Abscisic Acid Synthesis in Acer rubrum L. Leaves— A Vapor-pressure-defi cit-mediated Response

Quantitative differences in leaf abscisic acid (ABA) among Acer rubrum L. (red maple) ecotypes were investigated. This study tested the hypothesis that seedlings from wet and dry maternal sites display distinctly different capacities to synthesize ABA in response to atmospheric vapor pressure defi cits. The increased levels of ABA in leaf tissue in the red maple ecotypes were associated with atmospheric vapor pressure defi cit (VPD). Leaves on well-watered plants responded to VPD by increasing their ABA levels and reducing their photosynthesis (Anet) and stomatal conductance (gs). Both ecotypes appear to accumulate ABA at about the same rate as VPD increased. Despite the similar accumulation rates between ecotypes, wet site ecotypes consistently had a higher level of ABA present in leaf tissue under both low and high VPD conditions. Furthermore, wet site provenances appear to reduce Anet and gs in response to ABA accumulation, whereas dry sites do not present as clear an ABA/gs relationship. This study shows variation between wet and dry site red maple populations in physiological response to atmospheric vapor pressure defi cits, indicating that natural ecotypic variation in stomatal responsiveness to air humidity is likely mediated by ABA accumulation in leaf tissue. This research demonstrates that ecotypes of red maple may be selected for atmospheric drought tolerance based on site moisture conditions. Stomata are the primary passageway for CO2 diffusion into the substomatal cavity and allow the plant to regulate water loss and carbon gain. Recent evidence suggests that stomatal conductance (gs) correlates better with environmental variables such as air temperature, vapor pressure defi cit (VPD), and photosynthetic photon fl ux density (PPFD) than with plant variables such as shoot water potential, turgor potential, bulk leaf osmotic potential, xylem pH, and xylem sap abscisic acid (Augé et al., 2000). Plants are known to respond to a high leaf-to-air VPD by closing their stomata (Maroco et al., 1997; Meinzer et al., 1997; Thomas and Eamus 1999) and this response has been observed in a variety of species (Monteith, 1995; Sheriff, 1979). Studies of well-watered Douglas-fi r and alder roots indicate that the declines in leaf water potential associated with transpiration-induced water potential decline can induce stomatal closure without invoking a change in soil water potential (Fuchs and Livingston, 1996). Gollan et al. (1986) maintained the same leaf water potentials across the soil–plant–water continuum as the soil dried, so the closure of stomata occurred in drying soil despite the same demand/supply regime. Abscisic acid (ABA) has been shown to be a stress signal that permits communication between the roots and shoots, especially when the root system is under stress (Davies and Zhang, 1991; Davies et al., 1994). The hypothesis that ABA is the only root signaling molecule, however, has been challenged (Munns, 1992; Munns and King, 1988; Munns et al., 1993; Trejo and Davies 1991). Currently, there is a paucity of studies that investigate the effect of atmospheric water defi cit on the production of ABA. Gollan et al. (1986) and Turner et al. (1985) carefully unscrambled the effect of VPD and soil drying on photosynthesis and conductance that led to the subsequent work on root-shoot signals and ABA. Augé et al. (2000) found that VPD better described stomata sensitivity to ABA than shoot water potential (Ψw). Nonetheless, a nonhydraulic chemical factor that permits plants to sense and respond to soil drying seems to exist (Davies and Zhang, 1991; Davies et al., 1994; Liang et al., 1996; Loewenstein and Pallardy, 1998a, 1998b; Zhang et al., 1987). Acer rubrum L. (red maple) is one of the most widespread hardwood tree species in North America (Walters and Yawney, 1990). Not only does red maple have a wide range, it thrives under a wider variety of site/soil conditions than most other North American tree species. With respect to horticulture, it is a popular ornamental tree and is commonly placed on a variety of landscapes. Subsequent studies indicate that red maple populations are genetically variable for traits including