This study investigated the role of quercitol in osmotic adjustment in field-grown Eucalyptus astringens Maiden subject to seasonal drought stress over the course of 1 year. The trees grew in a native woodland and a farm plantation in the semi-arid wheatbelt region of south Western Australia. Plantation trees allocated relatively more biomass to leaves than woodland trees, but they suffered greater drought stress over summer, as indicated by lower water potentials, CO2 assimilation rates and stomatal conductances. In contrast, woodland trees had relatively fewer leaves and suffered less drought stress. Plantation trees under drought stress engaged in osmotic adjustment, but woodland trees did not. Quercitol made a significant contribution to osmotic adjustment in drought-stressed trees (25% of total solutes), and substantially more quercitol was measured in the leaves of plantation trees (5% dry matter) than in the leaves of woodland trees (2% dry matter). We found no evidence that quercitol was used as a carbon storage compound while starch reserves were depleted under drought stress. Differences in stomatal conductance, biomass allocation and quercitol production clearly indicate that E. astringens is both morphologically and physiologically ‘plastic’ in response to growth environment, and that osmotic adjustment is only one part of a complex strategy employed by this species to tolerate drought.
Most climate change scenarios suggest an increase in aridity in many areas of the globe (IPCC 2007), which makes investigations of plant responses to drought stress increasingly important. Higher plants have developed a range of mechanisms to adapt to soil moisture deficits, with escape, avoidance and tolerance strategies being the most important (Levitt 1972). Annual species such as grasses can escape drought completely by dying off during the driest times of the year. Perennial species such as trees and shrubs can avoid drought by maximizing water uptake (e.g. tapping permanent water sources via deep roots) or minimizing tissue water loss while maintaining high tissue water potentials (e.g. stomatal closure, sclerophyllous leaf morphology, leaf abscission). Other perennial plants have developed mechanisms for tolerating drought based on tolerating low tissue water potentials by having smaller cells, more rigid or more flexible cell walls or through osmotic adjustment (Morgan 1984).
Osmotic adjustment occurs when plants accumulate solutes for the purpose of maintaining positive cell turgor. Osmotic adjustment can occur either as an active process [i.e. the net accumulation of solutes on a dry weight (DW) basis] or as a passive process (i.e. a decrease in tissue water content that leads to an increase of solutes on a plant water basis, but not to a net increase on a DW basis) (Turner & Jones 1980). Adjustment is usually slow, but it is a potentially powerful adaptation to drought because it allows plants to maintain water absorption, cell turgor and metabolic activity during periods of drought stress, and also enables quick resumption of growth when water becomes available again (Chaves, Maroco & Pereira 2003).
Osmotic adjustment is a widely studied phenomenon, but the specific compounds responsible for maintaining cellular osmolarity remain unidentified and unstudied in most species. Previous work has shown that plants may regulate ion concentrations and/or draw on a variety of solutes, including organic acids, amino acids, quaternary ammonium compounds, carbohydrates and cyclitols (Hasegawa et al. 2000). Different solutes almost certainly vary in their metabolic cost, and it is reasonable to assume that plants have evolved to use solutes that provide the most benefits for investment. This assumption underpins recent work focusing on the role of ‘compatible’ solutes in osmotic adjustment, that is, solutes that do not inhibit cytosolic enzymes (as do high concentrations of ions), but rather protect cell membranes and metabolic processes under dehydration (Popp & Smirnoff 1995; Jaindl & Popp 2006).
Cyclitols are a group of compatible solutes that occur in appreciable concentrations in a wide range of plant species (Popp et al. 1997). Cyclitols such as pinitol, quebrachitol and quercitol can accumulate in relatively large amounts, up to several percent of dry matter on a whole-plant basis (Dittrich, Gietl & Kandeler 1971; Popp et al. 1997; Merchant et al. 2006a), and they have been shown to accumulate in response to drought (Nguyen & Lamont 1988) or salinity (Richter, Thonke & Popp 1990), which also suggests a role in stress tolerance. In the case of Eucalyptus, glasshouse studies have shown that concentrations of quercitol increase in the leaves of some species in response to salinity and drought stress (Adams et al. 2005; Merchant & Adams 2005). Species from low rainfall environments tend to produce significantly greater amounts of quercitol in leaf tissues than species from mesic environments (Merchant et al. 2006a), suggesting that quercitol accumulation may be correlated with species distribution as determined by water availability. However, quercitol is not always utilized as a dominant osmolyte, and its production varies greatly among different Eucalyptus species. Under some experimental conditions, quercitol concentrations (on a water content basis) may increase in leaf tissue only as a consequence of decreased leaf water content rather than active accumulation (Merchant & Adams 2005; Merchant et al. 2006b), and recent work has demonstrated that cyclitols can be a minor component of total solute concentration in species that maintain high leaf water contents and display strong stomatal sensitivity to water deficits (Warren, Bleby & Adams 2007).
