Cyclitols and carbohydrates in leaves and roots of 13 Eucalyptus species suggest contrasting physiological responses to water deficit



    Corresponding author
    1. School of Forest and Ecosystem Science, The University of Melbourne, Water St, Creswick, Victoria, 3363, Australia, and
    2. School of Biological, Earth and Environmental Sciences, The University of New South Wales, Sydney, NSW 2052, Australia
      A. Merchant. Fax: +61 3 5321 4166; e-mail:
    Search for more papers by this author

    1. School of Forest and Ecosystem Science, The University of Melbourne, Water St, Creswick, Victoria, 3363, Australia, and
    Search for more papers by this author

    1. School of Forest and Ecosystem Science, The University of Melbourne, Water St, Creswick, Victoria, 3363, Australia, and
    Search for more papers by this author

    1. School of Biological, Earth and Environmental Sciences, The University of New South Wales, Sydney, NSW 2052, Australia
    Search for more papers by this author

A. Merchant. Fax: +61 3 5321 4166; e-mail:


In many tree species, physiological adaptations to drought include the accumulation of osmotically active substances and/or the presence of particular compatible solutes, among them cyclitols. Recently, the cyclitol quercitol was identified in species of Eucalyptus, a diverse genus whose speciation is probably driven by adaptation to water availability.

We subjected seedlings of 13 Eucalyptus species from different ecosystems (‘mesic’ and ‘xeric’) and different sub-generic taxonomic groups to 10 weeks of water deficit (WD) treatment. Pre-dawn water potentials (ψpdwn) and relative water content (RWC) were determined in shoots, and total osmolality, soluble low-molecular-weight carbohydrates and cyclitols were measured in leaves and roots.

Responses to water deficit followed two distinct patterns: Eucalyptus species from ‘mesic’ environments adjusted concentrations of sucrose (through increased levels of sucrose and decreases in RWC) in response to water deficit, whereas ‘xeric’ species increased concentrations of quercitol (through reductions in RWC). In root tissues, only species from xeric environments contained high levels of quercitol and mannitol, increasing under WD conditions.

We suggest that the former (mesic) strategy may be beneficial to respond to short-lasting drought conditions, because sucrose is easily metabolized, whereas the latter (xeric) strategy may relate to an effective acclimation to longer-lasting drought. These physiological response groups are also related to taxonomic groups within the genus.


The genus Eucalyptus encompasses more than 700 species among 15 subgenera, majority of which form the subgenera Corymbia (∼ 70 spp.), Eucalyptus (∼ 110 spp.) and Symphyomyrtus (∼ 500 spp.). The ecological range of the genus reaches from regions with an annual rainfall of 250 mm in southern or 350 mm in northern Australia (Gillison 1994), to the wet sclerophyll forests of eastern and south-western Australia where rainfall exceeds 1500 mm year−1. The subgenus Corymbia dominates savannas in northern Australia, while species of the Eucalyptus subgenus dominate most of the coastal and upland regions of south-east and south-west Australia (Gill, Belbin & Chippendale 1985). Species of the subgenus Symphyomyrtus are widely distributed across the continent, particularly in more arid habitats and regions subject to prolonged periods of water deficit (WD) (for review, see Boland 1992). It was suggested that the distribution of Eucalyptus species is particularly dependent on the availability of water (Adams 1996). While many species of the genus are known for their tolerance of arid conditions, there has been strikingly little research into attributes – physical, chemical or biological – that might confer such tolerance.

For all trees, maintaining growth and surviving under dry conditions depend on a number of physiological processes and physical attributes. By definition, such features are less crucial for species endemic to more mesic regions and so within the Eucalyptus genus, we expect a high diversity in traits related to drought tolerance. Indeed, from the limited data available for eucalyptuses, changes in biomass allocation (Rawat & Banerjee 1998), stomatal control (Macfarlane, White & Adams 2004), cell wall reinforcement (Ladiges 1975), cell wall water storage (Tuomela 1997) and cellular osmolarity (Myers & Neales 1986; Lemcoff et al. 1994; Chen, Keiper & Filippis 1998; Lemcoff et al. 2002) have all been documented as responses to arid conditions. A number of authors (Ladiges 1975; Clayton-Greene 1983; Myers & Neales 1986; Lemcoff et al. 1994; Stoneman, Turner & Dell 1994; White, Turner & Galbraith 2000) investigated the contribution of changes in osmotic potential to drought and salt stress tolerance in Eucalyptus species. Regulation of osmotic potential allows plants to maintain cell turgor and thus, growth under drought conditions (for reviews, see Morgan 1980; Turner & Jones 1980; Chaves, Maroco & Pereira 2003). For Eucalyptus, many authors recorded (mainly through the indirect method of analysis of pressure–volume curves) that osmotic potential decreased (became more negative) in response to drought stress by between 0.1 and 0.8 MPa (Lemcoff et al. 1994; Stoneman et al. 1994; Tuomela 1997; Li 1998; White et al. 2000). These decreases may be attributed to one or both of two mechanisms: (1) active accumulation of solutes – commonly referred to as ‘osmotic adjustment’ (Turner & Jones 1980; Chaves et al. 2003); or (2) the increase in concentration of constitutive solutes through reductions in cellular water.

Among species, a wide range of solutes are responsible for changes in osmotic potential. The osmotic contribution of stress-induced accumulations of major inorganic ions (K+, Na+, Ca2+, Mg2+) and/or carbohydrates involved in primary metabolism (sucrose, glucose or fructose) vary, owing to the metabolic impact of substantial changes in pool sizes. Instead, several ‘compatible solutes’, slightly different to primary metabolites, were identified, particularly in drought-tolerant plants (for review, see Chaves et al. 2003). Cyclitols, represented by compounds such as quercitol (Paul & Cockburn 1989), pinitol (Nguyen & Lamant 1988) or quebrachitol (Popp et al. 1997), are highly stable and prevalent among the candidate classes of compatible solutes for woody plants.

