Photosynthesis, carbohydrate levels and chlorophyll fluorescence-estimated intercellular CO2 in water-stressed Casuarina equisetifolia Forst. & Forst.


Rafael Martínez-Carrasco Tel. + 34 923219606; fax: + 34 923219609; e-mail:


The effect of long-term water stress on photosynthetic carbon metabolism in Casuarina equisetifolia Forst. & Forst. was analysed by measuring CO2 assimilation, stomatal conductance, the quantum yield of photosystem II (ΦPSII), enzyme activities, and the levels of photosynthetic intermediates and carbohydrates. CO2 assimilation decreased under water stress while the intercellular CO2 concentration (Ci) as estimated by gas exchange measurements remained high. However, the estimates of Ci from measurements of ΦPSII suggest that the decrease in photosynthesis can be explained in terms of stomatal closure. Water stress decreased total stromal fructose-1,6-bisphosphatase activity and did not alter the activities and activation states of ribulose bisphosphate carboxylase oxygenase and NADP-dependent malate dehydrogenase (NADP-MDH). The concentration of photosynthetic metabolites, glucose, fructose and sucrose decreased, whereas starch concentrations increased under drought conditions.


Decreased photosynthesis under water stress is normally not attributable to restricted photochemical reactions, the photosynthetic apparatus being very resistant to desiccation (Cornic 1994). A direct inhibition of mesophyll photosynthesis in response to dehydration has been reported (Gunasekera & Berkowitz 1993; Gimenez, Mitchell & Lawlor 1992), although impaired carbon assimilation may be the result of low availability of CO2 to the chloroplast due to a water-stress-induced closure of stomata. Conventional methods of calculating the intercellular concentration of CO2 (Ci) from gas exchange data may result in an overestimation of the value of Ci if non-uniform closure of stomata occurs (Terashima 1992), although several forms of averaging Ci over regions of a leaf may be used (Farquhar 1989). Another complication when stomata close is that the larger conductance of the cuticle to water vapour than to CO2 leads to overestimations of the water-based calculations of Ci (Boyer, Wong & Farquhar 1997; Meyer & Genty 1998). When stomatal limitations are overcome by high CO2 concentrations (Quick et al. 1992; Brestic et al. 1995; Tourneaux & Peltier 1995), or when Ci is estimated from measurements of 18O exchange (Tourneaux & Peltier 1995) or chlorophyll fluorescence (Dai, Edwards & Ku 1992) it is concluded that moderate water deficits inhibit photosynthesis through a decrease in chloroplast CO2 concentrations. Fluorescence-based measurements of Ci assume that the relationships of the quantum yield of photosystem II (PSII) electron transport (ΦPSII) to carbon assimilation and Ci are the same in control leaves and those experiencing heterogeneous stomatal opening (Lal, Ku & Edwards 1996; Meyer & Genty 1998). A Ci-independent enhancement of electron flow to sinks other than carbon assimilation (e.g. the Mehler-peroxidase reaction) could change the above relationships. However, ΦPSII was not significantly increased in poplar transformants overexpressing iron superoxide dismutase (increased flow through the Mehler reaction) (Arisi et al. 1988). Moreover, under non-photorespiratory conditions, watered and dehydrated plants exhibited the same linear relationship between ΦPSII and the quantum yield of CO2 assimilation, and the relationship of ΦPSII to the rate of gross photosynthesis in 21% O2 was the same when the CO2 concentration was varied and when leaves were rapidly dehydrated (Cornic 1994).

Studies of photosynthetic carbon reduction during drought have shown that water stress has no effect on the carboxylation capacity of ribulose-1,5-bisphosphate carboxylase oxygenase (Rubisco) (Sharkey & Seemann 1989; Gimenez et al. 1992; Brestic et al. 1995; Sánchez-Rodríguez, Martínez-Carrasco & Pérez 1997) and that it decreases the ribulose-1,5-bisphosphate (RuBP) content (Sharkey & Seemann 1989; Gimenez et al 1992). The light-dependent reduction of the coupling factor reverses any observable effect of low leaf water potential on photophosphorylation (Ortíz-López, Ort & Boyer 1991). Moreover, in mildly stressed plants no decline in ATP levels was observed and the ratio of 3-phosphoglycerate (PGA) to triose-phosphates (TP) declined, also indicating that energy input into the photosynthetic carbon reduction cycle was not decreased by water stress (Sharkey & Seemann 1989). Gunasekera & Berkowitz (1993) have suggested that an enzyme involved in RuBP regeneration limits photosynthesis under water stress. In agreement with this, we have found a decrease in the total activity of stromal fructose-1,6-bisphosphatase (FBPase) activity under drought in Casuarina equisetifolia (Sánchez-Rodríguez et al. 1997).

