Influence of a drying cycle on post-drought xylem sap abscisic acid and stomatal responses in young temperate deciduous angiosperms

Authors

  • Nancy J. Loewenstein,

    1. Department of Forestry, 203 Anheuser-Busch Natural Resources Building, University of Missouri, Columbia, MO 65211, USA;
    2. present address, School of Forestry and Wildlife Sciences, Auburn University, Auburn, AL 36849-5418 USA
    Search for more papers by this author
  • Stephen G. Pallardy

    Corresponding author
    1. Department of Forestry, 203 Anheuser-Busch Natural Resources Building, University of Missouri, Columbia, MO 65211, USA;
    Search for more papers by this author

Author for correspondence: Stephen G. PallardyFax: +1 573 882 1977Email: PallardyS@missouri.edu

Summary

  • •   Post-drought patterns of water relations and gas exchange were studied in relation to xylem sap abscisic acid (ABA) concentration during recovery for young plants of five woody species. The potential role of xylem sap [ABA] in these responses was the object of study.
  • •   Potted plants were allowed to deplete soil water and then were rewatered. At peak drought and during recovery, predawn and midday leaf water potential (Ψl), stomatal conductance (gs), and xylem sap [ABA] were measured.
  • •   Water potentials recovered rapidly after rewatering but stomatal re-opening was delayed. Xylem sap [ABA] was elevated early in recovery and might have affected stomatal opening, but after 1 d at high soil water content [ABA] in recovering plants was equal to or lower than in control plants. Stomata appeared to be more sensitive to xylem sap [ABA] in recovering than droughted plants.
  • •   Xylem sap [ABA] may play some role in delayed recovery of stomatal opening after drought, but may not completely explain the responses.

Introduction

While the recovery phase of leaf level response to drought is physiologically and ecologically significant, it has received far less attention than the period of developing drought. The degree and rate of return of processes to pre-stress levels varies among species and can have a substantial impact on plant fitness (Quick et al., 1992; Roe et al., 1995). Recovery of stomatal opening after relief from drought is often not immediate and may vary among species (Davies & Kozlowski, 1977; Ni & Pallardy, 1992). For example, after 5 d at high soil water content following severe drought, stomatal conductance (gs) of sugar maple (Acer saccharum Marsh.) and black walnut (Juglans nigra L.) seedlings had not recovered to pre-stress levels, whereas gs had fully recovered in seedlings of post and white oak (Quercus stellata Wangenh. and Quercus alba L., respectively) (Ni & Pallardy, 1992). A similar uncoupling of stomatal aperture and soil water availability in plants recovering from drought has been reported elsewhere in woody (Pezeshki & Hinckley, 1982; Jackson et al., 1995; Ruiz-Sánchez et al., 1997) and herbaceous plants (Fischer et al., 1970; Beardsell & Cohen, 1975; Correia & Pereira, 1995). The mechanism(s) causing this delay in full stomatal recovery is not well understood. Several early experiments indicated that bulk leaf abscisic acid (ABA) concentration often recovers to pre-drought levels before complete stomatal recovery (Beardsell & Cohen, 1975; Bengston et al., 1977; Dörffling et al., 1977), although there may some persistent effects of drought on apoplastic vs symplastic partitioning of leaf ABA (Loveys & Kriedemann, 1973; Cornish & Zeevaart, 1985, see also Liu et al., 2001a). Prolonged elevation of xylem sap [ABA] is another potential source of stomatal regulation after relief from drought, as stomatal closure in a variety of herbaceous and woody species has been closely linked to elevated concentrations of ABA in the xylem sap during drought (Zhang et al., 1987; Liang et al., 1996; Loewenstein & Pallardy, 1998a,b; Liu et al., 2001b). This ABA signal may be modulated by numerous environmental and endogenous factors that vary with stress, many of which appear to act through direct and indirect influences on apoplastic pH and consequent partitioning of ABA between apoplastic and symplastic compartments of the leaf (Wilkinson & Davies, 2002). However, Wilkinson & Davies (2002) also noted that xylem sap pH in woody plants is unresponsive to soil drying, suggesting that regulation of stomatal aperture in this group of plants may be more closely controlled by xylem sap [ABA] alone. Previous work in our laboratory supports a central role for xylem sap [ABA] in control of stomatal aperture during drought in forest tree species (Loewenstein & Pallardy, 1998a,b). However, the role of xylem sap ABA in post-recovery delays in stomatal opening remains uncertain (Correia & Pereira, 1994, 1995) and has not been adequately studied in either herbaceous or woody species.