Quercitol functions as a stable osmoticum in Eucalyptus species and contributes significantly to the cellular osmotic potential in leaves, but its role in osmotic adjustment is not yet fully understood. Most studies investigating quercitol have been performed on juvenile trees in glasshouse studies, and most have only investigated its accumulation in leaves. It is currently unknown if quercitol concentrations show similar responses in mature trees in the field, where the environmental conditions are different and complex, and the onset of drought stress is less abrupt as compared to glasshouse experiments. Previous studies of Quercus robur indicate that quercitol can be translocated between leaves and woody above-ground tissues during the year (Popp et al. 1997), and appreciable amounts of quercitol have been measured in xylem sap of a range of Eucalyptus species (Merchant, unpublished results). It appears possible that quercitol may be translocated between different organs during the course of the year, and that it may therefore play a role in carbon transport or carbon storage.
The primary aim of this study was to investigate whether the concentration of quercitol in different plant tissues of a semi-arid Eucalyptus species changes in response to drought stress on a whole-plant level in the field. The objectives of the study were to investigate: (1) the concentration of quercitol in different tissues (leaves, branches, stems, roots) at different times of the year; (2) the contribution of quercitol to total osmolytes and possible osmotic adjustment in the different tissues; and (3) the translocation of quercitol between the different tree organs during the season. The species chosen for this study was Eucalyptus astringens (brown mallet), which grows naturally in the semi-arid wheatbelt region of southern Western Australia. This region has a typical Mediterranean-type climate, and E. astringens experiences significant drought stress during the hot, dry summer months and little or no drought stress during the cooler, wet winter months. We compared two neighbouring stands, a native woodland and a farm plantation, where preliminary work showed that plantation trees were more drought stressed than woodland trees, particularly in summer.
MATERIALS AND METHODS
We studied E. astringens Maiden stands at two sites in the Dryandra Woodland region, approximately 200 km south-east of Perth in Western Australia. Dryandra Woodland is the largest area (28 066 ha) of remnant woodland and associated vegetation in the Western Australian wheatbelt, comprising 17 discrete remnant woodland blocks separated by agricultural (cleared) land (CALM 1995). The region is characterized by a prolonged hot and dry summer (mean 23 °C) and a cool and wet winter with occasional severe frosts (mean 10 °C). The mean annual rainfall is approximately 500 mm, 77% of which falls between May and October. Meteorological data for the period of this study were sourced from a weather station located in the township of Narrogin through the Bureau of Meteorology.
The first site, located 20 km north-east of Narrogin, WA (S32°46′29″, E117°07′19″), was a natural woodland (5000–30 000 stems ha−1) that had naturally re-established from seed following a wild fire in 1991. The second site was a plantation (2000 stems ha−1) that had been established in 1999 on agricultural grazed land about 10 km south of the woodland (S32°51′45″, E117°07′44″). Trees at the woodland had an average height of 3.5 m, compared to 2.4 m at the plantation. Above-ground biomass at the woodland site with the older, denser stand of trees (∼20 Mg ha−1) was approximately double that of the plantation (∼10 Mg ha−1). The opposite was true for leaf area, in that the leaf area index (LAI) of the plantation (LAI = 2) was approximately twice that at the woodland site (LAI = 1). Preliminary studies showed that E. astringens exhibited greater drought stress in the plantation during summer months compared to trees growing in the woodland, as indicated by differences in pre-dawn and midday leaf water potentials, and mid-morning stomatal conductance (data not shown).