drought tolerance of substrate moisture (Abrams and Kubiske, 1990; Bauerle et al., 2003a, 2003b; Townsend and Roberts, 1973); however, a review of the literature reveals no investigations of atmospheric stress tolerance. The goal of the study was to examine the range of stomatal control displayed by red maple trees grown under well-watered conditions when subjected to extreme atmospheric VPD. The Received for publication 1 May 2003. Accepted for publication 10 Oct. 2003. The authors thank A. Lakso, and S. Maki for helpful discussions of earlier drafts of this manuscript, A. Leed for technical assistance and M. Compton and A. Roberts for plant care. We also thank Nursery Supply, Inc. for donating the pots. This work was supported by USDA Hatch Projects NYC-141370 and NYC-141406. 1Corresponding author; e-mail [email protected]. 089-Env 182 1/13/04, 10:00:35 AM 183 J. AMER. SOC. HORT. SCI. 129(2):182–187. 2004. horticultural popularity and natural variation within the species warrants red maple as a good model for this study. Red maple populations originating from water-limiting sites (dry site) osmotically adjust and have a lower capacity for growth when compared to sites were water is not limiting (wet site) (Bauerle et al., 2003b). The dry-site seedlings appear to have tissue better adapted to withstand dehydration and might be less sensitive to nonhydraulic signaling because they can osmotically adjust. Conversely, red maple populations originating from sites where water is not limiting do not osmotically adjust. We hypothesize that wet-site trees may have a greater ability to both biosynthesize and catabolize ABA, a nonhydraulic root-to-shoot signal, to control excess transpirational water loss during periods of excess atmospheric water demand. The altered stomatal control would constitute a sensitive mechanism by which transpiration is reduced in an ecotype that does not normally experience water defi cits and aid in selection strategies for atmospherically stressful landscape situations (e.g., urban heat islands). In this study, we address the following questions: 1) Does the ABA stress response quantitatively differ among red maple ecotypes, 2) What is the relative importance of leaves as sensors of atmospheric water defi cits in red maple, and 3) Does red maple show ecotypic differences in leaf synthesis of ABA? Specifi cally, this study examined physiological mechanisms of drought tolerance in red maple seedlings from eight provenances. Six provenances represented the wet and dry hydrologic extremes in New York. The other two sites, both in Virginia, represented among the wettest and driest sites inhabited by red maples in North America. Altogether then, trees from four wet and four dry sites were subjected to stress by creating atmospheric water defi cits and then relieving the water stress. Materials and Methods PLANT MATERIALS. The geographic origin, site descriptions, and seedling establishment have been described elsewhere, Bauerle et al. (2003b). Specifi c to this experiment, 144 seedlings comprised 18 replications from each of the four wet and four dry sites. Annual cohort 1998 plant material was randomly selected from the eight sites in May 2000. Briefl y, a wet site was defi ned by poorly drained soils that are saturated in the spring during seed development (a wetland or swamp), and a dry site was defi ned by well-drained upland soils. GROWTH CHAMBER CONDITIONS. Two walk-in growth chambers (Environmental Growth Chamber Co., Chagrin Falls, Ohio) were used to control light and temperature. Within the control room (C1), an ambient whole plant chamber (VPDA), dimensions 1 × 1 × 0.66 m3, was constructed from polyvinyl chloride (PVC) tubing covered with clear 2 mm Mylar. A data logger monitored relative humidity (model Hobo Pro; Onset Computer Corp., Pocasset, Mass.) and an infrared thermocouple (OS36, Omega Co., Stamford, Conn.), connected to a remote datalogger (CR21X; Campbell Scientifi c Inc., Logan, Utah), monitored temperature inside each individual chamber at minute intervals. Within the treatment room (C2), two Mylar chambers were constructed, identical to the chamber in C1. During VPD stress, moisture was removed from the ambient atmospheric air via a cold plate condenser and Drierite desiccant