In this paper we report the effects of long-term water stress on photosynthetic carbon assimilation and photosynthate levels in potted plants of Casuarina equisetifolia, a tree from arid regions that is able to fix atmospheric N in symbiosis with Frankia. Intercellular CO2 concentrations were estimated by a chlorophyll fluorescence method to distinguish between stomatal and non-stomatal limitations on photosynthesis, when patchy stomatal closure may occur. Fluorescence-based measurements of Ci have been used in studies on the response to abscisic acid feeding (Meyer & Genty 1998), to short-term responses to air water-vapour pressure deficits (Dai et al. 1992) and to water stress induced by withholding water (Lal et al. 1996). Here we have estimated Ci through chlorophyll fluorescence in plants under long-term water stress receiving limited weekly rewatering. The unchanged Rubisco and decreased stromal FBPase activities under drought found in a previous study (Sánchez-Rodríguez et al. 1997) are appraised along with the pool sizes of photosynthetic metabolites to determine whether carboxylation or RuBP availability limits photosynthesis under water stress.


Seeds of Casuarina equisetifolia Forst. & Forst. were surface-sterilized with 50% sodium hypochlorite, thoroughly washed with distilled water and germinated in trays containing perlite under 80 μmol m–2 s–1 illumination provided by fluorescent lights (Silvania Standard F36 W/154 Daylight) and 70% relative humidity at 25 °C under a 16 h photoperiod. The plants, which were irrigated with distilled water, were not inoculated with the Frankia microsymbiont. Ten days after germination, the seedlings were transferred to hydroponic culture in a 25% strength nutrient solution containing 4 mM CaCl2, 1·5 mM MgSO4, 0·34 mM HNa2PO4, 0·94 mM H2NaPO4, 10·1 mM KNO3, 0·1 mM iron citrate, and micronutrients (Hewitt 1966). Twelve weeks later, the seedlings were planted in 2·5 L pots containing 300 g of perlite (whose water-holding capacity was previously determined in the laboratory), placed in a glasshouse with minimum and maximum temperatures of 15 °C and 30 °C and 800 μmol m–2 s–1 maximum illumination, and supplied with a full-strength nutrient solution. Forty-six days after the plants had been transferred to pots, drought was imposed in half of the pots while the remaining plants were kept as controls. Initially, the pots with water-stressed and control plants were brought to 25 and 75% maximum water-holding capacity, respectively. At weekly intervals, the pots were weighed and water and nutrient solution were added to replace evapotranspiration losses. There were two blocks, each consisting of 16 replicate plants of each treatment.

For measurements of relative water contents [RWC = (fresh weight – dry weight) × 100/(fully turgid weight – dry weight)], recently matured branchlets from six replicate plants per block were detached and their fresh weights were recorded. They were then floated on water for 16 h, the full turgor weight recorded and, after drying at 60 °C for 24 h, were weighed again (Turner 1981).