Here we report on the dynamics of post-drought recovery of leaf water potential, stomatal aperture and xylem sap ABA for young plants of five woody temperate deciduous angiosperm species growing in soil that was allowed to dry gradually. We address the possible role of xylem sap ABA in delays in stomatal reopening in species of varying drought tolerance and we identify a post-drought shift in apparent sensitivity of stomata to xylem sap ABA that needs to be con-sidered in models of stomatal regulation. The five species chosen for study –Q. alba, Quercus velutina Lam. (black oak), J. nigra, black willow (Salix nigra Marsh.) and eastern cottonwood (Populus deltoides Bartr. ex Marsh.) – are extensively distributed in the eastern USA and vary widely in drought tolerance. The two upland, drought-tolerant oak species are anisohydric and possess adaptations that confer adaptive fitness in water-limited environments (e.g. deep and extensive root system development, osmotic adjustment, sustained capacity for photosynthesis and resistance to protoplasmic injury) (Bahari et al., 1985; Abrams, 1990). Similar to the oaks, J. nigra trees exhibit a superior capacity to avoid leaf drought (Ginter-Whitehouse et al., 1983) but only when growing on deep, moist soil. Unlike the oaks, J. nigra trees are isohydric and exhibit stomatal and photosynthetic sensitivity (Ni & Pallardy, 1991, 1992) and dehydration injury (Martin et al., 1987) in response to soil water depletion. Salix nigra and P. deltoides are riparian species usually found on wet soils near watercourses (Burns & Honkala, 1990), are isohydric and generally considered drought sensitive.

Materials and Methods

Plant materials

For both experiments, J. nigra, Q. velutina Lam. and Q. alba seeds from open-grown trees were obtained from the Missouri State Tree Nursery (Licking, MO, USA). Hardwood cuttings were collected from S. nigra and P. deltoides trees near Columbia, MO, USA (38° N, 92° W). Seeds and cuttings were planted in 2.5-l (10.5 cm diameter) plastic pots containing a 2 : 1 : 1 (v : v : v) mixture of sand, peat moss, and silt-loam soil. A top-dressing of slow-release fertilizer (14 : 14 : 14 N–P–K, Osmocote, Sierra Chemical Co., San Milpitas, CA, USA) was applied soon after plants were established and modified half-strength Hoagland's solution was applied approximately every 2 wk. Plants were grown in an evaporation-cooled greenhouse under 50% shade. Photoperiod was extended to 14 h with sodium vapor lamps during short-day periods of the year.

Drought cycle induction and physiological measurements

Experiments were conducted separately for each species and were initiated when plants were 3–6 months old. Young plants of each species were randomly assigned to either a drought treatment (n = 24–27) or to a treatment wherein plants were kept well-watered throughout the drought and recovery periods of other plants (n = 12, hereafter called control). Plants were placed on a greenhouse bench in a completely randomized design. Drought was imposed by withholding water until stomatal conductance (gs), measured with a steady-state porometer (LI-1600; Li-Cor, Lincoln, NE, USA), was much reduced and plants were showing midday wilting. At this point, predawn leaf water potential (Ψpd) of the drought-stressed plants was determined with a pressure chamber to provide an estimate of both the soil water potential to which the plants were exposed (Ritchie & Hinckley, 1975; Pallardy et al., 1991) and the severity of the imposed drought.

Upon completion of Ψpd measurements, all but four of the water-stressed plants were rewatered (this day upon which recovery began was designated Day 0). At midday of Day 0, leaf water potential (Ψmd), gs and whole-plant transpiration rate (E) were determined, and xylem sap was collected from the three plants per species that were not rewatered, four randomly selected rewatered plants and two randomly selected control plants. These same measurements, as well as Ψpd, were taken for two control and four rewatered plants 1, 2, 4, 6 and 11 d after watering was resumed. Control and rewatered plants were kept at high soil water content during the recovery period.

The sampling regime for the J. nigra experiment, which was conducted first, was slightly different from that for the other species. Predawn measurements were conducted 1 d before rewatering (Day −1 relative to other experiments); plants were rewatered at predawn on Day 0 and midday measurements were not conducted until Day 1. Upon completion of this initial experiment, it was decided that midday measurements on the day plants were rewatered (Day 0) would be of interest, so the design of the experiment for the other species was altered as described above.