and the temperature was set at 34 °C, high vapor-pressure-defi cit chamber (VPDH). A plenum heated with three 150-W incandescent lights brought the temperature back to 34 °C before entering the dehumidifi ed chamber. In addition, 7.25 kg of Drierite was spread over the bottom of the VPDH to aid in further removal of H2O. In the second Mylar chamber, water was added to the ambient atmospheric air stream via three ultrasonic humidifi ers (Sunbeam Inc., BocaRaton, Fla.). Again, air was passed through a plenum to bring the temperature back up to 34 °C before entering the humidifi ed chamber (VPDW). Each Mylar chamber held 3 maple tree seedlings at any given time. Environmental parameters were varied over a 12 h time course for each of the eight red maple ecotypes, which resulted in one ecotype per day being processed. The containers of the seedlings were sealed in white plastic bags to prevent evaporation. The treatments started at 0800 HR. Leaf gas exchange and ABA were measured to indicate the impact of ambient, atmospheric VPD and humidifi ed elevated temperature conditions, as well as post stress responses. Leaf disks were harvested every 4 h on all plants within the chambers (fi ve 1 cm disks per tree) and gas exchange data collected every 2 h from a subsample of plants (one per chamber) with a gas analyzer (LI-6400; LI-COR, Inc., Lincoln, Neb.). The gas exchange chamber was placed inside the growth chamber for a period of no less than 0.5 h to equilibrate with growth chamber conditions before the beginning of a measurement. At 1600 HR (hour 8) the VPDH and VPDW chambers were returned to ambient relative humidity until 2000 h in order to investigate the decline in ABA in the leaf tissue during the 4-h poststress time frame. ABA SAMPLING. Using a 1-cm cork borer, leaf disks were collected from the fi rst through third fully expanded leaves every 4 h. The protocols were simple modifi cations of Alves and Setter (2000). Briefl y, fi ve 1-cm-diameter leaf disks were harvested from each replicated seedling and immediately placed in a precooled (0 °C) 1.5-mL microcentrifuge tube containing 676 μL of extraction medium (80% v/v methanol, 20% v/v glass distilled H2O). Samples were stored at –18 °C until analysis. A 200 μL extract per sample was lyophilized and then redissolved in 150 μL of aqueous + 1% v/v glacial acetic acid and 10 μL of [3H] ABA with sonication. Chromatography columns were constructed with micropipette tips containing 0.15 g of silica C18 packing material (40 μm particle size). Columns were washed with 800 μL of 95% EtOH and then with 600 μL of 20% MeOH + 1% v/v glacial acetic acid with suction applied via a vacuum aspirator (Univac, Polyfi ltronic). As soon as washing was complete, the extract was loaded under constant vacuum at a rate of ≈5 μL·s–1. The column was then washed two times with 200 μL of 20% MeOH + 1% v/v glacial acetic acid. Columns were then eluted with 200 μL of 55% MeOH and the ABA extract collected. Upon collection of elute, the samples were again stored at –18 °C. ABA ASSAY. ABA was assayed by enzyme linked immunosorbent assay (ELISA) as described by Alves and Setter (2000). Briefl y, each well of a 96-well microtiter plate (Corning/Costar high binding #3366) was coated with 20 μL of ABA-bovine serum albumin conjugate. After incubation, for 24 h at 5 °C, the plate was decanted and washed four times with TBST (Tris buffered saline with 0.02% v/v Tween-20) with 5-minute incubations per wash. One hundred μL of TBSA (Tris buffered saline + bovine serum albumin) and 10 μL of eluted sample were added to each well. Then 100 μL of anti-ABA monoclonal antibody (clone 15-I-C5, currently available from Agdia Inc., Elkhart, Ind.) was added to each well. The plate was incubated for 24 h at 5 °C. After incubation, the plate was again decanted and washed with TBST a total of four times. One hundred and eighty microliters of diluted secondary antibody (antimouse-alkaline phosphotase conjugate, Sigma product A-3562, in TBST with 0.1% [w/v] 089-Env 183 1/13/04, 10:00:40 AM 184 J. AMER. SOC. HORT. SCI. 129(2):182–187. 