We have recently shown that cyclitols accumulate to osmotically significant concentrations in leaves of a range of Australian tree species (Adams et al. 2005; Merchant & Adams 2005; Merchant et al. 2006), raising suggestions for their function as cellular osmotica. In vitro, cyclitols show additional beneficiary properties such as the ability to protect tertiary protein structures and antioxidant capacity (see Orthen, Popp & Smirnoff 1994; Popp et al. 1997; Orthen & Popp 2000). The occurrence of quercitol in Eucalyptus species seems related to both the phylogenetic relationships within the genus and distribution of species with regard to water availability (Merchant et al. 2006). More specifically, Eucalyptus species originating from low-rainfall environments (< 500 mm year−1) contain substantial amounts of quercitol in leaf tissues (Merchant et al. 2006). We have also shown previously that experimental exposure to salinity and WD lead to increased quercitol concentrations in leaves of the dry land species Eucalyptus spathulata (Merchant & Adams 2005). Questions arise whether cyclitols are: (1) related to physiological responses to drought in a wider range of ecologically different Eucalyptus species; (2) whether cyclitols are present and accumulate in leaf and root tissues; and (3) whether drought responses and cyclitol accumulation show phylogenetic relationships that match the corresponding ecological niches of Eucalyptus species.

To address these questions, we subjected seedlings of a range of Eucalyptus species of different taxonomic groups and from different environments to an experimental drought treatment, and characterized water relations and low-molecular-weight metabolite (carbohydrates and cyclitols) concentrations in leaves and roots.


Plant selection

In the selection of species for the present experiment, the following aspects were taken into account: (1) species should encompass a diverse taxonomy representing major subgenera and series classifications to include taxa found to contain quercitol in leaves and others where no detectable amounts of quercitol were observed (see Merchant et al. 2006); (2) species should originate from contrasting environments (particularly regarding rainfall), and encompass variation in potential growth rate. Species selected are shown in Table 1 and are subsequently referred to as originating from ‘mesic’ (> 500 mm annual rainfall) or ‘xeric’ (< 500 mm annual rainfall) environments.

Table 1. Eucalyptus species selected for this study
SpeciesSubgenusSectionSeriesEnvironmental classificationQuercitol (Merchant et al., 2006)Growth habitMin Annual Rainfall (mm)Distribution
Eucalyptus maculata Hook.CorymbiaSeptentrionalesNavicularesMesicNo. . . tall straight tree (20–45+ m)a800Common in coastal regions . . . mainly in tall open forests on slopesb
Eucalyptus obliqua L Her’EucalyptusEucalyptusRegnantesMesicNo. . . a medium-sized to very tall tree (15–60 m)a600Common and widespread in moister open forestsb
Eucalyptus macrorhyncha F. MuellEucalyptusCapilliusPachyphloiusMesicN/AA small- to medium-sized tree . . . to 30 ma500. . . on drier, well-drained hilly sites (foothills and tablelands to about 1000 mm)b
Eucalyptus rubida Deane and MaidenSymphyomyrtusMaidenariaViminalesMesicN/Afrom small to taller (to 35 m) forests of rangesa800Mainly on drier and shallower soils of the foothills and tablelands (up to 1400 m)b
Eucalyptus ovata LabillSymphyomyrtusMaidenariaFoveolataeMesicNoSmall- to medium-sized treea600Common on periodically swampy flats and poorly drained swamps and hollowsb
Eucalyptus globulus LabillSymphyomyrtusMaidenariaGlobularesMesicNoUsually a tall tree (25–60 m)a600Distributed in . . . moist or dry hill country (below subalps)b
Eucalyptus camaldulensis DehnhSymphyomyrtusExsertariaRostrataePhreatophyticNoMedium . . . impressive tree (to 45 m) . . . open spreading crowna200along inland rivers or dry watercourses and floodplains . . . preferring deep, moist subsoils with clay contentb
Eucalyptus polyanthemos SchauerSymphyomyrtusAdnatariaHeterophloiaeXericYesSmall- to medium-sized tree (8–25 m) . . .a500drier open forests . . . up to 650 m mostly on shallow soils of sedimentary origin . . . on hillsides, gullies or open flatsb
Eucalyptus tricarpa L.A.S JohnsonSymphyomyrtusAdnatariaMelliodoraeXericYes  (Eucalyptus sideroxylon,  see note)Medium-sized tree (10–30 m)a500 . . . in drier, open forests on well-drained skeletal soils (often gravely with quartz on low ridges or adjacent flat country)b
Eucalyptus cladocalyx F. MuellSymphyomyrtusSejunctae XericYesA small to large tree (8–35 m)a400Natural distribution on sandy soils of Kangaroo Island and Flinders Rangesa
Eucalyptus astringens MaidenSymphyomyrtusBisectaeErectaeXericYesSmall to fairly tall tree (8–35 m)a300Tree of south-western Australia typically on lateritic soils of low nutrient statusa
Eucalyptus viridis R.T.BakerSymphyomyrtusAdnatariaBuxealesXericYesSmall tree to 10 ma300Inland Australiaa typical on low fertile soilsb
Eucalyptus oleosa F. MuellSymphyomyrtusBisectaeSubulataeXericNot testedSmall tree (3–8 m)a200Chiefly on brown, loamy soils of flatter areasa

Seeds of Eucalyptus species were sown in early August 2003 and raised for 8 weeks at the Mildura Native Plant Nursery, Mildura (34.18° S, 142.16° E). The seedlings were then transferred to and raised in an open-ended shade house in Creswick, Victoria Australia (37.43° S, 143.89° E) for 3 months prior to the commencement of treatments. Irradiance inside the shade house was 85–90% of ambient (in the 400–700 nm range), and temperatures were a maximum of 2 °C higher than outside. The seedlings were potted into 9 L pots in a mix of 50:50 peat/coarse river sand (v/v). Macronutrients and micronutrients were supplied with the addition of 5 g of the slow-release fertilizer Nutricote (Yates, Homebush, NSW, Australia) per litre of potting mix. The plants were watered daily to soil capacity prior to the commencement of treatments.