Soils at both sites were yellow duplex soils (sand over clay), and the soil bulk density was similar at both sites (1.3–1.5 g cm−3 in the top 10 cm). Topsoil (0–7 cm) at the woodland site had a total soil C content of 1.01% and a total N content of 0.03%, whereas at the plantation site, topsoil total C ranged between 0.51% (on the mounds of tree rows) and 0.83% (in the furrows between rows), and total N ranged between 0.25% (mounds) and 0.55% (furrows). The soils in the plantation had greater mineral nitrogen contents, probably as a consequence of frequent fertilization of the pasture prior to afforestation and other agricultural practices. Available soil nitrate (NO3-), as determined from in situ incubations of ion exchange resin bags, was less in the E. astringens woodland (≤1.2 µg NO3-–N kg−1 resin d−1) than in the newly established plantation (≤8.7 µg NO3-–N kg−1 resin d−1), while available soil ammonium (NH4+) was similar, peaking at 2.3 µg NH4+–N kg−1 resin d−1.
A total of five destructive harvests were undertaken in March, May, July, September and December 2002 to study seasonal variations in total osmolytes, quercitol, total carbohydrates and starch in leaves, branches, stems and roots of E. astringens. Five trees at each site were selected at random covering the range of stem diameters at both sites within a 100 m radius, and harvested and measured for total above-ground biomass at each harvesting event. Above-ground components of each tree were harvested in the morning (between 0900 and 1300 h) and separated into leaves, branches and stems. Roots were excavated later in the day, and all main roots extending from the tree bole, including tap roots, were exposed to a depth of 0.5 m within a lateral 1.5 × 1.5 m quadrant around the stem. All roots > 5 mm in diameter were harvested. Preliminary work showed that the vast majority (85–90%) of the lateral root mass of trees occurred in this volume of soil. The fresh weight (FW) of each tissue fraction (leaves, branches, stems and roots) was determined in the field. Tissue fractions of woody tissues were coarsely chopped, and each tissue fraction from each individual tree was thoroughly mixed, and subsamples were taken for solute analyses. Subsamples were rapidly heated to 100 °C in a microwave oven in the field (Popp et al. 1996) to ensure minimal change in solute composition, and then dried to constant weight at 80 °C in a drying oven in the laboratory. Total DWs for each tree were calculated from the FW:DW ratios of each tissue subsample and total FW of each plant tissue fraction.
Gas exchange and water relation measurements
Net rate of carbon assimilation and stomatal conductance of sunlit and fully expanded leaves were measured with a CIRAS-1 portable infra-red gas analyser (PP Systems, Hitchin, Herts, UK). The CIRAS-1 was equipped with automatic CO2 control and a standard broadleaf cuvette that can measure gas exchange of an illuminated 2.5 cm2 area of a leaf. The cuvette headspace CO2 concentration was controlled by the CIRAS-1 at approximately 360 ppm. The cuvette lacked temperature control but contained ventilated internal and external heat exchange surfaces. Leaf temperature in the cuvette was measured with an inbuilt infrared temperature sensor. All measurements were made at ambient light conditions on sunny days without cloud cover. The measurements were taken between 1000 and 1200 h on the day prior to harvesting. We measured gas exchange of three fully expanded, outer canopy leaves on each of the five trees subsequently harvested at each site. Pre-dawn leaf water potential was also measured with a pressure chamber (PMS Instrument Co., Corvallis, OR, USA) on fully expanded, outer canopy shoots of three to four leaves from the same five trees at each site. Pre-dawn leaf water potential and leaf gas exchange were measured on the harvesting dates stated earlier and additionally on 15 February 2003.
Dried leaf, branch, stem and root samples were ground to a fine powder in a ball mill (Retsch MM2; Vienna, Austria) and extracted with hot water (4:1, w/v, 95 °C, 1 h). Total osmolyte concentration (i.e. that of all osmotically active substances) was determined by freezing point osmometry of the hot water extracts (HWEs) (Callister, Arndt & Adams 2006; Merchant et al. 2006b). The osmotic potential of the HWE was determined using an Osmomat 030 freezing point depression osmometer (Gonotec, Berlin, Germany) that was zeroed and calibrated daily with water and a standard solution provided by the manufacturer. Tissue carbohydrate and polyol content were extracted using an initial methanol–chloroform–water mixture (MCW) (12:5:3, v/v/v, 80 °C, 30 min) and isolated using phase separation. Aliquots of the MCW extracts were deionized and silylated using a 10:1 mixture of bis-trimethylsilyl-trifluoroacetamide (BSTFA) and trimethylchlorosilane (TMCS). Carbohydrates and polyols were analysed by capillary gas chromatography (Shimadzu 17A; Shimadzu Corporation Limited, Columbia, MD, USA) using a DB1 (30 m) column (Merchant et al. 2006b). To determine starch content, leaf and root samples were homogenized and repeatedly extracted with ethanol:water (8:2 v/v) to remove low-molecular weight carbohydrates. Starch was enzymatically degraded to glucose with heat-stable α-amylase and amyloglucosidase. Glucose concentration was then measured by high-performance anion exchange chromatography (HPAEC-PAD, DX 500 and ED 40; Dionex, Lane Cove, NSW, Australia) (Arndt et al. 2000).
