carrier. and Biochemistry 38, 177–185. Blanke MM, Pring RJ, Baker EA. 1999. Structure and elemental xylem conduits, membrane integrity, and thus cellular composition of grape berry stomata. Journal of Plant Physiology compartmentation, also apparently remain intact in ripening 154, 477–481. grape berries. Although this contradicts the conclusion Bondada BR, Matthews MA, Shackel KA. 2005. Functional reached by Lang and Thorpe (1989) and Lang and Du¨ring xylem in the post-veraison grape berry. Journal of Experimental (1991), their results are actually consistent with apoplastic Botany 56, 2949–2957. Bradford KJ. 1994. Water stress and the water relations of seed phloem unloading, and continued hydraulic connection development: a critical review. Crop Science 34, 1–11. and cell compartmentation. While our interpretation allows Chrispeels MJ, Crawford NM, Schroeder JI. 1999. Proteins for for some diffusional loss of solutes from the berries, the transport of water and mineral nutrients across the membranes of ‘damage’ should be small, since the proportion of the plant cells. The Plant Cell 11, 661–675. berry volume occupied by the apoplast is very small Considine JA, Kriedemann PE. 1972. Fruit splitting in : determination of the critical turgor pressure. Australian Journal of (Hardie et al., 1996; Diakou and Carde, 2001). Solutes Agricultural Research 23, 17–24. leaking back to the pedicel also could be actively reab- Creasy GL, Lombard PB. 1993. Vine water stress and peduncle sorbed ‘en route’ by the transport phloem (Patrick and girdling effects on pre- and post-veraison grape berry growth and Offler, 1996; van Bel, 2003). Moreover, diffusion would be deformability. American Journal of Enology and Viticulture 44, far slower than active removal into the sink symplast 193–197. Creasy GL, Price SF, Lombard PB. 1993. Evidence for xylem by transporters and channels. In fact, the expression in discontinuity in Pinot noir and Merlot grapes: dye uptake and ripening grape berries of membrane proteins, such as mineral composition during berry maturation. American Journal sucrose (Davies et al., 1999; Ageorges et al., 2000) and of Enology and Viticulture 44, 187–192. 2586 Keller et al. Davies C, Wolf T, Robinson SP. 1999. Three putative sucrose Matthews MA, Cheng G, Weinbaum SA. 1987. Changes in water transporters are differentially expressed in grapevine tissues. Plant potential and dermal extensibility during grape berry development. Science 147, 93–100. Journal of the American Society for Horticultural Science 112, Davies WJ, Bacon MA, Thompson DS, Sobeih W, Rodriguez LG. 314–319. 2000. Regulation of leaf and fruit growth in plants growing in Mu¨nch E. 1930. Die Stoffbewegungen in der Pflanze. Jena, drying soil: exploitation of the plants’ chemical signalling system Germany: Fischer. and hydraulic architecture to increase the efficiency of water use in Ollat N, Diakou-Verdin P, Carde JP, Barrieu F, Gaudille`re JP, agriculture. Journal of Experimental Botany 51, 1617–1626. Moing A. 2002. Grape berry development: a review. Journal Diakou P, Carde JP. 2001. In situ fixation of grape berries. International des Sciences de la Vigne et du Vin 36, 109–131. Protoplasma 218, 225–235. Patrick JW. 1997. Phloem unloading: sieve element unloading and Drazˇeta L, Lang A, Hall AJ, Volz RK, Jameson PE. 2004. Causes post-sieve element transport. Annual Review of Plant Physiology and effects of changes in xylem functionality in apple fruit. Annals and Plant Molecular Biology 48, 191–222. of Botany 93, 275–282. Patrick JW, Offler CE. 1996. Post-sieve element transport of Du¨ring H, Lang A, Oggionni F. 1987. Patterns of water flow in photoassimilates in sink regions. Journal of Experimental Botany

Riesling berries in relation to developmental changes in their 47, 1165–1177. Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021 xylem morphology. Vitis 26, 123–131. Picaud S, Becq F, De´dalde´champ F, Ageorges A, Delrot S. 2003. Esau K. 1977. Anatomy of seed plants, 2nd edn. New York: John Cloning and expression of two plasma membrane aquaporins Wiley & Sons. expressed during the ripening of grape berry. Functional Plant Fillion L, Ageorges A, Picaud S, Coutos-The´venot P, Lemoine R, Biology 30, 621–630. Romieu C, Delrot S. 1999. Cloning and expression of a hexose Possingham JV, Chambers TC, Radler F, Grncarevic M. 1967. transporter gene expressed during the ripening of grape berry. Cuticular transpiration and wax structure and composition of Plant Physiology 120, 1083–1093. leaves and fruit of Vitis vinifera. Australian Journal of Biological Findlay N, Oliver KJ, Nii N, Coombe BG. 1987. Solute Sciences 20, 1149–1153. accumulation by grape pericarp cells. IV. Perfusion of pericarp Pratt C. 1971. Reproductive anatomy in cultivated grapes—a review. apoplast via the pedicel and evidence for xylem malfunction in American Journal of Enology and Viticulture 22, 92–109. ripening berries. Journal of Experimental Botany 38, 668–679. Radler F. 1965. Reduction of loss of moisture by the cuticle wax Greenspan MD, Schultz HR, Matthews MA. 1996. Field evalu- components of grapes. Nature 207, 1002–1003. ation of water transport in grape berries during water deficits. Rogiers SY, Hatfield JM, Jaudzems VG, White RG, Keller M. Physiologia Plantarum 97, 55–62. 