Experimental design

Twelve uniform plants from each species were selected and randomly assigned to two treatments, ‘well watered’ (WW) and ‘water deficit’ (WD). WW plants were watered to field capacity of the potting mix at all times. WD-treated plants received 20% of the water volume used by WW plants (measured gravimetrically). Watering was conducted on a daily basis at dusk. Pots were arranged in a fully randomized fashion and were rotated so as to minimize incident light effects and any non-random sources of error.

Plant height, diameter, leaf area and total leaf weight measurements

Plant height, diameter, leaf area and total leaf weight measurements were made after 10 weeks of treatment. Height was measured as the distance between the stem base and the shoot tip. Stem diameter was measured 2 cm above the soil surface with electronic callipers precise to 0.1 mm. Leaf area was measured with a leaf area meter (model 3100, Li-Cor, Lincoln, NE, USA) precise to 0.1 mm2.

Collection of leaf and root samples

Samples were collected immediately before the treatment started (WW0 and WD0) and after 10 weeks of treatment (WW10 and WD10). Leaf samples were collected from six trees per group, consisting of the youngest fully expanded leaf on terminal or first-order branchlets. Because of the diversity of growth habit, the height of collected foliage varied considerably. The leaf samples were placed in 2 mL microtubes and transferred immediately to liquid nitrogen. In the laboratory, the samples were microwaved (Popp et al. 1996) for 30 s, in a 650 W conventional microwave oven (Panasonic, Secaucus, NJ, USA) and oven-dried at 85 °C for 48 h. Samples were then stored at −86 °C.

Leaf water characteristics

At the time of each sampling, pre-dawn shoot water potential (ψpdwn) of each tree was measured on a first-order branch using a Scholander pressure bomb (PMS Corvallis, Albany, OR, USA) as per Scholander et al. 1965. Leaf relative water contents (RWC) were determined gravimetrically. Three leaves from each plant were removed, sliced into 5 mm sections and weighed (fresh weight, FW). These were then placed in deionized water at 4 °C for 12 h to allow full hydration. The leaf samples were then lightly blotted dry with tissue paper and reweighed (saturated weight). The samples were then dried at 36 °C for 48 h and reweighed (dry weight, DW). RWC was calculated as a proportion based on the formula:


Solute analysis

Approximately 40 mg of dried plant material was weighed into a 2 mL screw-capped microtube. One millilitre of hot methanol–chloroform–water (MCW, 12:5:3, v/v/v) was added and incubated at 80 °C for 30 min. The water fraction of the extraction mix included a 0.1% (w/v) solution of internal standard. The internal standard used was a mixture of 0.1% penta-erythritol (Pierce Endogen, Rockford, IL, USA) and 0.1% xylitol (Pierce Endogen) for gas chromatography (GC) analysis. External standards of quercitol were made from solutions supplied by A. Richter (University of Vienna, Austria)

After cooling, the samples were centrifuged (11 400 g), and 800 µL of the supernatant was removed and placed into a 2 mL round-bottomed microtube. A further 200 µL of chloroform and 500 µL of deionized water were added to facilitate the separation of phases. The samples were left to stand for 15 min to allow phase separation and then centrifuged at 11 400 g for 3 min. Seven hundred microlitres of the aqueous supernatant (called MCW extract) was then transferred to a 1.5 mL microtube to which 300 µL of mixed bed resin (MBR) had already been added. MBR consisted of one part Dowex 1 × 8 (anion exchange, Cl- form; Dow Chemical Company, Michigan, USA) and one part Dowex 50W (cation exchange, formate form). The samples were agitated for 2 h at room temperature. Following pulse centrifugation, 400 µL of the supernatant was transferred to a microtube and stored at −80 °C until analysis.

Carbohydrates were separated and quantified using GC. Sixty microlitres of deionized MCW extracts was dried and resuspended in 400 µL of anhydrous pyridine to which 50 µL of trimethylchlorosilane (TMCS)/bis-trimethylsilyl-trifluoroacetamide (BSTFA) mix (1:10 v/v; Sigma Aldrich, Sydney, Australia) was added. The samples were incubated for 1 h at 75 °C and analysed by GC within 24 h. GC analysis was performed using a Shimadzu 17A series gas chromatograph (Shimadzu Corporation Limited, Columbia, MD, USA) using a DB1 (30 m) column. Split injection was made at 300 °C, with an initial oven temperature program of 60 °C for 2 min ramping to 300 °C at a rate of 10 °C min−1 and maintained for 10 min. Column flow rate was 1.5 mL min−1. Peak integration was made using Class VP analysis software (Shimadzu Corporation Limited).

Concentrations of solutes in leaf tissue are reported in units of both mmol g−1 DW and mmol L−1 leaf water (using RWC measurements to calculate tissue water content). Concentrations of solutes in root tissues are given in mmol g−1 DW only.