Concentrations of total osmolytes are reported in units of both mmol kg−1 DW and mmol L−1 plant water (using field-measured water content data). Concentrations of the other solutes are reported in mmol kg−1 DW. Root samples from woodland trees sampled in December 2002 were not analysed for solute content owing to technical difficulties.
All data were analysed using the Statgraphics Plus software (Statistical Graphics Corp., Herndon, VA, USA). One-way analysis of variance (anova) was used to identify statistically significant differences in gas exchange and water relation parameters, and concentration of osmotically active substances among sampling times within each site. Significant differences were determined using post-hoc tests [Tukey honestly significantly different (HSD)] at P < 0.01.
The Dryandra woodland region experienced typical Mediterranean-type climatic conditions over the course of this study (Fig. 1). The hot and dry summer period lasted from December 2001 to April 2002, and the first substantial rainfalls occurred in May 2002. Total precipitation for 2002 was 350 mm, and most (70%) of that rainfall occurred in the winter months between June and September. There were a few frosty days in late July and early August, where minimum temperatures were below zero.
Stands of E. astringens were relatively homogenous in stand structure and even aged at both sites, but there were clear differences between the sites with respect to tree age, stand structure and biomass allocation. Plantation trees were 4 years old compared to around 10 years old in the woodland, had been planted at a lower stem density than the natural regeneration in the woodland and clearly had greater access to light and nutrients (NO3- and NH4+) compared to the woodland trees. Plantation trees allocated approximately 35% of their biomass to the stem, 20% to branches, 20% to leaves and the remaining 20% to roots (Table 1). In comparison, woodland trees allocated relatively less biomass to branches and leaves and more to stems, approximately 50% to stems, 15% to branches, 15% to leaves and 20% to roots (Table 1). Patterns of biomass allocation did not change significantly during the year at either site (Table 1).
Table 1. Allocation of biomass and quercitol among different plant tissues (leaves, branches, stems and roots) of Eucalyptus astringens grown in a native woodland or a plantation in Dryandra, Western Australia
Five trees were harvested at each site in different months of the year. The values for each plant tissue are percent of total biomass or percent of total quercitol content; different letters indicate statistically significantly differences, analysis of variance (anova), Tukey honestly significantly different (HSD) 0.01.
na, data not available.
The long dry period over the summer months had a profound effect on the water relations and gas exchange of all E. astringens trees measured in this study (Fig. 2). Trees at both sites had very negative pre-dawn leaf water potentials in the summer months and high leaf water potentials in winter (Fig. 2a), and this trend was mirrored in our gas exchange data in that photosynthesis and stomatal conductance were lowest at the height of summer and highest in winter (Fig. 2b,c). Plantation trees were considerably more stressed in summer compared to woodland trees, as evidenced by the very negative pre-dawn leaf water potentials (down to −6 MPa), very low CO2 assimilation rates (<2 µmol m−2 s−1) and extremely low stomatal conductance (<10 mmol m−2 s−1) of the plantation trees in February and April 2002. In contrast, the woodland trees had less negative pre-dawn water potentials in summer (down to −4 MPa) as well as appreciable CO2 assimilation (5–10 µmol m−2 s−1) and at least fivefold greater stomatal conductance (50–70 mmol m−2 s−1). In winter, trees at both sites recovered from drought stress and had similar water potentials (−1 to −2 MPa), CO2 assimilation rates (12–23 µmol m−2 s−1) and stomatal conductance (120–240 mmol m−2 s−1).