Stomatal conductance measurements were typically conducted under natural lighting; however, when ambient photosynthetic photon flux density (PPFD) was low because of cloud cover, sodium-vapor lamps were turned on approximately 30 min before beginning measurements. Whole-plant E was determined gravimetrically. Before initiating midday measurements, pots were enclosed in plastic bags to eliminate water loss from the soil. Pots were weighed, then reweighed after approximately 30 min.

Upon completion of gs, E and Ψmd measurements, xylem sap was collected from excised stems with a pressure chamber, as previously described (Loewenstein & Pallardy, 1998a; Wartinger et al., 1990). The first microlitre (approximately) of sap was blotted from the stem surface to minimize possible contamination of the sap from injured tissues. Fifty microlitres of sap was collected from each stem, with the exception of some water-stressed plants for which less (or no) sap was available. Sap was immediately frozen in liquid nitrogen and stored at −75°C until analysis (see below). Previous experiments indicated that our collection technique did not significantly influence [ABA] in the xylem sap (Loewenstein & Pallardy, 1998a).

Following collection of xylem sap, leaf area was determined with a LI-3000 Leaf Area Meter (Li-Cor) for calculation of whole-plant E rates.

Analysis of ABA

Concentration of ABA in the xylem sap was determined by enzyme-linked immunoassay (ELISA; Sigma, St Louis, MO, USA) (Loewenstein & Pallardy, 1998a). Dilution/spike recovery tests (Jones, 1987) indicated the presence of nonspecific interference in the sap of all five species. Lyophilization sufficiently reduced this interference that quantification of ABA in samples was possible (Loewenstein & Pallardy, 1998a). Lyophilized samples were reconstituted with 25 mm Tris buffered saline adjusted to pH 7.5 with reagent-grade HCl. All samples were run in duplicate and ABA standards were included, in triplicate, in each assay for the construction of a standard curve.

The ABA flux was calculated by multiplying the xylem sap [ABA] by the whole-plant E after Tardieu et al. (1993).

Statistical analysis

Student's t-test was used to compare means of physiological variables for rewatered and control plants. General patterns of response across species were further investigated by combining probabilities from independent tests of significance for each species, as described by Sokal & Rohlf (1995). In this test, the natural logarithms of individual test probabilities are summed and multiplied by −2. The result is distributed as χ2 with 2k degrees of freedom, where k is the number of tests combined. This procedure (also called meta-analysis) allows overall conclusions to be drawn from separate analyses.

Results

Water relations

Predawn water potential and gs of all study species were reduced in response to the drought treatment (Table 1). Because of plant-to-plant variation in size and transpiration rates, it was not possible to achieve identical reductions of soil water content in all species. However, stomatal closure and midday wilting were observed in the majority of droughted plants in all species before termination of the drought treatment.

Table 1.  Mean stomatal conductance (gs) and predawn water potential (Ψpd) of control (n = 24–27 per species) and water-stressed plants (n = 12 per species) of five species before recovery (± SE)
 Drought treatmentControl
 gs (mmol m−2 s−1) Ψ pd (MPa)gs (mmol m−2 s−1) Ψ pd (MPa)
Salix nigra77.9 ± 16.3 −1.31 ± 0.11498.1 ± 54.1 −0.12 ± 0.02
Populus deltoides73.9 ± 23.4 −1.42 ± 0.65338.5 ± 57.3 −0.12 ± 0.02
Juglans nigra11.9 ± 1.7 −0.92 ± 0.39144.9 ± 15.5 −0.27 ± 0.02
Quercus velutina37.8 ± 6.6 −1.40 ± 0.17  84.6 ± 3.6 −0.17 ± 0.00
Quercus alba42.8 ± 8.9 −1.77 ± 1.19267.3 ± 16.8 −0.08 ± 0.02