Measurements of total osmolality

Total cellular osmolality was measured using freeze-point depression from 4% (w/v) hot water extracts. Approximately 40 mg of dried ground leaf tissue (as prepared earlier) was placed in a 2 mL screw-capped microtube to which 1 mL of hot water (80 °C) was added. The samples were incubated for 30 min at 80 °C and agitated twice during this period. The samples were left to cool to room temperature, centrifuged and the top 800 µL was re-moved to a microtube. Osmolality of the solution was measured using an OSMOMAT 030 cryoscopic osmometer (Gonotec, Berlin, Germany).

Statistical analysis

Effects of treatments on tissue water relations and solute concentrations were analysed by analysis of variance using Statistica analytical software (version 6, StatSoft, Tulsa, OK, USA). Drought treatment and species were used as fixed effects; sampling date (before and after 10 weeks of treatment) was treated as repeated measures factor. Homogeneity of variance was tested by Levene’s test; it was assured that means and standard deviations were not correlated across groups, and distributions of residuals were visually checked for any deviation from normality. P-values were calculated using Fisher’s least square difference post hoc test.


Leaf water characteristics

The ψpdwn showed a significant reduction in all species in the WD treatment (Table 2) after 10 weeks. Leaf RWC was significantly reduced compared to control values in WD plants of all species except for Eucalyptus maculata and the two species originating from the lowest rainfall environments –Eucalyptus viridis and Eucalyptus oleosa (Table 2). Values of ψpdwn and leaf RWC at the beginning of the experiment (WW0 and WD0) did not differ from those of WW plants at the end of the experiment (WW10) for any species (statistics not shown).

Table 2.  Pre-dawn water potential (ψpdwn, MPa) and relative water content (RWC, proportion) of 13 Eucalyptus species in well-watered (WW) and water deficit (WD)-treated plants at week 0 (prior to the commencement of water deficit treatment, 18 December 2003) and at week 10 (post-water deficit treatment, 2 March 2004)
ψpdwn Speciesψpdwn at week 0ψpdwn at week 10Significance
Week 10
Eucalyptus maculate−0.25 ± 0.04−0.23 ± 0.03−0.29 ± 0.05−1.79 ± 1.42***
Eucalyptus obliqua−0.19 ± 0.04−0.19 ± 0.04−0.40 ± 0.05−3.00 ± 0.37***
Eucalyptus macrorhyncha−0.16 ± 0.02−0.21 ± 0.04−0.47 ± 0.22−2.82 ± 0.68***
Eucalyptus camaldulensis−0.28 ± 0.04−0.23 ± 0.03−0.35 ± 0.09−2.64 ± 1.00***
Eucalyptus ovata−0.29 ± 0.08−0.23 ± 0.05−0.82 ± 0.49−3.61 ± 0.31***
Eucalyptus globulus−0.22 ± 0.03−0.26 ± 0.01−0.37 ± 0.08−2.45 ± 0.84***
Eucalyptus rubida−0.21 ± 0.02−0.28 ± 0.07−0.40 ± 0.07−3.72 ± 0.29***
Eucalyptus polyanthemos−0.25 ± 0.04−0.28 ± 0.07−0.39 ± 0.15−4.34 ± 0.64***
Eucalyptus tricarpa−0.31 ± 0.04−0.40 ± 0.09−0.35 ± 0.09−1.96 ± 1.08***
Eucalyptus cladocalyx−0.33 ± 0.04−0.33 ± 0.05−0.55 ± 0.21−4.04 ± 0.41***
Eucalyptus astringens−0.36 ± 0.06−0.33 ± 0.09−0.50 ± 0.07−4.66 ± 0.51***
Eucalyptus viridis−0.32 ± 0.02−0.36 ± 0.07−0.72 ± 0.24−2.65 ± 1.78***
Eucalyptus oleosa−0.27 ± 0.02−0.28 ± 0.04−0.36 ± 0.21−1.50 ± 0.88***
RWCSpeciesRWC at week 0RWC at week 10Significance
Week 10
  1. Means ± SE. Asterisks indicate significant differences (***< 0.001) between WW and WD (n = 6). All plants at week 0 were watered to field capacity of the potting material.

E. maculata0.95 ± 0.020.94 ± 0.020.91 ± 0.050.84 ± 0.08
E. oblique0.95 ± 0.010.92 ± 0.070.91 ± 0.060.65 ± 0.04***
E. macrorhyncha0.91 ± 0.030.89 ± 0.030.92 ± 0.070.74 ± 0.07***
E. camaldulensis0.94 ± 0.020.94 ± 0.050.93 ± 0.080.77 ± 0.13***
E. ovata0.79 ± 0.350.91 ± 0.050.91 ± 0.050.66 ± 0.08***
E. globulus0.95 ± 0.010.90 ± 0.030.91 ± 0.060.78 ± 0.15***
E. rubida0.97 ± 0.020.90 ± 0.040.96 ± 0.010.57 ± 0.05***
E. polyanthemos0.96 ± 0.030.87 ± 0.050.96 ± 0.030.66 ± 0.1***
E. tricarpa0.95 ± 0.020.92 ± 0.040.96 ± 0.030.84 ± 0.04***
E. cladocalyx0.89 ± 0.050.92 ± 0.010.94 ± 0.010.68 ± 0.05***
E. astringens0.71 ± 0.460.93 ± 0.020.94 ± 0.030.58 ± 0.04***
E. viridis0.80 ± 0.350.79 ± 0.360.93 ± 0.070.84 ± 0.12
E. oleosa0.95 ± 0.020.95 ± 0.020.92 ± 0.040.84 ± 0.06

