Volume 175, Issue 1 p. 81-93
Free Access

Photosynthetic limitations in response to water stress and recovery in Mediterranean plants with different growth forms

Jeroni Galmés

Jeroni Galmés

Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Spain

Search for more papers by this author
Hipólito Medrano

Hipólito Medrano

Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Spain

Search for more papers by this author
Jaume Flexas

Jaume Flexas

Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Spain

Search for more papers by this author
First published: 23 April 2007
Citations: 429
Author for correspondence: Jeroni Galmés
Tel: +34 971 259556
Fax: +34 971 173184
Email: [email protected]


  • Whether photosynthesis is limited during water stress and recovery because of diffusive or biochemical factors is still open to debate, and apparent contradictions appear when various studies on species with different growth forms are compared.

  • Ten Mediterranean species, representing different growth forms, were subjected to different levels of water stress, the most severe followed by rewatering. A quantitative limitation analysis was applied to estimate the effects of water stress on stomatal (SL), mesophyll conductance (MCL) and biochemical limitations (BL).

  • Results confirmed a general pattern of photosynthetic response to water stress among C3 plants when stomatal conductance (gs) is used as a reference parameter. As gs values decreased from a maximum to approx. 0.05 mol H2O m−2 s−1, the total photosynthetic limitation rose from 0 to approx. 70%, and this was caused by a progressive increase of both SL and MCL limitations, while BL remained negligible. When lower values of gs were achieved (total photosynthetic limitation increased from 70 to 100%), the contribution of SL declined, while MCL still increased and BL contributed significantly (20–50%) to the total limitation.

  • Photosynthetic recovery of severely stressed plants after rewatering showed a dominant role of MCL, irrespective of the degree of photosynthesis recovery.


Low water availability is considered the main environmental factor limiting plant growth and yield in semiarid areas (Boyer, 1982). The water stress-induced limitation on plant growth is mainly caused by reductions in plant carbon balance, which depends on the balance between photosynthesis and respiration (Flexas et al., 2006a). The response of photosynthesis to water stress has received considerable attention in the past, and there has been a long-standing controversy regarding which is the primary limitation on photosynthesis: stomatal closure or metabolic impairment (Chaves, 1991; Lawlor, 1995; Cornic & Massacci, 1996). In recent years, efforts have been made to generalize the responses to water stress of photosynthetic parameters in higher plants (Flexas & Medrano, 2002; Lawlor & Cornic, 2002; Chaves et al., 2003). As a result, there is now some consensus that diffusion limitations on photosynthesis predominate under most water-stress situations. These limitations involve not only stomatal closure, but also decreased mesophyll conductance to CO2 (gi), an important but sometimes neglected process (Roupsard et al., 1996; Flexas et al., 2002; Ennahli & Earl, 2005). Regardless of the species analysed, a general failure of metabolism occurs only when daily maximum stomatal conductance (gs) drops below 0.1 mol H2O m−2 s−1 (Flexas et al., 2004; Grassi & Magnani, 2005).

However, this general response pattern has been tested mostly in crops, and few data are available for natural vegetation of different origins. Because natural environments offer a range of microhabitats and ecological niches, it is likely that particular adaptations can be found, among which exceptions to the general rule may exist (Schulze, 1988). The Mediterranean climate is characterized by a hot, dry period in summer and a cool, wet period in winter, as well as by high interannual variability. The variability and unpredictability of precipitation impose strong constraints on plants and could represent an important evolutionary pressure (Joffre et al., 1999). As a consequence, natural vegetation from the Mediterranean area seems an appropriate genetic background to search for adaptations that may represent exceptions to the established pattern of photosynthesis response to water stress. The natural vegetation of the Mediterranean area has developed an array of adaptations to water stress, resulting in a high diversity of growth forms. The vegetation consists of deep-rooted evergreen sclerophyll trees and shrubs, which tolerate and/or avoid water stress and maintain green leaves during the summer drought period; semideciduous shrubs, which lose some of their leaves during summer; and geophytes and winter annual and biennial herbs, which escape drought by finishing their annual cycle before summer (Ehleringer & Mooney, 1983). In addition to this diversity of morpho-phenological forms, there is a strong diversity in ecophysiological traits that are likely to be of adaptive value, such as the specificity factor of Rubisco (Galmés et al., 2005a); the response of relative growth rate and its components to water stress (Galmés et al., 2005b); or leaf water relations and stomatal control (Galmés et al., 2006). A primary objective of the present study was to test the generality of the pattern of photosynthetic response to water stress described above, using the natural plant diversity of the Mediterranean area. While gas-exchange analysis of photosynthetic limitations under water stress in Mediterranean plants has been investigated previously in evergreen sclerophyll and summer semideciduous shrubs (Tenhunen et al., 1985; Harley et al., 1986; Harley et al., 1987a, 1987b; Gulías et al., 2002; Peña-Rojas et al., 2004), none of these studies took variations in gi into account, therefore conclusions from these studies regarding biochemical limitations must be viewed with care.