Between 75 and 102 d after the start of drought, gas exchange was measured in three replicate plants. Before the measurements, which were performed with an infrared gas analyser (LCA-2; Analytical Development Co., Hoddesdon, UK) with differential operation in an open system, each plant was irrigated to the set water content of the root medium, transferred to a growth chamber with 300 μmol quanta m–2 s–1 (incandescent and cool white fluorescent lamps) at 25 °C and 79% relative humidity and allowed to equilibrate. Air was mixed with CO2 from a gas-cylinder, humidified by passage through a bubbler, and adjusted for the set humidity (1200–1400 Pa vapour-pressure deficit) and CO2 concentration by passage through columns filled with silica gel and soda lime, respectively. The mass flow of the resulting gas was measured using a mass-flow regulator (ASUM; Analytical Development Co., Hoddesdon, UK). The temperature of the leaf chamber [PLC-2 (N); Analytical Development Co., Hoddesdon UK] and the humidity of the gas stream entering and leaving the chamber were measured with the in-built thermistor and capacitance sensors. Recently matured branchlets were placed in the chamber, avoiding mutual shading, and were homogeneously illuminated either with 650 or 1200 μmol quanta m–2 s–1 provided by halogen lamps. The CO2 concentrations were varied from 50 to 800 μmol mol–1. The leaf area of the enclosed branchlet was measured by analysis of digital images (Cyclops; Confocal Technologies, Liverpool, UK) recorded with a still-video camera (Ion-250; Canon, Spain). The quantum yield of photosystem II electron transport (ΦPSII), defined as (FmFs)/Fm (where Fm and Fs are the maximal and steady-state fluorescence values under illumination, respectively) was measured along with gas exchange under 650 μmol quanta m–2 s–1 by adapting the fibre optics of a modulated fluorometer (PAM-2000; Waltz, Effeltrich, Germany) to the leaf chamber and following the procedure described by Genty, Briantais & Baker (1989). Calculations of net CO2 assimilation (A), transpiration, stomatal conductance and Ci were based on the equations of von Caemmerer & Farquhar (1981). The relationship between ΦPSII/A and Ci measured in well-irrigated plants, where stomatal conductance was high and accurate calculations of Ci from gas exchange data were available, was used to estimate Ci under water stress from the measured ΦPSII/A ratios (Dai et al. 1992).

At 104 d from the start of drought (day 0), two samples were harvested during the central hours of the photoperiod. The first sample consisted of four replicate plants per block, which were separated into green branchlets, stem and root. A subsample of the green branchlets was weighed and the projected area measured by image analysis. The total green area was estimated from this subsample and the fresh weight of all the green branchlets. Then, all the plant fractions were oven dried at 80 °C for 24 h and weighed. A second sample, consisting of green branchlets detached from six replicate plants from each block, was immediately transferred to liquid nitrogen under illumination and then kept in liquid nitrogen until enzyme activities and amounts of carbohydrates, metabolites and chlorophyll were analysed.

Metabolites from branchlets stored in liquid nitrogen were extracted in 1 M HClO4 as described by Labate, Adcock & Leegood (1990). The amounts of glucose-6-phosphate (G6P), fructose-6-phosphate (F6P), fructose-1,6-bisphosphate (FBP), PGA, TP (glyceraldehyde 3-phosphate and dihydroxyacetone phosphate) and RuBP were determined enzymatically according to Kobza & Edwards (1987), except that TP levels were analysed in a fluorometer with 360 nm and 485 nm excitation and emission wavelengths, respectively. Starch in the residue from the HClO4 extraction was digested with amyloglucosidase and amylase and determined according to ap Rees et al. (1977). For measurements of glucose, fructose and sucrose, frozen branchlets were extracted with 4·2 cm3 extraction medium [1·2 cm3 CHCl3, 2·4 cm3 CH3OH, 0·6 cm3 buffer containing 50 mM NaF, 10 mM EGTA and 50 mM Hepes (pH 8·5)] as described by Labate et al. (1990) and analysed according to Jones, Outlaw & Lowry (1977). The recovery of representative amounts of metabolites added to the extracts was greater than 87%. Chlorophyll (phaeophytin) in the perchloric acid extract was analysed by the method of Vernon (1960) and that present in the chloroform extract by the method of Arnon (1949).