Growth characteristics

After 10 weeks, plant height and diameter were significantly lower in WD treatments compared with WW for most species with the exceptions of Eucalyptus tricarpa, Eucalyptus astringens and E. oleosa (height), and E. tricarpa and E. oleosa (stem diameter) (Table 3). The leaf area of WD seedlings was always significantly less than that of WW seedlings, except for E. oleosa. Ratios of leaf area to mass were also reduced in WD versus WW treatments for Eucalyptus obliqua, Eucalyptus globulus, Eucalyptus rubida, Eucalyptus polyanthemos and E. astringens. Other species showed no difference between WW and WD treatments. Proportional distribution of growth towards leaf, stem and roots was not affected by treatment, with the exception of Eucalyptus ovata (lesser mean stem mass), E. oleosa (greater root mass) and E. rubida (lesser stem, increased root mass) (data not shown).

Table 3.  Average (x) SE and significance (P) of differences between treatment groups (WD10–WW10) in average height, diameter at 2 cm height, total leaf area and leaf area to weight ratio of 13 Eucalyptus species subject to well-watered and water deficit treatments over a 10-week-long period (n = 6)
SpeciesHeight (cm)Diameter (cm)Leaf area (m2)Leaf area : weight
  1. Significance of the difference between treatments was obtained from Fisher’s least square difference post hoc test.

Eucalyptus maculate−15.332.880.004−3.770.910.028−0.400.050.0000.220.920.872
Eucalyptus obliqua−31.836.760.010−3.970.850.015−0.570.080.0018.761.630.004
Eucalyptus machrorhyncha−24.506.460.023−3.800.560.001−0.340.030.0001.881.470.415
Eucalyptus camaldulensis−52.4312.570.021−4.340.800.006−0.480.050.0002.112.340.555
Eucalyptus ovata−36.506.400.002−4.930.580.000−0.820.080.000−1.685.270.846
Eucalyptus globulus−34.477.500.022−4.630.460.000−0.860.110.0018.592.300.037
Eucalyptus rubida−37.837.230.007−3.530.660.004−0.440.030.00010.971.140.000
Eucalyptus polyanthermos−38.805.870.003−3.200.670.025−0.300.030.0007.161.060.004
Eucalyptus tricarpa−40.1712.900.066−2.700.850.053−
Eucalyptus cladocalyx−24.202.470.000−4.120.640.005−0.420.030.0001.311.620.598
Eucalyptus astringens−8.833.810.137−2.470.530.014−
Eucalyptus viridis−33.973.860.000−2.150.630.046−
Eucalyptus oleosa−12.508.640.362−2.130.810.103−

Changes in total osmolality of leaf tissues

Total tissue osmolality increased significantly in leaves of WD plants compared to WW plants over the 10-week-long period, with the exception of E. globulus and the two most xeric species, E. viridis and E. oleosa (Fig. 1e). Differences in cellular osmolality between WW and WD plants were generally between 500 and 1000 mmol L−1 leaf water (Fig. 1e).

Figure 1.

Change in solute concentrations of 13 Eucalyptus species from the beginning to the end of the experiment in well-watered (WW, ▪) and water deficit (bsl00005)-treated plants. (a) quercitol; (b) myo-inositol; (c) glucose; (d) sucrose; (e) total osmolality. Bars represent means values (n = 6). Error bars represent 1 SE. Asterisks indicate significant differences (**P < 0.01; ***P < 0.001) from WW plants. All plants at week 0 were watered to field capacity of the potting material.
E. macror, Eucalyptus macrorhyncha; E. camal, Eucalyptus camaldulensis.

Changes in carbohydrate and cyclitol concentrations

At the beginning of the experiment, the leaves of all plants contained detectable levels of the cyclitols quercitol and myo-inositol and the major sugars glucose and sucrose (Fig. 1). There were no significant differences in concentrations among species. Other common sugars (mannitol, fructose, etc.) were detected in low amounts both pre- and post-treatments.

The responses of leaf sugar and cyclitol contents to the WD treatment revealed two different patterns distinguishing two groups of Eucalyptus species (Fig. 1a–d). One group – composed of E. polyanthemos, E. tricarpa, Eucalyptus cladocalyx, E. astringens and E. viridis– responded to the WD treatment with significant increases in quercitol concentrations in leaf water by 80–100 mmol L−1 (Fig. 1a), but mostly (except for E. cladocalyx) no changes in sucrose concentrations and other investigated sugars and cyclitols (Fig. 2b–d). Species in this group are all from xeric environments (‘xeric species’ as defined in Table 1, generally from areas with an annual rainfall below 500 mm).

Figure 2.

Solute concentrations and total osmolality in leaves of 13 Eucalyptus species in well-watered (WW, ▪) and water deficit (bsl00005)-treated plants at the end of the 10 weeks treatment. (a) quercitol; (b) myo-inositol; (c) glucose; (d) sucrose; (e) total osmolality. Bars represent mean values (n = 6). Error bars represent 1 SE. Asterisks indicate significant differences (** = P < 0.01, *** = P < 0.001) from WW plants. All plants at week 0 were watered to field capacity of the potting material.
E. macror, Eucalyptus macrorhyncha; E. camal, Eucalyptus camaldulensis.

A second group of species –E. maculata, Eucalyptus macrorhyncha, Eucalyptus camaldulensis, E. ovata and E. globulus– showed no change of quercitol concentrations relative to the WW seedlings (Fig. 1a), but responded to WD treatment with increases in sucrose concentrations in leaf water by up to 150 mmol L−1 (Fig. 1d) and with a tendency towards increased myo-inositol concentrations (Fig. 1b). Species belonging to this group are all from higher rainfall areas (> 500 mm year−1) or dependent on accessible groundwater (‘mesic species’, Table 1).