In general, Ψl recovered rapidly after droughted plants were provided moist soil. Predawn Ψl of droughted plants recovered to at least control levels within 1 d of rewatering for all five species (Figs 1 and 2, P. deltoides and Q. alba). Further, significant over-recovery of Ψpd was observed for rewatered S. nigra and P. deltoides plants on Day 1 (Fig. 1). Similarly, midday leaf water potential recovered to at least control levels within 6 h of rewatering in S. nigra, P. deltoides, Q. velutina and Q. alba plants (Figs 1 and 2). (As noted, midday data were not available for J. nigra on Day 0.) Moreover, mean Ψmd of rewatered S. nigra plants was significantly higher than that of controls on Day 0 and remained somewhat higher on Day 1 (data not shown). Similarly, mean Ψmd of rewatered P. deltoides plants was higher than controls on Days 1 and 2 (Fig. 1). Midday Ψl of J. nigra plants, first measured 18 h after watering, had also fully recovered (data not shown). As was observed for S. nigra and P. deltoides, Ψmd of rewatered J. nigra plants tended to be somewhat higher than for controls during the first 4 d of recovery (data not shown), but differences were not significant. Treatment differences in Ψmd were not observed for Q. velutina and Q. alba plants at any time during the recovery, or after Day 2 in the other three species.

Figure 1.

Patterns of predawn and midday leaf water potential (Ψl), stomatal conductance (gs) and xylem sap abscisic acid concentration over 11 d of rewatering after drought in plants of Populus deltoides. Diamonds, droughted; closed squares, rewatered; open squares, control; error bars indicate ± 1 SE. Inset: xylem sap abscisic acid (ABA) concentration during the first 2 d of recovery; note change in [ABA] axis scale. *, P < 0.10; **, P < 0.05.

Figure 2.

Patterns of predawn and midday leaf water potential (Ψl), stomatal conductance (gs) and xylem sap abscisic acid concentration over 11 d of rewatering after drought in plants of Quercus alba. Diamonds, droughted; closed squares, rewatered; open squares, control; error bars indicate ± 1 SE. Inset: xylem sap abscisic acid (ABA) concentration during the first 2 d of recovery; note change in [ABA] axis scale. **P < 0.05.

Stomatal conductance

Recovery of gs of rewatered plants to levels exhibited by control plants lagged behind recovery of Ψl. Despite the restoration of Ψmd to control levels, complete recovery of gs was not observed on Day 0 (Fig. 1, Table 2) in three of the four species for which midday measurements were conducted. In the fourth species (Q. alba), gs of rewatered plants was less than 60% of that of control plants (Fig. 2), but this difference was not statistically significant. Comparisons of gs for rewatered and control plants over subsequent days suggested recovery at varying rates from one (S. nigra) to 4 d (Q. alba, Fig. 2). Combined analysis of gs data across species for Days 0 and 1 indicated a significant trend toward reduced stomatal opening in rewatered plants on Day 0, but not on Day 1 (Table 2). Overall, recovery time for stomatal conductance was not closely linked with species drought tolerance. However, isohydric, riparian J. nigra was the only species that showed a lag in recovery in gs on Day 1 that approached statistical significance (P = 0.07; Table 2).

Table 2.  Mean values of stomatal conductance (gs) and xylem sap abscisic acid (ABA) concentration on the day of rewatering (Day 0) or 1 d after rewatering had begun (Day 1)
 Day 0 gs (mmol m−2 s−1)  [ABA] (mmol m−3) Day 1 gs (mmol m−2 s−1)  [ABA] (mmol m−3) 
 RCRCRCRC
  1. R, rewatered, previously water-stressed plants (n = 4); C, plants kept at high soil moisture through the experiment (n = 2). −2 Σ ln p indicates χ2-value of combined probabilities for independent tests of significance for each species. *, P < 0.1; **, P < 0.05; ***, P < 0.01.

Salix nigra  78.4**459.6  1.781.01267.3219.3  0.50**1.26
Populus deltoides  31.5***410.6  3.640.14271.3547.2  0.130.22
Juglans nigra     –   –  59.8150.5  0.280.22
Quercus velutina  27.8**  80.7  0.380.04  69.1  77.5  0.090.11
Quercus alba163.2286.0  0.320.07165.8232.1  0.240.02
−2 Σ ln p  34.0*** 16.3*   11.0 13.5 

Xylem sap [ABA]

Mean baseline xylem sap [ABA] of control plants varied considerably among species, ranging from between 0.1 mmol m−3 in the drought-tolerant oaks to between 0.2 mmol m−3 and 0.9 mmol m−3 in the riparian species (Table 3). Xylem sap [ABA] increased significantly (P ≤ 0.05) in response to the drought treatment in all five species.