On the other hand, the carbon balance of a plant enduring a water-stress period may depend as much on the rate and degree of photosynthetic recovery as on the rate and degree of photosynthetic decline during water depletion. While many studies have addressed different aspects of photosynthetic limitations during water-stress imposition, analyses of the photosynthetic limitations during photosynthetic recovery after water stress are scarce (Miyashita et al., 2005; Flexas et al., 2006a). An early study by Kirschbaum (1987, 1988) suggested that photosynthesis during recovery was colimited by incomplete stomatal opening and a metabolic component. Recently, Ennahli & Earl (2005) have suggested that limited recovery of photosynthetic biochemistry was the most important limitation for photosynthetic recovery in cotton plants subjected to severe water stress. Therefore another objective of the present work was to perform an analysis of photosynthetic limitations after rewatering different species exposed to severe water stress. In particular, in view of the recently highlighted importance of decreased gi in the regulation of photosynthesis during water stress, we test the hypothesis that limited recovery of gi after rewatering may contribute to incomplete recovery of photosynthesis.

Materials and Methods

Plant material

Ten Mediterranean species occurring naturally in the Balearic Islands were selected for this study (Table 1), representative of different growth forms and leaf habits: two evergreen sclerophyll shrubs (Pistacia lentiscus and Hypericum balearicum), two evergreen sclerophyll semishrubs (Limonium gibertii and Limonium magallufianum), three summer semideciduous shrubs (Lavatera maritima, Phlomis italica and Cistus albidus), two perennial herbs (Beta maritima ssp. maritima and B. maritima ssp. marcosii), and an annual herb (Diplotaxis ibicensis).

Table 1. Species studied, family and brief description
Species Family Description
Diplotaxis ibicensis Pau Brassicaceae Annual herb, endemic to the Balearic Islands and inhabiting a few coastal locations.
Beta maritima L. ssp. marcosii A. Juan & M. B. Crespo Chenopodiaceae Perennial herb. Endemic to the Balearic Islands, inhabiting a few small islets subjected to strong saline spray.
Beta maritima L. ssp. maritima Chenopodiaceae Perennial herb inhabiting coastal ecosystems. Widespread in Mediterranean and temperate climates.
Lavatera maritima Gouan Malvaceae Semi-deciduous shrub up to 2 m, densely covered with hairs. Inhabits coastal locations.
Phlomis italica L. Labiatae Semi-deciduous shrub up to 1 m, densely covered with hairs.
Endemic to the Balearic Islands. The biggest populations are found 500 m above sea level, where they coexist with C. albidus.
Cistus albidus L. Cistaceae Semi-deciduous shrub up to 1 m. Commonly found in the Mediterranean garigue. Leaves densely covered with hairs.
Hypericum balearicum L. Guttiferae Woody evergreen shrub up to 2 m, endemic to the Balearic Islands.
Largest populations found in the garigue 500 m above the sea level, where it competes with P. lentiscus.
Pistacia lentiscus L. Anacardiaceae Woody evergreen shrub up to 5 m, commonly found in the Mediterranean garigue.
Limonium magallufianum L. Llorens Plumbaginaceae Woody evergreen semishrub, in cushion-like rosettes. Endemic to the Balearic Islands, inhabiting just one coastal marsh located in Magalluf, Mallorca.
Limonium gibertii (Sennen) Sennen Plumbaginaceae Woody evergreen semishrub, in cushion-like rosettes. Occurring in west Mediterranean rocky and sandy coastal areas.