Subsamples of the stored green branchlets were ground in liquid nitrogen and extracted in buffer for enzyme assays. An aliquot of the whole extract was used to determine chlorophyll contents (Arnon 1949) and the remainder was centrifuged at 13 000 g. The total time from extraction to the assay of initial enzyme activity was less than 2·5 min. NADP-MDH and stromal FBPase were extracted in the buffer described by Harbinson & Foyer (1991) except that 1% Triton X-100 (v : v) was used. Rubisco was extracted in 100 mM Bicine-NaOH (pH 7·8) containing 10 mM MgCl2, 10 mM 2-mercaptoethanol, 2% PVPP (w : v), 1% BSA (w : v) and 1% Triton X-100 (v : v). This high concentration of Triton X-100 was found to be necessary to obtain good enzyme activity owing to the presence of large amounts of phenols, tannins and other compounds, which may inhibit or precipitate the enzymes (Pitel & Cheliak 1985). Initial and total NADP-MDH, FBPase and Rubisco activities were measured in a spectrophotometer at 340 nm as described by Scheibe, Fickenscher & Ashton (1986), Holaday et al. (1992) and Sharkey, Savitch & Butz (1991), respectively. Activation states were calculated by expressing initial activities as percentages of total activity. Checks were made for linearity of enzyme activities over time and for proportionality between rate and amount of extracts.

Analyses of variance and comparisons of regression lines of A versus Ci were made according to Snedecor & Cochran (1967). The SIMFIT (W.G. Barsley, University of Manchester) package was used to fit a function to the ΦPSII/A versus Ci data.


The green area per plant at 104 d after the start of drought showed a four- to 22-fold increase over the 24 ± 3 cm2 value at day 0. Water stress decreased the green area and dry matter of plants at 104 d by 79 and 73%, respectively (Table 1). Drought stress had little effect on specific leaf area (the area/dry weight ratio for the green branchlets) and significantly decreased the leaf area ratio (green area/total plant dry weight), indicating that less dry matter was invested in the production of green area. Water stress decreased the shoot/root ratio (Table 1). The relative water content of green tissues was decreased by drought to a varying extent depending on the day in the irrigation cycle, although it did not reach the low values found in other studies (Brestic et al. 1995). Table 1 shows the relative water contents at 104 d. The difference between control and stressed plants was larger at other dates [e.g. 81·2 and 77·7, standard error of the difference between two means (s.e.d.) 1·16, respectively, at 45 d]. The rate of water consumption by plants was decreased by drought (Fig. 1). Water consumption increased with time in irrigated plants, probably as a result of greater plant size. By contrast, water use in the stressed plants varied little over time.

Table 1.  . Green area (cm2 plant–1), total dry weight (g plant–1), specific leaf area (SLA, cm2 g–1), leaf area ratio (LAR, cm2 g–1), ratio of shoot to root dry weights, and relative water content (RWC, %) of droughted and well-watered Casuarina equisetifolia plants 104 d after the start of drought. Each value is the mean of eight replicates (12 for RWC) Thumbnail image of
Figure 1.

. Time course (days from the start of the drought period) of water consumption by droughted (closed symbols) and well-watered (open symbols) Casuarina equisetifolia plants.

Drought decreased A with CO2 concentrations in the bulk air (Ca) higher than about 100 μmol mol–1 (Fig. 2). Increasing the photon flux density led to an increase in photosynthesis in control plants, but drought inhibited the response of A to increased irradiance. By contrast, ΦPSII was similar for both control and water-stressed plants at any given Ca under moderate light intensity (data not shown), resulting in higher ΦPSII/A ratios under water stress (Fig. 3). To distinguish between the stomatal and non-stomatal effects of drought on photosynthesis, avoiding the possible artefacts caused by differential cuticular conductance to water and CO2 and/or by non-uniform stomatal closure, Ci was estimated from measurements of the relationship of ΦPSII/A to Ci in control plants {y = 0·033 ( ± 0·0032) + {1·65 [± 0·47] exp[–0·026 ( ± 0·0031) x]}, R2 = 0·92} and the ΦPSII/A ratio for water-stressed plants. Figure 4 shows the relationships to Ca of the Ci estimated by this procedure (Ci’) and the Ci calculated with gas exchange parameters in drought-stressed and control plants. Whereas Ci increased linearly with Ca and was similar in both drought-stressed and control plants, Ci’ showed little increase with Ca above approximately 300 μmol mol–1 and was lower than Ci for control plants with a Ca value higher than about 150 μmol mol–1. The relation of photosynthesis to Ci’ in plants under drought conditions (Fig. 5) was linear across the whole experimental range (up to approximately 200 μmol mol–1). In control plants, which had Ci values up to about 600 μmol mol–1, a typical asymptotic response of A to Ci was found. Regressions were fitted to the linear portion of the ACi response of control and drought-stressed plants (Fig. 5). Statistical comparison (Snedecor & Cochran 1967) of these regressions revealed that the A versus Ci values for the two treatments were very close, the regression slope being significantly, although only slightly, greater for water-stressed plants. This indicates that for any given Ci below about 200 μmol mol–1, photosynthesis under water deficit is similar to that of control plants.