Table 3.  Mean xylem sap abscisic acid (ABA) concentration (mmol m−3) ± SE of young plants of five species at the end of the drought treatment (Day 0), when kept well-watered as controls throughout the experiment, and during 11 d of recovery from an imposed drought
 Water-stressed (n = 3)Control (n = 12)Rewatered (n = 24)
  1. The values for the rewatered and control treatments are the means for the entire treatment period. Differences between water-stressed and control seedlings are significantly different (P < 0.05) for all species.

Salix nigra15.36 ± 3.370.91 ± 0.140.88 ± 0.01
Populus deltoides40.84 ± 2.960.20 ± 0.020.73 ± 0.45
Juglans nigra  2.52 ± 0.620.33 ± 0.060.23 ± 0.03
Quercus velutina  1.15 ± 0.680.08 ± 0.010.15 ± 0.03
Quercus alba  1.21 ± 0.920.05 ± 0.010.14 ± 0.03

After morning rewatering, the [ABA] in the xylem sap dropped by midday (Figs 1 and 2; Table 2), but remained at least two to three times higher than that in control plants. Combined analysis of experiments suggested a general trend across species toward higher xylem sap [ABA] in rewatered than in control plants at midday of Day 0 (0.05 < P < 0.10, Table 2). With the exception of Q. alba, xylem sap [ABA] of rewatered plants was similar to or lower than that in control plants by Day 1 (Fig. 1, Table 2). Mean xylem sap [ABA] of rewatered Q. alba plants on Day 1 remained more than double that of control plants, but the difference was not statistically significant. Treatment differences in xylem sap [ABA] were not observed in any of the five species during the remainder of the recovery period.

ABA flux

Flux of ABA was significantly higher (P < 0.05) in droughted plants than in control plants in the two Quercus species and P. deltoides, but not in J. nigra and S. nigra (Fig. 3). In the four species for which data were available at midday on Day 0, ABA flux declined substantially upon rewatering. By midday of Day 1, ABA flux of rewatered S. nigra, J. nigra and Q. velutina plants was lower than that of control plants. However, ABA flux of rewatered Q. alba plants was somewhat elevated compared with control plants through Day 2. During the remainder of the recovery period, no significant differences among treatments in ABA flux were observed for any species.

Figure 3.

Patterns of flux of abscisic acid (ABA) over 11 d of rewatering after drought in plants of five temperate deciduous angiosperm species. Diamonds, droughted; closed squares, rewatered; open squares, control; error bars indicate ± 1 SE. **, P < 0.05.

gs–Xylem sap [ABA] relationships in water-stressed and rewatered plants

Because recovery of gs appeared to lag the decline in xylem sap [ABA] after rewatering, we explored the association of these variables during and after drought to assess whether the sensitivity of stomata to [ABA] might be different before and after rewatering. Although only a small number of water-stressed plants per species were available in this experiment, gs and xylem [ABA] data for additional water-stressed plants grown under similar conditions in the same greenhouse were available. Plots of the relationships between gs and [ABA] for rewatered plants and the combined populations of water-stressed plants suggested that stomatal aperture was more sensitive to a given level of xylem sap [ABA] after water-stressed plants had been brought to high soil water content than during a developing drought. Altered apparent sensitivity to ABA ranged from subtle (Fig. 4a, J. nigra) to obvious (Fig. 4b,Q. velutina).

Figure 4.

Reduction in gs as a percentage of maximum gs vs xylem sap abscisic acid concentration for droughted (closed symbols) and rewatered (open symbols) plants of Juglans nigra (a) and Quercus velutina (b). Values for xylem sap abscisic acid (ABA) concentration < 4 and < 10 mmol m−3 are shown in (a) and (b), respectively.

To test this apparent difference in stomatal sensitivity statistically, data for the portion of the gs–[ABA] relationship representing the stomatal closing phase (e.g. up to 3 mmol m−3 and 6 mmol m−3[ABA] for J. nigra and Q. velutina, respectively; Fig. 4) were used for each species to calculate an index of apparent stomatal sensitivity to xylem sap [ABA]. The index was computed for each individual plant by dividing the reduction of gs below the maximum observed (as a percentage) by plant xylem sap [ABA]. Two-sample t-tests of these data indicated that the reduction in gs per unit [ABA] was significantly greater in rewatered than water-stressed plants for four of the five species (Table 4). The apparent sensitivity of stomata of rewatered plants of S. nigra to xylem sap [ABA] was also greater than that of water-stressed plants, but the difference only approached statistical significance (P = 0.06). Combined analysis indicated a highly significant trend toward increased apparent stomatal sensitivity to xylem sap [ABA] in rewatered plants (P < 0.001).