Contributions of sugars and cyclitols to total osmolality

Under WW conditions, quercitol, glucose, sucrose and myo-inositol together contributed between 30 and 50 mmol L−1 in leaf water to overall cellular osmolality of approximately 1000 mM (3–5%) across all species (Fig. 2). Under WD conditions, the most notable change was that quercitol contributed between 100 and 150 mmol L−1 to the overall leaf osmolality of 1500 mmol L−1 in xeric species (5–10%), but only between 0 and 3.5% in mesic species. In mesic species, sucrose accounted for between 100 and 200 mmol L−1 to an overall osmolality of around 1700 mmol L−1 (5–10%).

Characteristics of solute accumulation

The increase of sucrose concentration in the leaf water of mesic species (with the exception of E. maculata and E. obliqua) is attributed to the ‘active’ accumulation of sucrose as demonstrated by increases in concentration on both a leaf water basis and on a dry weight basis (Fig. 3).

Figure 3.

Comparison of leaf water-based (a) and dry weight (DW) (b) sucrose concentrations in leaves of pot-grown Eucalyptus in well-watered (WW, bsl00077) and water deficit (WD, bsl00005)-treated plants at week 0, and WW (bsl00008) and WD-treated (▪) plants after 10 weeks of WD treatment. Only those species showing no significant responses of leaf quercitol concentrations to the treatment (see Fig. 2) are included. Bars represent mean values (n = 6). Error bars represent 1 SE. Asterisks represent significance of the difference between WD-treated plants at week 0 and after 10 weeks of treatment (*P < 0.1; **P < 0.01). All plants at week 0 were watered to field capacity of the potting material.
E. macror, Eucalyptus macrorhyncha; E. camal, Eucalyptus camaldulensis.

In contrast, increases in quercitol in xeric species were observed only on a leaf water basis (Fig. 4), illustrating that the increase in concentration was a result of reductions in cellular water.

Figure 4.

Leaf water (a) and dry weight (DW) (b) quercitol concentrations in leaves of pot-grown Eucalyptus in well-watered (WW, bsl00077) and water deficit (WD, bsl00005)-treated plants at week 0, and WW (bsl00008) and WD-treated (▪) plants after 10 weeks of stress treatment. Only those species showing significant responses of leaf quercitol concentrations to the treatment (see Fig. 2) are included. Bars represent mean values (= 6). Error bars represent 1 SE. Asterisks represent significance of the difference between WD-treated plants at week 0 and after 10 weeks of treatment (*P < 0.1; ** = P < 0.01). All plants at week 0 were watered to field capacity of the potting material.

Carbohydrates and cyclitols in roots

With the exceptions of E. obliqua, E. camaldulensis, E. astringens and E. viridis, the investigated species showed significant increases in total concentration of osmotically active substances in root tissues in response to drought treatment (Figure 5). All species had approximately the same level of root osmolality between 1200 and 1500 mmol g−1 root DW in WW plants and between 1500 and 2000 mmol g−1 root DW in WD-treated plants.

Figure 5.

Concentrations of total osmotically active solutes in roots of 13 Eucalyptus species in well-watered (WW, ▪) and water deficit (bsl00005)-treated plants after 10 weeks of treatment. Bars represent means values (n = 6). Error bars represent 1 SE. Asterisks indicate significant differences (**P < 0.01; ***P < 0.001) from WW plants. All plants at week 0 were watered to field capacity of the potting material.
DW, dry weight; E. macror, Eucalyptus macrorhyncha; E. camal, Eucalyptus camaldulensis.

Quercitol was present in root tissues of xeric species at concentrations between 20 and 200 mg g−1 DW (Fig. 6a), and mannitol in concentrations between 10 and 85 mg g−1 root DW (Fig. 6c). These compounds were almost entirely absent in roots of mesic species. In contrast, roots of mesic species contained between 10 and 40 mg g−1 DW myo-inositol, while xeric species contained little to no myo -inositol in root tissues (Fig. 6b). Concentrations of fructose, glucose and sucrose showed no such trends (Fig. 6d–f).

Figure 6.

Carbohydrate and cyclitol concentrations in roots of 13 Eucalyptus species in well-watered (WW, ▪) and water deficit (bsl00005)-treated plants after 10 weeks of treatment. (a) quercitol; (b) myo-inositol; (c) mannitol; (d) fructose; (e) glucose; (f) sucrose. Bars represent means values (n = 6). Error bars represent 1 SE. Asterisks indicate significant differences (*P < 0.1; **P < 0.01; ***P < 0.001) from WW plants. All plants at week 0 were watered to field capacity of the potting material.
DW, dry weight; E. macror, Eucalyptus macrorhyncha; E. camal, Eucalyptus camaldulensis.

Differences in root quercitol concentrations of between 50 and 150 mmol g−1 DW were detected between WD10 and WW10 plants of xeric species with the exception of E. tricarpa (Fig. 6a). This pattern was also seen in mannitol concentrations of E. oleosa, E. viridis and E. polyanthemos, which increased by between 30 and 80 mg g−1 root DW (Fig. 6c). In mesic species, average myo-inositol concentrations were higher in WD10 plants compared with WW10 ranging between 5 and 25 mg g−1 root DW (Fig. 6b). No consistent trends in root glucose, fructose and sucrose levels were observed in response to WD treatment, with values varying between 10 and 150 mg g−1 DW (Fig. 6d–f).