Plants were grown outdoors at the University of the Balearic Islands (Mallorca, Spain) in pots (25 l, 40 cm high) containing a 40 : 40 : 20 mixture of clay-calcareous soil, horticultural substrate (peat) and pearlite (granulometry A13). The experiment was performed in five rounds, each with a pair of the species at the same time. Plant ages at time of measurement differed because of the different life cycles of the species selected. Plants of P. lentiscus, H. balearicum, C. albidus, P. italica and L. maritima were 3 yr old; plants of L. magallufianum and L. gibertii were 1.5 yr old; and plants of D. ibicensis, B. maritima ssp. marcosii and B. maritima ssp. maritima were 6 months old at the onset of the experiments.

Four weeks before starting the experiment, 10 plants per species were placed in a controlled growth chamber with a 12-h photoperiod (26°C day: 20°C night) and a photon flux density at the top of the leaves of approx. 600 µmol m−2 s−1. Plants were irrigated daily with 50% Hoagland's solution. Measurements corresponding to control treatments were made during the first day of the experiment, when all the plants were well watered. Thereafter, irrigation was stopped in five plants for each species. Pots were weighed every day to determine the amount of water loss. The water available for plants with respect to the control was determined after measurement of soil dry weight in four samples representative of the substrate mixture used in the experiment. Measurements were made on days 4, 8 and 13–17 after the last irrigation, when plants were subjected to mild, moderate and severe water stress, respectively. Severe water stress was considered to be when stomatal conductance (gs) was close to zero, which was achieved 13–17 d after withholding water, depending on the species. At this time, pots were rewatered to field capacity, and the extent of photosynthesis recovery was determined on the next day. Control plants were watered daily throughout the experiment and measured every 5–6 d to ensure they had maintained constant values.

Chlorophyll fluorescence measurements

Chlorophyll fluorescence parameters were measured on attached leaves using a portable pulse amplitude modulation fluorometer (PAM-2000, Walz, Effeltrich, Germany). For each sampling time and treatment, six measurements were made on different plants.

A measuring light of approx. 0.5 µmol photon m−2 s−1 was set at a frequency of 600 Hz to determine, at predawn, the background fluorescence signal (F0). To obtain maximum fluorescence (Fm), saturation pulses of approx. 10 000 µmol photon m−2 s−1 were applied for 0.8 s. The maximum quantum efficiency of PSII was calculated as Fv/Fm = (Fm – Fo)/Fm. At mid-morning, the steady-state fluorescence signal (Fs) and the steady-state maximum fluorescence yield (inline image) were determined on the same leaves measured at predawn, using an actinic photon flux density approx. 1500 µmol m−2 s−1. The PSII photochemical efficiency (inline image, Genty et al., 1989) was then calculated as:

image(Eqn 1)

and used for calculation of the relative linear electron transport rate (ETR) according to Krall & Edwards (1992):

image(Eqn 2)

where PPFD is the photosynthetically active photon flux density, α is the leaf absorptance, and β is the distribution of absorbed energy between the two photosystems. β was assumed to be 0.5 (the actual factor has been described as ranging between 0.4 and 0.6; Laisk & Loreto, 1996). Leaf absorptances were determined for all 10 species in 10 replicates on leaves of well irrigated plants, using a spectroradiometer coupled to an integration sphere (UniSpec, PP-Systems, Amesbury, MA, USA). A value of 0.84 was obtained for all species, except for C. albidus and P. italica (0.74 and 0.77, respectively). Potential changes in leaf absorptance with water stress were not assessed but, because changes in chlorophyll content were nonsignificant (data not shown), they were assumed to be small and to induce no important biases in the calculations of ETR.