Figure 2.

. Response of A to varying Ca under 650 (triangles) and 1200 (squares) μmol m–2 s–1 photon flux density in branchlets of droughted (closed symbols) and well-irrigated (open symbols) Casuarina equisetifolia plants. Data from three replicate plants.

Figure 3.

. Changes in the ΦPSII/A ratio with varying Ca in branchlets of droughted (□) and well-watered (□) Casuarina equisetifolia plants under 650 μmol m–2 s–1 photon flux density. Data from three replicate plants.

Figure 4.

. Changes in Ci calculated from gas exchange (squares) and from the ΦPSII/A ratio (triangles) with varying Ca in branchlets of droughted (closed symbols) and well-watered (open symbols) Casuarina equisetifolia plants under 650 μmol m–2 s–1 photon flux density. Data from three replicate plants.

Figure 5.

. Response of A to Ci in branchlets of drought-affected (□) and well-watered (□) Casuarina equisetifolia plants under 650 μmol m–2 s–1 photon flux density. The values of Ci under water stress were calculated from the ΦPSII/A ratio. The regression lines for droughted (filled line) and irrigated (dashed line) plants are: y = –3·75 (± 0·63) + 0·066 (± 0·004) x, r2 = 0·84, and y = –3·33 (± 0·42) + 0·059 (± 0·003) x, r2 = 0·978, respectively. The value of F for the difference between slopes was 17·9 (P < 0·01) and that for the difference between elevations was 0·73 (not significant). Data from three replicate plants.

Water stress had no effect on chlorophyll a and b contents (data not shown). Drought decreased the leaf levels of glucose and fructose and tended to decrease that of sucrose (Table 2). By contrast, starch levels increased significantly under drought, and thus the sucrose/starch ratio (Table 3) was drastically lowered. The concentration of intermediate carbon metabolites (Table 2) also decreased under drought, although the decreases in PGA and FBP were not significant. Water stress slightly, but significantly, increased the G6P/F6P and FBP/F6P ratios (Table 3). The RuBP/PGA ratio was not affected by drought, and the TP/PGA and TP/RuBP ratios showed a non-significant decrease.

Table 2.  . Carbohydrate (mmol m–2) and photosynthetic metabolite (μmol m–2) levels in droughted and well-watered Casuarina equisetifolia plants 104 d after the start of drought. Starch is presented as glucose equivalents. Each value is the mean of eight replicates Thumbnail image of
Table 3.  . Ratios of photosynthetic metabolites in droughted and well-watered Casuarina equisetifolia plants 104 d after the start of drought. Each value is the mean of eight replicates Thumbnail image of

Drought had no significant effect on the activities or activation states of NADP-MDH and Rubisco (Table 4). By contrast, initial and total stromal FBPase activities were reduced under drought whereas the activation state of this enzyme was not affected by water stress.

Table 4.  . Effects of water stress on the initial and total activities (μmol s–1 m–2) and the activation states of NADP-dependent malate dehydrogenase (NADP-MDH), stromal fructose-1,6-bisphosphatase (FBPase) and ribulose-1,5-bisphosphate carboxylase oxygenase (Rubisco) in Casuarina equisetifolia plants 104 d after the start of drought. Data are means of eight (NADP-MDH) and 12 (FBPase and Rubisco) replicates Thumbnail image of