The distinct seasonal cycle observed in leaf water relations and gas exchange was not generally reflected in the osmolality of tree tissues (Fig. 3). In woodland trees, the concentration of osmotically active substances in leaves, stems and roots did not change during the course of the year when calculated on a dry matter basis. However, when calculated on a plant water basis, the concentration of total osmolytes in woodland trees tended to be greater in the summer months compared to the winter months. The seasonal differences in osmolyte concentrations in the woodland trees indicated there was a change in tissue water content during the year, with lower tissue water content in the summer months resulting in greater tissue osmolality. In plantation trees, on the other hand, concentrations of osmotically active substances did not vary seasonally and showed the same pattern on a DW or plant water basis, except in leaves, which had significantly greater concentrations of osmotically active substances in summer compared to winter.
Substantial quercitol concentrations were measured in all plant tissues, with leaves showing the greatest concentrations (between 150 and 270 mmol kg−1 DW), and branches, stems and roots containing around 70 mmol kg−1 DW (Fig. 4). On average, quercitol contributed about 15% to the total osmotically active substances in leaves at both sites. Quercitol concentrations did not change substantially between seasons in tissues of woodland or plantation trees, except in the leaves of plantation trees, where the summer concentration of quercitol (270 mmol kg−1 DW) was double that in winter (140 mmol kg−1 DW). For these leaves, the contribution of quercitol to osmotic adjustment in response to drought was approximately 25% (calculated from the difference in the foliar concentration of total osmolytes between winter and summer).
Woodland trees and plantation trees showed marked differences in the distribution of the total amount of quercitol per tree (Table 1). Woodland trees stored around one-third of total quercitol in their leaves and one-third in the stem, while the remaining third was equally divided between branches and roots. In contrast, plantation trees stored half or more of their quercitol in their leaves, while the remaining half was roughly equally divided between the stem, branches and roots. There was no significant change in quercitol allocation among plant tissues during the year in trees at either site.
In the neutral fraction of solutes, the concentration of total soluble carbohydrates was much less than the concentration of quercitol in trees from both sites (Fig. 4). Total soluble carbohydrates did vary according to the time of measurement, but there was no clear seasonality or relationship with water deficit or gas exchange parameters (Fig. 2). Starch concentrations, on the other hand, showed a clear seasonality in trees at both sites. Starch concentrations in all tissues were low in the summer months, increased in autumn to a maximum in winter and decreased again in spring to return to a minimum in summer (Fig. 4). This pattern was most pronounced in the roots and leaves, but also evident in other tree tissues. It was noticeable that starch reserves in plantation trees were almost completely depleted in February 2002, when these trees were most drought stressed.
This study is the first comprehensive investigation of cyclitols at a whole-plant level under naturally occurring stress conditions in the field. The results of this study provide important insights into the production of quercitol during drought stress, and its functional role in maintaining plant water and carbon balances, or lack thereof. Our data point to four key findings: (1) quercitol was accumulated in large concentrations in all tree tissues of E. astringens; (2) quercitol contributed significantly to osmotic adjustment, but only under severe stress conditions; (3) the allocation of quercitol to the different tree organs did not change during the year; and (4) quercitol was not catabolized under stress and therefore did not act as a carbon storage compound. We address each of these findings in detail as follows.
This study confirms that quercitol is a major organic constituent of E. astringens. Quercitol occurred in all tissues, including leaves, branches, stems and roots, with the greatest concentrations in the leaves. Woodland trees had an average quercitol concentration of around 140 mmol kg−1 dry matter in leaves, which corresponds to 23 mg g−1 or 2.3% of dry matter, whereas trees in the plantation accumulated up to 280 mmol kg−1 or 4.6% of leaf dry matter (46 mg g−1). Quercitol concentrations in other tissues were significantly lower, but nonetheless appreciable at ∼ 1.0%. Previous studies have found similar quercitol concentrations in leaves of other eucalypt (Merchant et al. 2006a) as well as oak species (Popp et al. 1997; Passarinho et al. 2006).