Gas-exchange measurements

Light-saturated net CO2 assimilation rates (AN) and stomatal conductance (gs) were measured at mid-morning on attached, fully developed young leaves of four to five plants per species and treatment, using a gas-exchange system (Li-6400, Li-Cor, Lincoln, NE, USA) equipped with a light source (6200-02B LED, Li-Cor). Environmental conditions in the leaf chamber consisted of a photosynthetic photon flux density of 1500 µmol m−2 s−1, a vapour pressure deficit of 1.0–1.5 kPa, an air temperature of 25°C and an ambient CO2 concentration (Ca) of 400 µmol mol−1 air.

After inducing steady-state photosynthesis, the photosynthesis response to varying substomatal CO2 concentration (Ci) was measured. The Ca was lowered stepwise from 360 to 50 µmol mol−1 and then returned to 360 µmol mol−1 to re-establish the initial steady-state value of photosynthesis. The Ca was then increased stepwise from 360 to 1500 µmol mol−1. Gas-exchange measurements were determined at each step after maintaining the leaf for at least 5 min at the new Ca. Measurements consisted of 12–13 measurements for each curve. ANCi curves were transformed to ANCc curves, as described in the following section.

Estimations of CO2 concentration at the site of carboxylation and mesophyll conductance

From combined gas-exchange and chlorophyll fluorescence measurements, the CO2 concentration in the chloroplasts (Cc) was calculated according to Epron et al. (1995). This model works on the assumption that all the reducing power generated by the electron transport chain is used for photosynthesis and photorespiration, and that chlorophyll fluorescence gives a reliable estimate of the quantum yield of electron transport. Thus the ETR measured by chlorophyll fluorescence can be divided into two components:


(Eqn 3)

where ETRA is the fraction of ETR used for CO2 assimilation, and ETRP is the fraction of ETR used for photorespiration. ETRA and ETRP can be solved from data of AN, the rate of nonphotorespiratory CO2 evolution in the light (RL) and ETR, and from the known stochiometries of electron use in photosynthesis and photorespiration, as follows (Epron et al., 1995; Valentini et al., 1995):

ETRA = 1/3(ETR + 8(AN + RL)); ETRP = 2/3(ETR – 4(AN + RL))

(Eqn 4)

The ratio ETRA to ETRP is related to the Cc/O ratio in the chloroplast (where O represents the oxygen molar fraction at the oxygenation site) through the Rubisco specificity factor τ, as follows (Laing et al., 1974):

τ = (ETRA/ETRP)/(Cc/O)

(Eqn 5)

Using the values of τ previously determined in vitro for each species (Galmés et al., 2005a), and assuming O to be equal to the molar fraction in the air, the above equation was solved for Cc. The mesophyll conductance to CO2 was then calculated as:

g i = AN/(Ci – Cc)

(Eqn 6)

In principle, combined gas exchange and chlorophyll fluorescence should be performed simultaneously and over the same leaf area (Warren, 2006). We could not do this during the experiments, as the chlorophyll fluorescence head of the Li-6400 was not available at that time. However, the values may be comparable as they were taken at saturating light, one immediately after the other. We have previously shown (Flexas et al., 1998) that light-saturated values of ETR depend little on possible slight variations of ambient factors affecting gs, such as vapour-pressure deficit or leaf temperature, and even on variations of gs itself, unless the changes are strong. As the measurements were made one immediately after the other, and inside a growth chamber with controlled environmental conditions, it is unlikely that environmental conditions between the two measurements had changed enough to induce variations in ETR. Later, and in several species not included in the present study, we have measured ETR using both the PAM-2000 and the Li-6400 equipped with the 6400-40 leaf-chamber fluorometer, finding no significant differences between them whenever light was saturating (data not shown).