Osmoregulation in leaves

Our data reveal that different compounds contribute to overall changes in concentrations of osmotically active substances among these eucalypt species. Significant concentrations (up to 150 mmol L−1) of quercitol were found in leaf water of E. polyanthemos, E. tricarpa, E. cladocalyx, E. astringens, E. viridis and E. oleosa (all species considered as ‘xeric’; Table 1). In contrast, most species regarded as more ‘mesic’– namely E. maculata, E. macrorhyncha, E. camaldulensis, E. ovata, E. globulus and E. rubida– contained no quercitol under WD conditions; however, they showed equally significant concentrations of sucrose (up to 200 mmol L−1) in leaf water.

In addition to the contribution of quercitol and sucrose to osmotic potential, the contrasting biochemistry of these compounds (and their accumulation) has important consequences for their function in plant tissues. While ‘osmotic adjustment’ (defined as the accumulation of solutes under WD; Turner 1986) can be achieved in both cases, sucrose is a vital primary metabolite, readily metabolized, and therefore not a stable osmoticum. In E. globulus, it was shown that during periods of drought stress, concentrations of sucrose may rapidly increase (Quick et al. 1992) presumably because of: (1) increased activity of sucrose–phosphate synthase (Quick et al. 1992) at the expense of starch biosynthesis; (2) decreased rate of export from the cell either by compartmentalization away from transport loading sites and/or a consequence of inhibition of phloem transport (Quick et al. 1992); or (3) starch hydrolysis. Importantly, sucrose can be metabolically recycled to facilitate growth during periods of relief from stress. Such characteristics may be advantageous under the temporary nature of osmotic stress experienced by trees from high-rainfall areas.

In contrast to sucrose, quercitol and other cyclitols are highly stable and metabolically inactive because of the absence of reactive aldehyde or ketone groups. We hypothesize that for low-rainfall Eucalyptus species, the continuation of physiological activity during prolonged dry conditions requires the presence of such stable osmotica. The stability of cyclitols – such as quercitol – improves their function as an osmolyte as they do not undergo short-term fluctuations (Paul & Cockburn 1989).

This pattern supports previous data (Adams et al. 2005; Merchant et al. 2006) that Eucalyptus species from contrasting environments may adopt contrasting responses to arid conditions. White et al. (2000) recently suggested that inherently low osmotic potentials may confer advantages in low-rainfall environments. A review of previous investigations reveals that inherently lower water potentials are commonly observed in eucalyptuses growing in low-rainfall environments, including Eucalyptus melliodora, Eucalyptus microcarpa (Clayton-Greene 1983), E. polyanthemos, Eucalyptus behriana, Eucalyptus macrocarpa (Myers & Neales 1986), Eucalyptus microtheca (Tuomela 1997) and Eucalyptus leucoxylon, (White et al. 2000). Our data suggest that such osmotic adaptations are specifically related to changes in quercitol and sucrose concentrations. Our tentative conclusion is that we have identified a solute (quercitol) responsible for significant species and edaphic condition-dependent variation in osmotic potential in eucalyptuses.

Subcellular compartmentalization of sucrose and quercitol further enhances osmolytic significance (see Paul & Cockburn 1989; Popp et al. 1997). Quantification of cytoplasmic osmotic potential during WD is difficult given the current inability to determine the extent of subcellular compartmentalization of solutes. To date, only Paul et al. (1989) has presented evidence of subcellular compartmentalization of cyclitols. He showed accumulations of up to 230 mol m−3 in the chloroplast and about 100 mol m−3 in the cytosol of the herbaceous halophyte Mesembryanthemum crystallinum. Concentrations of this magnitude make a highly significant contribution to osmoregulation (Popp et al. 1997). For trees more generally, including those investigated here, cyclitols commonly reach over 10 times these concentrations (Popp et al. 1997).

Measurement of total cellular osmolality (osmoles of solute per kilogram of solvent) is an approximation of the sum of individual solute concentrations. Nevertheless, significant changes in solute concentrations are reflected in cellular osmolality. We showed that in all but three species (E. globulus and the two most xeric species, E. viridis and E. oleosa), the pool of osmotically active substances in leaves increased in response to WD, as might be expected from previous studies on osmotic relations (see Morgan 1980; Turner & Jones 1980; Chaves et al. 2003). Increases in tissue concentrations of osmotically active substances can be attributed to either an accumulation of solutes and/or decreased water content. It is noteworthy that water deficit led to significantly decreased RWC in all but three species –E. maculata, and the two lowest rainfall species, E. oleosa and E. viridis. The lack of significant increases in cellular osmolality in E. oleosa and E. viridis are thus largely explained by the maintenance of high RWC. For these two species, changes in osmotic potential may arise as a result of changes in cell wall elasticity as drought-tolerant species tend to have lower bulk modulus of elasticity (BEM) (Clayton-Greene 1983) such as those found in Eucalyptus platypus (White et al. 2000). Mallees – a particular growth form of Eucalyptus– are generally regarded as particularly adapted to dry conditions. The two mallee species included in this experiment (E. viridis and E. oleosa) originate from low-rainfall environments (Table 1) and display structural mechanisms such as thick cuticles, high leaf surface to volume ratios and sunken stomata, congruent with water conservation.

In E. globulus, the absence of changes in cellular osmolality was surprising given that: (1) it is native to high-rainfall and cool environments; and (2) reductions in osmotic potential have been reported in previous studies (White, Beadle & Worledge 1996; Pita & Pardos 2001). Previous studies have also detected significant decreases in transpiration and stomatal conductance in E. globulus under drought conditions (Quick et al. 1992; Pita & Pardos 2001; Pita, Gasco & Pardos 2003), a response that may be common to related species such as Eucalyptus grandis (Whitehead & Beadle 2004). In our study, stomatal sensitivity to stress seems the likely explanation of the ability of E. globulus to maintain a relatively high RWC.