This study confirms that quercitol plays an important role in osmoregulation in E. astringens. Quercitol contributed 15% of total leaf solutes under no drought stress conditions in winter (Fig. 4), and up to 25% in the leaves of plantation trees in response to drought stress, indicating that quercitol was a major contributor to osmotic adjustment. These results contrast to earlier studies of field-grown chestnut oak (Quercus prinus) where quercitol concentrations did not change in response to water deficit (Gebre & Tschaplinski 2002). While increased leaf quercitol concentrations have been reported in response to drought and salt stress in a range of Eucalyptus species in the field and in the glasshouse, including in seedlings of E. astringens, increases were not absolute (Merchant & Adams 2005; Merchant et al. 2006b). Rather, increases were relative as a result of reductions in the volume of cellular water. It has long been established that creating drought stress conditions in the glasshouse that adequately mimic field conditions is very difficult, and that results can be skewed dependent on the rate at which drought stress develops (Jones & Rawson 1979). Glasshouse studies alone do not suggest a major role for quercitol in osmotic adjustment, but it is clear from our field-based study that quercitol does play a functionally important role in E.astringens when drought stress is slower to develop, and when more sustained and intense than can be achieved under controlled conditions. Quercitol may play an important role in decreasing cell osmotic potential and thus allowing the maintenance of water absorption and cell turgor under water deficit. This will be important in leaves to guarantee the function of gas exchange and photosynthesis. Increasing concentrations of osmolytes in response to drought is perhaps of even greater significance in roots, as it is a major facilitation mechanism for water uptake from drying soils (Chaves et al. 2003). However, we note that quercitol concentrations increased only under severe stress conditions, suggesting that other mechanisms were relatively more important in preventing dehydration at moderate levels of water stress. In fact, our results point to stomatal closure as being the dominant drought avoidance/tolerance mechanism on display in this study (Fig. 2).
The different responses of E. astringens trees at the two study sites to drought stress in summer indicate that this species can employ different strategies to tolerate severe drought stress, and that responses such as quercitol accumulation are to some degree ‘plastic’. Trees in the plantation endured much lower water potentials and lower gas exchange rates compared to the woodland tress, which was a consequence of the marked differences in biomass allocation. Plantation trees had a greater leaf area at similar-sized root systems, probably because of their younger age, greater access to light and nutrients and previous access to larger amounts of stored soil water during early stages of development. Leaves of plantation trees responded to increased water deficit (an increase of −5 MPa between summer and winter) with active osmotic adjustment by increasing the concentration of osmotically active substances. Conversely, leaves of woodland trees did not exhibit active osmotic adjustment but adjusted their leaf water content to a small degree in response to a significant change in leaf water potentials between summer and winter of −3 MPa (Fig. 4). It is possible, however, that leaves of woodland trees underwent morphological changes by way of an elastic adjustment, thereby adjusting cell wall turgor in response to water deficit. More detailed studies of water relations (e.g. pV curves) are required to elucidate the exact nature of these differences.
Trees at both sites maintained high concentrations of osmolytes throughout the year in all tree organs, and increases in osmolytes were observed despite low carbon assimilation rates. This was achieved through stable osmolytes such as quercitol, decreases in starch concentrations in all organs and probably an inhibition of tree growth.
The allocation of quercitol among the different tree organs of E. astringens did not change significantly during the course of the year (Table 1). One exception was observed in plantation-grown E. astringens in summer, where an increase in the proportion of quercitol allocated to leaves coincided with a lower proportion of quercitol in the stems and roots; however, these trends were not statistically significant. Previous studies have reported significant translocation of cyclitols throughout the year in other tree species. Both quebrachitol in Acer pseudoplatanus saplings and quercitol in Q. robur saplings were translocated between roots and leaves throughout the year, and from branches and stems to leaves between spring and summer (Popp et al. 1997). However, these studies were performed on deciduous species where the cyclical emergence and shedding of leaves drive translocation of carbon and other compounds. Eucalyptus astringens is an evergreen species, and we did not observe significant adjustments in leaf area during the study period. Significant concentrations of quercitol have been observed in xylem and phloem sap of eucalypts (Merchant, unpublished results), which implies that transportation of quercitol between different tree organs is possible. However, our results clearly demonstrate that if transport did occur in E. astringens, it did not result in significant changes of quercitol allocation on a whole-tree level.
Although there were no changes in allocation of quercitol between different organs of E. astringens, we did observe significant differences in the pattern of quercitol allocation between trees at the two different sites. Plantation trees experienced greater drought stress and had a greater proportion of their total quercitol allocated to leaves (∼60%) whereas the woodland trees only allocated a third of their quercitol to leaves and more to stems (Table 1), again confirming a high degree of physiological plasticity in E. astringens in response to stress.