Another key point concerning the validity of the estimations of gi is the accuracy of the estimated values of τ and RL. We are quite confident of the values used for τ, as these were determined in vitro for each species (Galmés et al., 2005a). However, many uncertainties have been highlighted regarding mitochondrial respiration in the light, regardless of the method used for its estimation (Harley et al., 1992; Warren, 2006). Nevertheless, Harley et al. (1992) showed that misleading RL estimations effects on gi are of importance only when gi is high. In this sense, we have selected the species with the highest gi, L. maritima, to check the importance of possible RL deviations for gi. To cope with the overall range of treatment-based variability, the analysis has been made considering two single measurements, one corresponding to a well watered plant and the other to a severely stressed plant. Under- and overestimations of RL by 50 and 150% were assessed (Table 2). As shown in Table 2, a 50% change of RL suggests a change in gi of only up to 8.3% in well watered plants, and even less in stressed plants. Therefore important biases on RL would not lead to critical errors in gi estimations in the ranges obtained in the present study.

Table 2. Assessment of the influences of the rate of nonphotorespiratory CO2 evolution in the light (RL) deviations on the mesophyll conductance estimations (gi) for Lavatera maritima
Parameter A N (µmol CO2 m−2 s−1) ETR (µmol e m−2 s−1) R L measured (µmol m−2 s−1)
L. maritima well watered 33.3 270 1.3
R L measured 50% RL 150% RL
g i (mol m−2 s−1) 0.488 0.458 0.528
Percentage change with respect to RL measured 6.9 8.3
L. maritima severe water stress 4.2 131 1.3
R L measured 50% RL 150% RL
g i (mol m−2 s−1) 0.032 0.031 0.033
Percentage change with respect to ‘RL measured’ 2.5 2.8
  • The analysis considers two single measurements corresponding to a well watered plant and a severely stressed plant. Net photosynthetic rates (AN) and electron transport rates (ETR) are also shown for each measurement.

Finally, it is worth mentioning that the method of Epron et al. (1995) used here, and the variable chlorophyll fluorescence method of Harley et al. (1992), resulted in almost identical gi values (data not shown), but the former was preferred because values of Rubisco specificity factor were obtained directly for each species by Galmés et al. (2005a) (as in Epron et al., 1995), and not derived from CO2 photocompensation point estimations (as in Harley et al., 1992).

Quantitative limitation analysis

At ambient CO2 concentration, light-saturated photosynthesis is generally limited by substrate availability, which was verified by ANCi curves in the present data for each species and treatment (not shown). Under CO2-limited conditions, photosynthesis can be expressed as (Farquhar et al., 1980):

A N = ((Vc,maxCc)/(Cc + Kc(1 + O/Ko)))(1 – (Γ*/Cc)) – RL

(Eqn 7)

where Vc,max is the maximum rate of carboxylation of Rubisco, Kc and Ko are the Michaelis–Menten constants for CO2 and O2, respectively, and Γ* is the CO2 compensation point in the absence of mitochondrial respiration. Estimations of Vc,max were derived from ANCc curves. The treatment average of Γ* for the species was obtained, according to Brooks & Farquhar (1985):

Γ* = 0.5O

(Eqn 8)

from specific τ-values for each species (Galmés et al., 2005a). Kc, Ko and their temperature dependencies were taken from Bernacchi et al. (2002). RL was calculated for the ANCi curve from the same treatment, as given by Grassi & Magnani (2005).

To compare relative limitations on assimilation caused by water stress, photosynthetic limitations were partitioned into their functional components following the approach proposed by Grassi & Magnani (2005). This approach, which requires the measurement of AN, gs, gi and Vc,max, makes it possible to partition photosynthesis limitations into components related to stomatal conductance (SL), mesophyll conductance (MCL) and leaf biochemical characteristics (BL), assuming that a reference maximum assimilation rate can be defined as a standard. The maximum assimilation rate, concomitantly with gs and Vc,max, was reached under well watered conditions, therefore the control treatment was used as a reference.