Table 4.  Mean values of the index of stomatal sensitivity to abscisic acid (ABA) (reduction in gs as a percentage of maximum observed gs divided by xylem sap ABA concentration)
 Rewatered plants (% × (mmol ABA m−3)−1)Droughted plants (% × (mmol ABA m−3)−1)
  1. Significance of the difference between rewatered and droughted plants of each species is indicated with asterisks. −2 Σ ln p indicates χ2-value of combined probabilities for independent tests of significance for each species. *, P < 0.1; **, P < 0.05; ***, P < 0.01.

Salix nigra −0.086* −0.040
Populus deltoides −0.322*** −0.059
Juglans nigra −0.164** −0.090
Quercus velutina −0.501*** −0.036
Quercus alba −0.532** −0.255
−2 Σ ln p39.67*** 

Discussion

Plant water status recovered from an imposed drought in a matter of hours after the soil had been moistened and by the morning following re-irrigation, differences in Ψpd between rewatered and control plants were negligible. This result is consistent with reports in the literature of relatively rapid rehydration where moderately water-stressed plants were relieved from stress by irrigation of the soil (Aloni et al., 1991; Quick et al., 1992; Liang et al., 1996). Post-stress Ψmd in rewatered plants was actually higher than in control plants for some species in the days after rewatering had begun, a result that was most likely attributable to residual suppression of transpiration and consequent moderating effects on flow-induced Ψ gradients within the plant (Slatyer, 1967). This is a significant result, as rapid recovery of water status in the rewatered plants allowed us to compare gs–xylem sap [ABA] relationships without any confounding effects of drought differences between control and rewatered plants (Tardieu & Davies, 1992).

As has been shown for young plants and mature trees of several of these species, elevated xylem sap [ABA] is associated with stomatal closure during drought (Loewenstein & Pallardy, 1998a,b). While it is possible that this association is coincidental, the fact that these species showed stomatal closure when detached leaves were fed ABA in artificial sap supports a functional role for xylem [ABA] in control of stomatal aperture (Loewenstein & Pallardy, 1998a).

Stomatal reopening was inhibited following relief from drought despite rapid recovery of Ψl to prestress levels. Whereas leaf water potential had largely recovered within 6 h of rewatering in all five study species, gs did not recover to control levels for at least one full day. This lag in stomatal reopening after recovery of water status has also been reported in other woody species (Vitis vinifera, Quick et al., 1992; Litchi chinensis, Roe et al., 1995; Salix dasyclados, Liu et al., 2001b). Xylem sap [ABA] also recovered upon rewatering but remained two to three times above control levels at midday of Day 0, a result that may at least partly explain persistent stomatal closure early in recovery. However, xylem sap [ABA] recovered to control levels in all species before complete recovery of gs. Recovery of xylem sap or leaf apoplast [ABA] to that of well-watered controls before complete stomatal reopening has also recently been reported for Salix dasyclados plants subjected to partial root drying by exposure of roots to air (Liu et al., 2001a) and previously reported for rehydrated Leucaena leucocephala (Lam.) de Wit (Liang et al., 1997) and Lupinus albus L. (Correia & Pereira, 1995). These results and those reported here for five additional species suggest a general pattern of plant response wherein xylem sap ABA concentration in rehydrated plants returns to control levels before complete recovery of gs.

Two alternative mechanisms might be advanced to account for the longer-term lag in recovery. First, during the recovery period the stomata might be rendered more highly sensitive to xylem [ABA] by the episode of drought itself, perhaps even showing some closure at the baseline [ABA] levels that elicit no closure in unstressed plants. Our data pertaining to increased apparent sensitivity of stomata to xylem [ABA] in all species are consistent with this mechanism. Further, while previous work has indicated that stomatal sensitivity to xylem [ABA] appears to increase at low Ψ (Dodd et al., 1996), our data suggest that stomata are even more sensitive to xylem [ABA] in the recovery period than during the period of drought itself. If catabolism and/or sequestration of ABA are inhibited during the period of recovery, xylem-transported ABA may remain elevated at the site of action in guard cells despite a return of xylem sap [ABA] to pre-stress levels (Correia & Pereira, 1995; Liu et al., 2001b; Wilkinson & Davies, 2002). These metabolic shifts could explain the apparent change in stomatal sensitivity to ABA in our study.