The decrease in photosynthetic carbon assimilation found under drought conditions implies reductions in photosynthetic electron transport or in the rate of the Rubisco-catalysed reaction, the latter effect being due to lowered Rubisco activity or to a reduced availability of RuBP or CO2. The observed inhibition of carbon assimilation under drought was not accompanied by parallel decreases in the PSII quantum yield, suggesting an increased partitioning of electrons to oxygen (Lawlor & Fock 1977; Dai et al. 1992; Cornic 1994; Brestic et al. 1995), probably in the Mehler-peroxidase reaction rather than in photorespiration (Biehler & Fock 1996). This flux of electrons to oxygen may support PSII activity and avoid an over-reduction of the stroma, as suggested by the unaltered activation state of NADP-MDH (Scheibe et al. 1986) under drought (Table 4). This contrasts with the increased reduction of the acceptor sides of photosystem I (unchanged activation, but increased total NADP-MDH activity) and PSII (decreased qP) and with the lower ΦPSII observed under drought in previous studies on Casuarina (Sánchez-Rodríguez et al. 1997); the reasons for these differences between the two studies are not clear. The notion that the photosynthetic apparatus is very resistant to drought (Cornic 1994) is supported by the similarity of the ΦPSII values in control and water-stressed plants in the present experiments.

Water stress did not affect the activity or the activation state of Rubisco (Table 4) which therefore did not limit photosynthesis, in agreement with other studies (Sharkey & Seemann 1989; Gimenez et al. 1992; Brestic et al. 1995; Sánchez-Rodríguez et al. 1997). Total stromal FBPase activity was inhibited by drought, as observed previously in Casuarina (Sánchez-Rodríguez et al. 1997). In addition, although the FBP level (Table 2) did not increase but rather decreased, the lowered F6P level was accompanied by an increased FBP/F6P ratio (Table 3), consistent with a decreased FBPase activity. Decreased levels of RuBP and other intermediate metabolites under drought conditions (Table 2) have also been reported by others (Sharkey & Seemann 1989; Gimenez et al. 1992). However, the unchanged TP/RuBP, TP/PGA and RuBP/PGA ratios under drought (Table 3) did not point to one-sided limitations to carbon flux through the Calvin cycle. In our experiments, drought increased the cytosolic compartmentation of hexose phosphates (higher G6P/F6P ratio) and caused a decrease in the levels of sucrose and other soluble sugars (Table 2), in agreement with Lawlor & Fock (1977) and with Huber, Rogers & Mowry (1984), who found a stronger inhibition of photosynthesis than of carbohydrate export from leaves. By contrast, drought increased the levels of starch. On the contrary, Quick et al. (1992) observed decreases in starch, increases in sucrose and restricted export under drought. Further analyses of carbohydrate contents throughout the day are required to assess the effects of water stress on leaf carbon export and partitioning between sucrose and starch in Casuarina.

In contrast with the linear increase in Ci calculated from gas exchange data under drought, the Ci estimated from measurements of the ΦPSII/A ratio did not vary with Ca increases above approximately 300 μmol mol–1 (Fig. 4), consistent with the finding (Quick et al. 1992; Tourneaux & Peltier 1995) that very high CO2 concentrations (2–10%) in the air are required to increase the intercellular CO2 concentration in stressed leaves and to be able to distinguish between stomatal and non-stomatal effects of drought. The unchanged Ci of drought-stressed plants when Ca was increased shows that calculations from gas exchange overestimate the internal CO2 concentration owing to non-uniform stomatal closure (Terashima 1992) or, more likely, to differences in cuticular conductance to water and CO2 under water stress. (Lal et al. 1996; Boyer et al. 1997; Meyer & Genty 1998). The low value of Ci estimated from the ΦPSII/A measurements (Fig. 4) indicates that drought induces an important decrease in CO2 availability. The similarity of the A versus Ci relationships in control and drought-stressed leaves when Ci was estimated from the measured ΦPSII/A ratios (Fig. 5) suggests, in agreement with other studies (Sharkey & Seemann 1989; Cornic 1994; Tourneaux & Peltier 1995), that a reduced CO2 influx would be the main cause of the limitation of net CO2 assimilation during long-term water stress. Further measurements of ΦPSII and A at low O2 (Cornic 1994) and spatial resolution of ΦPSII from chlorophyll fluorescence images (Meyer & Genty 1998) can be used for improving the accuracy of calculations of Ci and gas exchange parameters with non-uniform stomatal closure.


This work was funded by the European Commission, STD3 programme (TS3*-ct91–0027). J.S.-R. was supported by a grant from the CSIC.