In every water deficit experiment, the imposition of a consistent level of physiological drought across species that differ in structure and physiology, is a difficult, if not impossible task. Species-specific variation in growth habits and water use frequently serves to confound experimental designs. For example, while ψpdwn is often used as a surrogate measure of physiological stress, there are numerous associated difficulties (Flexas & Medrano 2002). In isohydric species (such as grapevine), leaf water potential may remain high despite severe stress (Flexas & Medrano 2002) while in our experiment with 13 eucalyptuses, ψpdwn dropped significantly by between 2 and 4 MPa in all species. Reductions in ψpdwn of this magnitude encompass the range of osmotic pressures experienced by a number of Eucalyptus species under field conditions (eg, White et al. 1996, 2000). The applied treatments generally reduced (but did not stop) growth – a response often associated with the need to maintain turgor through a diversion of carbon to non-growth processes (Chaves et al. 2003). Hence, this experiment allowed study of acclimation processes critical to survival and growth under water-limited conditions.

Osmoregulation in roots

Increasing concentrations of osmotically active substances in response to water deficit is perhaps of even greater significance in roots, as it is a major facilitation mechanism for water uptake from dry(ing) soils (Chaves et al. 2003). The present study showed increases in concentrations of low-molecular-weight metabolites in root tissues of Eucalyptus species subjected to WD. The accumulation of quercitol and mannitol accounted for approximately 10–20% of observed solute accumulation (on a DW basis) in root tissues of xeric species. Previous detections of reductions in osmotic potential in root tissues of Pinus radiata (Zou, Sands & Sun 2000), Pinus pinaster (Nguyen & Lamant 1989), Pinus banksiana, Picea glauca (Koppenaal, Tschaplinski & Colombo 1991) and Prunus avium × pseudacerasus and Prunus cerasus (Ranney, Bassuk & Whitlow 1991) indicate that such a response mechanism may be common to a number of tree species.

The only comparative data for eucalyptuses are those of Morabito et al. (1996), who demonstrated increased K+ concentrations in root tissues of E. microtheca during osmotic stress. In that study, the influence of the rooting medium and potential confounding by nutrient status precluded clear conclusions. We note that expression of solute concentrations on a DW basis limits our ability to determine the overall contributions to cellular osmolarity. Nevertheless, quercitol represents approximately 10% of the DW of osmotically active substances in root tissues of low-rainfall species and quercitol and mannitol together accumulate in response to water deficit. These solutes therefore play significant roles in the amelioration of the effects of low external water potentials.

Potential role in chemotaxonomy

Interestingly, seedlings of all species in the present study contained at least some quercitol in leaf or root tissues at the beginning of the experiment. The major distinction was that ‘xeric’ species from low-rainfall environments maintained high quercitol concentrations during growth under drought stress. Assuming that quercitol accumulation is beneficial, it is the retention of quercitol in leaf tissues that provides a link with aridity tolerance. Quercitol has been isolated previously in drought/salt-tolerant Eucalyptus species (Adams et al. 2005), and initial results of a genus-wide screening of the occurrence of cyclitols among Eucalyptus species corroborates a link to drought adaptation within the genus (Merchant et al. 2006). The presence of quercitol in tissues of all 13 Eucalyptus species selected in this study, even those from mesic environments, indicates that while the physiological and genetic capacity to synthesize this compound is widely present in eucalyptuses, it is expressed differently in species from differing habitats.

Several authors have attempted to delineate Eucalyptus species based on initial growth rates (Davidson & Reid 1980; Duff, Reid & Jackson 1983), foliar nutrient concentrations (Lambert & Turner 1983), volatile leaf oils (Li, Madden & Potts 1995, 1996) and respiratory metabolism (Anekonda et al. 1999) and combinations of these parameters (Noble 1989). It is likely, however, that the geographical distribution of Eucalyptus species is dependent on a variety of these factors, but particularly related to the availability of water (Adams 1996). Our results suggest the inclusion of an additional taxonomic component of the responses of quercitol concentrations to WD.

In conclusion, we can answer the questions addressed in the present study as follows: (1) cyclitols respond to drought exposure, but responses are different in different groups of Eucalyptus species. In ‘xeric’ species, tissue concentrations of quercitol on a leaf water basis increased upon exposure to WD, and contributed to an overall increase in leaf osmolality. On a leaf DW basis, quercitol concentration in leaves decreased during the growth of WW plants and WD-treated plants of ‘mesic’ species, whereas it was maintained in water deficit treated ‘xeric’ species; (2) cyclitols (mainly quercitol) are present and accumulate in both leaves and roots of these ‘xeric’ species; and (3) seedlings of all 13 Eucalyptus species investigated contained at least low amounts of cyclitols. The constitutive presence and significant concentration of quercitol in eucalyptus leaf and root tissues demonstrated here point towards a biochemical link between taxonomy, physiology and acclimation to aridity.


We specifically thank the Landcare Nursery in Creswick for providing space for the experiment. Charles Warren is cordially thanked for assistance with the watering regime along with Najib Ahmady, Douglas Rowell, Andrew Callister and Katherine Whittaker for their assistance with physiological measurements. This work was supported by funding from the School of Forest Ecosystem Science (SFES). SFES is supported by the Victorian Governments Department of Sustainability and Environment and The University of Melbourne. A. Merchant gratefully acknowledges the support of a School of Forest and Ecosystem Science scholarship.