2004. Grape berry cv. Shiraz epicuticular wax and transpiration Greenspan MD, Shackel KA, Matthews MA. 1994. Developmen- during ripening and preharvest weight loss. American Journal of tal changes in the diurnal water budget of the grape berry exposed Enology and Viticulture 55, 121–127. to water deficits. Plant, Cell and Environment 17, 811–820. Rogiers SY, Smith JA, White R, Keller M, Holzapfel BP, Hardie WJ, O’Brien TP, Jaudzems VG. 1996. Morphology, Virgona JM. 2001. Vascular function in berries of Vitis vinifera anatomy and development of the pericarp after anthesis in grape, (L) cv. Shiraz. Australian Journal of Grape and Wine Research Vitis vinifera L. Australian Journal of Grape and Wine Research 7, 247–51. 2, 97–142. Ruan YL, Patrick JW. 1995. The cellular pathway of postphloem Ho LC, Grange RI, Picken AJ. 1987. An analysis of the sugar transport in developing tomato fruit. Planta 196, 434–444. accumulation of water and dry matter in tomato fruit. Plant, Cell Sarry JE, Sommerer N, Sauvage FX, Bergoin A, Rossignol M, and Environment 10, 157–162. Albagnac G, Romieu C. 2004. Grape berry biochemistry revisited Keller M, Koblet W. 1995. Dry matter and leaf area partitioning, upon proteomic analysis of the mesocarp. Proteomics 4, 201–215. bud fertility and second-season growth of Vitis vinifera L.: Scholander PF, Love WE, Kanwisher JW. 1955. The rise of sap responses to nitrogen supply and limiting irradiance. Vitis in tall grapevines. Plant Physiology 30, 93–104. 34, 77–83. Sperry JS, Holbrook NM, Zimmermann MH, Tyree MT. 1987. Lang A. 1990. Xylem, phloem and transpiration flows in developing Spring filling of xylem vessels in wild grapevine. Plant Physiology apple fruits. Journal of Experimental Botany 41, 645–651. 83, 414–417. Lang A, Du¨ring H. 1991. Partitioning control by water potential Thomas TR, Shackel K, Matthews MA. 2004. Direct measurement gradient: evidence for compartmentation breakdown in grape of berry turgor pressure in Vitis vinifera: relationship to vine water berries. Journal of Experimental Botany 42, 1117–1122. status and deformability, and evolution throughout development. Lang A, Thorpe MR. 1989. Xylem, phloem and transpiration flows Proceedings of the 7th International Symposium on Grapevine in a grape: application of a technique for measuring the volume Physiology and Biotechnology. Davis, CA, USA, pp. 35. of attached fruits to high resolution using Archimedes’ principle. Thompson MV, Holbrook NM. 2003. Scaling phloem transport: Journal of Experimental Botany 40, 1069–1078. water potential equilibrium and osmoregulatory flow. Plant, Cell Lebon E, Dumas V, Pieri P, Schultz HR. 2003. Modelling the and Environment 26, 1561–1577. seasonal dynamics of the soil water balance of vineyards. Tyerman SD, Tilbrook J, Pardo C, Kotula L, Sullivan W, Functional Plant Biology 30, 699–710. Steudle E. 2004. Direct measurement of hydraulic properties in Li ZS, Delrot S. 1987. Osmotic dependence of the transmembrane developing berries of Vitis vinifera L. cvs Shiraz and Chardonnay. potential difference of broadbean mesocarp cells. Plant Physiology Australian Journal of Grape and Wine Research 10, 170–181. 84, 895–899. van Bel AJE. 2003. The phloem, a miracle of ingenuity. Plant, Cell Liu WT, Pool R, Wenkert W, Kriedemann PE. 1978. Changes in and Environment 26, 125–149. photosynthesis, stomatal resistance and of Vitis Van Ieperen W, Volkov VS, Van Meeteren U. 2003. Distribution labruscana through drought and irrigation cycles. American of xylem hydraulic resistance in fruiting truss of tomato in- Journal of Enology and Viticulture 29, 239–246. fluenced by water stress. Journal of Experimental Botany 54, Malone M, Andrews J. 2001. The distribution of xylem hydraulic 317–324. resistance in the fruiting truss of tomato. Plant, Cell and Wei C, Tyree MT, Steudle E. 1999. Direct measurement of Environment 24, 565–570. xylem pressure in leaves of intact maize plants. A test of the Grape berry hydraulics 2587 cohesion–tension theory taking hydraulic architecture into consid- Zhang LY, Peng YB, Pelleschi-Travier S, Fan Y, Lu YF, eration. Plant Physiology 121, 1191–1205. Lu YM, Gao XP, Shen YY, Delrot S, Zhang DP. 2004. Xia GH, Zhang DP. 2000. Intercellular symplastic connection and Evidence for apoplasmic phloem unloading in developing apple isolation of the unloading zone in flesh of the developing grape fruit. Plant Physiology 135, 574–586. berry. Journal of Integrative Plant Biology 42, 898–904. Zwieniecki MA, Melcher PJ, Feild TS, Holbrook NM. 2004. Yong JWH, Wong SC, Letham DS, Hocart CH, Farquhar GD. A potential role for xylem–phloem interactions in the 2000. Effects of elevated [CO2] and nitrogen nutrition on hydraulic architecture of trees: effects of phloem girdling cytokinins in the xylem sap and leaves of cotton. Plant Physiology on xylem hydraulic conductance. Tree Physiology 24, 124, 767–779. 911–917. Downloaded from https://academic.oup.com/jxb/article/57/11/2577/676356 by guest on 27 September 2021