2004. BSA) was added to each well. The plate was incubated at 5 °C for 24 h. Once the fi nal incubation was complete, the plate was decanted and again washed four times. Colorimetric reagent, containing para-nitrophenylphosphate, PNPP (Sigma N3129) in diethanol-amine buffer was added and the plate was left to develop for 1 h at room temperature. After 1 h, the plate was read with a plate reader at a wavelength of 405 nm (model 750; Cambridge Technology, Watertown, MA). (+)ABA content was determined by calculations based on (+)ABA calibration standards. A spreadsheet macro written in Excel (Microsoft, Seattle, Wash.) provided a logit-transformed plot of the standard curve, calculated regressions, and predicted pmol ABA per well. Samples were replicated three times in the assay and averaged. Once the assay was complete, a 70-μL aliquot was mixed with 400 μL of scintillation fl uid (Ecocint H; National Diagnostic, Manville, N.J.), and the radioactivity was measured in a scintillation counter (model LS 5000 TD; Beckman Inc., Fullerton, Calif.). Individual ABA samples were then corrected for loss that occurred in the chromatography clean up process. DATA ANALYSIS. Gas exchange and ABA data were analyzed as a repeated measure using the GLM (General Linear Model) module (SPSS, Inc., 1999). To address the potential for between chamber variation in Anet, gs, and ABA a Fisherʼs least signifi cant difference test (P < 0.01) was used to characterize the relationship between treatment chambers within a time interval. Paired ABA, Anet, and gs data were analyzed with a two level hierarchical model as covariates at the tree level (SAS proc mixed module; SAS Institute, Inc. 2000). Fig. 1. Net photosynthesis (Anet), stomatal conductance to water vapor (gs), abscisic acid (ABA), and vapor pressure defi cit of the atmospheric air (VPD) (♦) of high VPD and air temperature maintained at 34 °C (A), moderate VPD and high air temperature maintained at 34 °C (B), and moderate VPD and air temperature maintained at 22 °C (C). Photosynthetically active radiation (PAR) was provided by fl uorescent lamps at a maximum photosynthetic photon fl ux of 375 μmol·m–2·s–1 at the terminal tip of the canopy. Photoperiod was varied diurnally for a photoperiod of 14 h each day. The lamps were set to come on at 0600 HR and shut off at 2000 HR. Data represent four consecutive times on wet-site (■) and dry-site (●) seedlings of red maple. Each ABA point represents the mean of six replicate plants for each of four provenances per wet or dry site ± SE. As described in the methods section, data are a subset for gas exchange measurements. 089-Env 184 1/13/04, 10:00:45 AM 185 J. AMER. SOC. HORT. SCI. 129(2):182–187. 2004. Results GAS EXCHANGE AND ABA RELATIONSHIP. The ≈5 kPa VPD treatment resulted in increased leaf ABA over time, relative to control VPD levels of ≈1.2 kPa (P < 0.01) (Fig. 1). No treatment differences were observed at 0 h in any of the measured variables, however, signifi cant differences were observed over time between treatments. In addition, Anet, gs, and ABA differed between wet versus dry sites in response to varied VPD (P < 0.01). Leaf gas exchange and ABA indicate the effect of atmospheric VPD demand and VPD relief (Fig. 1). Before the initiation of the treatments, wet-site ecotypes had a signifi cantly higher leaf ABA level, Anet rate, and gs regardless of the treatment chamber (P < 0.01). EFFECT OF HIGH VPD AND HIGH TEMPERATURE. Within the high VPD chamber, Anet decreased 17% in the wet-site ecotype during the fi rst 4 h of stress while dry-site plants remained unchanged (Fig. 1A1). Stomatal conductance remained relatively unchanged after the fi rst 4 h of stress, whereas leaf ABA increased signifi cantly (Fig. 1A2 and A3). The greatest change was observed between 4 to 8 h, with a decrease of 36% and 15% in wetand dry-site gs respectively (Fig. 1A2). Photosynthesis refl ected the decline in gs. However, wetand dry-site declines in Anet were similar during this period (Fig. 1A1). ABA accumulation, conversely, continued to increase between 4 h and 8 h. Wet-site trees maintained higher leaf ABA but the rate of accumulation was similar for both ecotypes during the fi rst 8 h (Fig. 1A3). EFFECT OF MODERATE VPD AND HIGH TEMPERATURE. To elucidate temperature infl uence, samples were collected from plants grown in a chamber where VPD was maintained at moderate levels (2.6 ± 0.2 kPa). Within the high temperature and moderate VPD chamber, Anet