This mechanism be would of interest to those researchers developing and applying models of stomatal regulation by endogenous and environmental factors, as these models are based on an assumption that the sensitivity of gs to xylem [ABA] increases as leaf water potential declines (Tardieu & Davies, 1992; Tardieu et al., 1993; Tardieu & Simonneau, 1998). As noted, our data indicate that once relieved from drought, stomata appear at least for a time to be more sensitive to xylem [ABA] even though Ψl has fully recovered. Studies of the temporal dynamics of this apparent change in sensitivity during recovery are needed to develop appropriate descriptive functions for these models, which could improve their predictive and explanatory powers.

An alternative mechanism consistent with our results would involve possible short-term inhibition of stomatal reopening by xylem [ABA] on the day of rewatering and more extended impacts of other potential sources of stomatal inhibition. Upon rehydration, some time may be required for restoration of ion transport in the guard cell plasma membrane, thus also possibly explaining lags in recovery of stomatal function. Cellular membrane injury is a common result of dehydration stress, especially in sensitive plants (Levitt, 1980). Pearce (1985) observed changes in distribution patterns of intra-membranous particles (IMP) in freeze-fractured membranes of water-stressed wheat leaves. Whereas IMP were evenly distributed across membranes in well-watered plants, some areas of plasma membranes became barren of IMP as plants were water-stressed. These particles likely include membrane channel and H+-ATPase proteins that are involved in guard cell solute movements. Within a given species, loss of membrane integrity is positively correlated with stress intensity (Martin et al., 1987; Gebre & Kuhns, 1991; Tan & Blake, 1993). Therefore, if the stomatal ‘after-effect’ is a function of membrane injury of the guard cell, one would expect to find a positive correlation between degree of stress and rate of recovery following relief from the stress. Indeed, this was the case for all species except J. nigra, as the correlation between stomatal conductance during recovery and peak-drought Ψpd tended to be positive during the first days after watering was resumed (data not shown).

Similarly, it is plausible that a reduced supply of cytokinins may contribute to the stomatal after-effect, given the capacity of cytokinins to promote stomatal opening (Fußeder et al., 1992). However, Fischer (1970) found that solutions of kinetin ranging from 0.1 to 17 mg l−1 did not stimulate stomatal recovery. In addition, Correia & Pereira (1995), observed that while the gs of moderately droughted lupin plants recovered within 2 h when leaves were excised and placed in water, the gs of severely droughted plants did not recover as quickly. Short-term feeding with benzyladenine had no effect on recovery of gs in severely water-stressed plants.

It was unlikely that impaired hydraulic conductance (Correia & Pereira, 1995; Tognetti et al., 1995; Liu et al., 2001a) contributed to the lags in stomatal recovery observed here. While hydraulic conductance should not directly affect stomatal conductance, it can indirectly influence gs through water status and supply of nutrients and root signals. Given the rapid recovery in leaf water potentials in this experiment, it is clear that hydraulic conductances of the rewatered plants in this experiment were not catastrophically impaired. Neither is it probable that variation in greenhouse conditions, particularly vapor pressure deficit (VPD), could account for changes in apparent stomatal sensitivity to ABA (Tardieu & Davies, 1992). Levels of light, VPD and leaf temperature in the present study were moderate and did not vary in a fashion that was correlated with trends in apparent stomatal sensitivity to xylem [ABA] (data not shown).

Stomatal recovery and drought tolerance

Differences in rate of recovery of stomatal opening did not appear to be closely linked with field drought tolerance; however, it should be noted that the plants studied here were subjected to moderate levels of drought and that the rate of recovery is likely to be inversely proportional to the level of drought induced (Fischer et al., 1970). The rate of recovery observed for drought-sensitive J. nigra in this experiment was more rapid than observed in the same species in a previous experiment (Ni & Pallardy, 1992), but water potentials induced in the latter study were lower than those seen here. As we wished to avoid a confounding effect of the leaf senescence/abscission response that rapidly follows drought-induced stomatal closure in isohydric plants (i.e. J. nigra, S. nigra and P. deltoides; Pallardy & Rhoads, 1993, 1997), we may have been unable to detect differences among species in recovery associated with more severe levels of stress.

Acknowledgements

Research supported by the USDA National Research Initiative under Grant USDA-CSREES 96-35100-3215.

Ancillary