Early studies suggested that cyclitols may function as carbon storage compounds (Diamantoglou 1974). Our results indicate that quercitol is a highly stable osmolyte that is not catabolized under stress. Starch reserves, on the other hand, decreased with drought stress in the woodland trees, and were almost completely exhausted under severe drought stress in all tree tissues of the plantation trees (Fig. 4). Decreases in starch concentrations in response to soil drying have frequently been reported in other tree species such as Eucalyptus globulus, Quercus petraea or apple (Quick et al. 1992; Wang & Stutte 1992; Epron & Dreyer 1996), but the extent of starch exhaustion in the plantation trees is remarkable. Quercitol concentrations either remained steady or increased under stress conditions, and there was no decrease in the amount of quercitol during the year, clearly indicating that quercitol does not function as a carbon storage compound.
Compatible solute and stress metabolite
The large quercitol concentrations in all tree organs at all times of the year raise an interesting question: Why does E. astringens invest 2–4% of its dry matter in such a stable and an inert substance? It is most likely that quercitol primarily functions as a compatible solute and stable osmolyte in E. astringens. We know very little about subcellular allocation of cyclitols in plants, although early studies suggested that cyclitols are primarily located in the cytosol and chloroplasts (Paul & Cockburn 1989). However, in eucalypts, it is very unlikely that quercitol in leaves would only be located in the cytosol and chloroplasts at a concentration of 350 mmol L−1 plant water. Assuming that the cytosol occupies only around 10% of the cell volume, the actual concentration of quercitol would be 3.5 mol L−1 of cytoplasm. A concentration of that magnitude is highly unlikely given that this concentration would have to be osmotically counterbalanced by osmolytes in the vacuole. The alternative is that quercitol is located in both the vacuole and the cytoplasm in eucalypt leaves.
Quercitol distributed throughout the cell in the cytosol and the vacuole would be well positioned to act as an osmotic insurance strategy against severe environmental stresses such as drought or freezing stress. Movement of quercitol through the tonoplast would enable the deployment of a stable and compatible solute into the cytoplasm in times of high stress that would act as an osmoticum and would have osmoprotective properties. The substantial leaf gas exchange rates in the woodland trees even when under drought stress confirm that E. astringens was able to maintain significant physiological activity despite pre-dawn leaf water potential as negative as −4 MPa.
It is also worth noting that both amounts and concentrations of quercitol remained high in the winter months despite increased growth rates, suggesting that E. astringens also metabolizes quercitol in winter. Previous studies have shown that cyclitols can have cryoprotective functions (Orthen & Popp 2000), and the greatest quercitol concentrations in Quercus suber leaves were found in the winter months (Passarinho et al. 2006), indicating a role in freezing tolerance. Eucalyptus astringens experienced several frost events during the study period. Unlike drought stress, freezing stress is a spontaneous event and as such, high osmolyte and quercitol concentrations in winter may act as an insurance against occasional freezing damage in the winter months.
In conclusion, this study confirmed the importance of quercitol as a stable osmolyte and compatible solute in the drought tolerance of field-grown eucalypts. It also indicates that quercitol is only one part of a complex, plastic strategy employed by E. astringens to tolerate drought. Phenotypic plasticity, in this case adjustment of biomass allocation according to environmental constraints, is accompanied by a physiological plasticity. There was greater allocation of biomass to the leaves in the plantation-grown trees, which leads to greater carbon gain at the expense of greater drought stress in summer. However, the plantation trees were able to tolerate these long periods of drought stress with the help of quercitol as a stable osmolyte, and allocated most of the quercitol in the tree to the leaves. The smaller proportional allocation of biomass to the leaves in woodland trees reduced drought stress in summer and led to a smaller, but constant, carbon gain during the year. Quercitol contributed significantly to the osmolality in the leaves and providing a safety net against freezing stress in winter, but a smaller fraction of quercitol was allocated to the leaves in comparison to the plantation trees.
This study was supported by a University of Western Australia small research grant and the Victorian Department of Sustainability and Environment. We would like to thank Benjamin Arndt, Kenny Heard, Becky Livesley and Kelly Whyte for assistance in the field; Steve Gorton for assistance in field site selection and assistance with field trips; and two anonymous reviewers for constructive comments.