and gs increased in both the wetand dry-site ecotypes during 0 to 4 h (Fig. 1B1 and B2), while ABA levels either declined or remained constant. In dry-site trees, levels remained stable from hour 0 through 4 and then both ecotypes accumulated ABA similarly from hour 4 to 8 (Fig. 1 B3). Wet-site trees declined in ABA levels from 0 to 4 h with increases in gs and Anet. After the initial increase, stomatal conductance and Anet remained relatively unchanged between hours 4 to 8, while an increase in leaf ABA levels occurred in both ecotypes (Fig. 1B3). In contrast to the VPD effects, elevated temperature resulted in a nearly opposite response with respect to Anet and gs. Both ecotypes had elevated Anet and gs during the period when temperature was raised. Stomatal conductances in particular were about two times that of hour 0 when temperature was 22 °C and VPD was 1.2 ± 0.2 kPa. Once temperature was returned to ambient conditions (hour 8 to 12), gs of trees in the VPDW chamber declined to near 0 h levels. In comparison, the ambient control chamber did not show a signifi cant change in Anet, gs, or ABA through 0 to 8 or 8 to 12 h at a VPD of 1.2 ± 0.2 kPa. Regardless of ecotype, the stomatal conductance to water vapor and Anet both increased during the poststress recovery period in the VPD chamber (Fig. 1A1 and A2). Moreover, hour 12 values of Anet and gs were similar to hour 0 values for both ecotypes, with wet-site ecotypes persisting at higher Anet and gs levels throughout the time frame (Fig. 1A1, and A2). Anet and gs declined in the VPDW and returned to prestress levels after 4 h of recovery. EFFECT OF AMBIENT VPD AND AMBIENT TEMPERATURE. Within the VPDA chamber, Anet and gs remained different between ecotypes with wet sites having both higher Anet and gs (Fig. 1C1 and C2). On the other hand, chamber effects through the entire time course were not signifi cant on either ecotype. Moreover, ABA levels remained relatively unchanged (Fig. 1C3). STOMATAL RESPONSES. Figure 2 illustrates the sensitivity of Anet and gs to ABA accumulation and that gs and Anet decline more with ABA accumulation in wet, as opposed to dry sites. Additionally, Fig. 2 illustrates the mean of wet and dry sites across all conditions from hour 0 to 12 where linear regression indicates a signifi cantly steeper slope for wet sites (P < 0.01). Wet-site ecotypes reduced Anet and gs in association with increased levels of ABA. Wet-site ecotypes had higher levels of ABA and as ABA accumulated, Anet and gs declined, whereas dry sites did not decrease or increase Anet and gs in association with ABA. Despite the fact that the rate of ABA accumulation in the wet-site plants was not signifi cantly greater than the rate observed from the dry-site plants, signifi cantly higher levels were still maintained in wet-site ecotypes. The steeper slope for wet sites illustrates the response to ≈5 pmol·cm–2 of ABA accumulation, whereas the dry sites have less of a response with more ABA accumulation (≈6 pmol·cm–2). Additionally, poststress gs for wet-site ecotypes in the VPDW were similar to prestress, but dry sites remained elevated (0.01 versus 0.06). Lastly, poststress Anet recovered to prestress rates across treatments with wetand dry-site ecotypes behaving differently. Figure 3 illustrates the relationship among physiological variables Anet and gs. The Loess method (Cleveland, 1979) was used for locally weighted scatter plot smoothing in Fig. 3 and weights Fig. 2. The mean of wet and dry sites across all conditions from hour 0 to 12. Mean net photosynthesis (Anet) and stomatal conductance to water vapor (gs) versus mean abscisic acid (ABA) concentration. Ecotypes are wet sites (■) and dry sites (●). A linear regression was individually fi t to both wet and dry sites with three iterations. Solid lines represent the linear regression curve for wet sites and broken lines represent dry sites. 089-Env 185 1/13/04, 10:00:48 AM 186 J. AMER. SOC. HORT. SCI. 129(2):182–187. 2004. the proximate nearest neighboring points more heavily. Moreover, the Loess curve was fi t to the data to compare the relationship between wet-site and dry-site Anet response to gs. Wet-site plants had higher Anet per unit gs and wet-site gs values extended beyond the maximum dry-site values.