Calculations of gi (and therefore Vc,max calculations) may be impaired if heterogeneous stomatal closure affects Ci calculations significantly (Laisk, 1983; Beyschlag et al., 1992). This may impair the application of limitation analysis. However, the effect of heterogeneous stomatal closure is negligible for gs values above 0.03 mol H2O m−2 s−1 (Flexas et al., 2002; Grassi & Magnani, 2005). In the present study, values lower than 0.03 mol H2O m−2 s−1 were obtained only under severe water stress, and in some of the species analysed (see Results and Discussion). Even in these cases, gi estimations were considered a good approximation of actual values because: (i) Cc calculations are unaffected by Ci in the model of Epron et al. (1995); and (ii) at low values of gi, the results are much less affected by errors in Ci. For instance, under severe water stress treatment, with a gs of 0.017 mol H2O m−2 s−1, L. magallufianum showed an AN of 1.6 µmol CO2 m−2 s−1, an ETR of 148 µmol e m−2 s−1, and a Ci of 222 µmol mol−1 (Table 3). Patchy stomatal closure usually results in some overestimation of Ci (Terashima, 1992). Even in the case of 50% overestimation of the measured Ci, the differences between gi were no greater than 0.015 mol H2O m−2 s−1, very small compared with control values (approx. 0.120 mol CO2 m−2 s−1), which may produce only a 15% difference in the calculated MCL, SL and BL (Table 3).

Table 3. Assessment of the influences of substomatal CO2 concentration (Ci) estimations on the mesophyll conductance (gi) and photosynthetic limitations for Limonium magallufianum
Parameter C i measured C i 50%
g i (mol CO2 m−2 s−1) 0.010 0.025
S L 28 37
MC L 33 18
B L 28 33
  • S L, stomatal limitation; MCL, mesophyll limitation; BL, biochemical limitation.
  • A possible overestimation of Ci by 50% because of heterogeneous stomatal closure was considered, to analyse how much it would affect limitation calculations.
  • Values for the main photosynthetic parameters were as follows: net photosynthetic rate (AN), 1.6 µmol CO2 m−2 s−1; stomatal conductance (gs), 0.077 mol H2O m−2 s−1; electron transport rate (ETR), 148 µmol e m−2 s−1; substomatal CO2 concentration (Ci), 222 µmol mol−1 air.

Statistical analysis

Regression coefficients between gs and AN, ETR, gi and Vc,max were calculated with the 8.0 SigmaPlot software package (SPSS, Chicago, IL, USA). Differences between means were revealed by Duncan analyses (P < 0.05) performed with the spss 12.0 software package (SPSS).


Stomatal conductance and photosynthesis responses to water stress and recovery

The response of leaf water potential (Ψ) and relative water content (RWC) to water stress and recovery during this experiment has been reported previously (Galmés et al., 2006). In most species, both Ψ and RWC decreased progressively but slightly from control to moderate water stress, followed by a larger decrease at severe water stress. Three of the species (D. ibicensis and the two Limonium spp.) showed almost isohydric behaviour (very small, usually nonsignificant changes in Ψ throughout the experiment), while the other seven species showed a marked anisohydric behaviour (progressive decreases in Ψ as water stress intensified) (Table 4). The day after rewatering, the recovery of leaf water status was almost complete in all species except C. albidus and P. lentiscus, which showed only approx. 50% recovery.

Table 4. Maximum (under control conditions) and minimum (under severe water stress conditions) predawn leaf water potential (ΨPD) for the 10 selected species (data from Galmés et al., 2007)
Species Maximum ΨPD
Minimum ΨPD
Diplotaxis ibicensis –0.43 ± –0.03 –1.00 ± –0.11
Beta maritima ssp. marcosii –0.33 ± –0.01 –3.34 ± –0.10
B. maritima ssp. maritima –0.333 ± –0.01 –3.738 ± –0.09
Lavatera maritima –0.41 ± –0.05 –3.54 ± –0.25
Phlomis italica –0.33 ± –0.01 –5.00 ± –0.01
Cistus albidus –0.513 ± –0.04 –4.117 ± –0.56
Hypericum balearicum –0.38 ± –0.01 –2.97 ± –0.16
Pistacia lentiscus –0.300 ± –0.03 –4.550 ± –0.45
Limonium magallufianum –0.53 ± –0.03 –1.38 ± –0.14
Limonium gibertii –0.550 ± –0.03 –1.050 ± –0.09

Despite the observed interspecific differences in water potential and relative water content, all 10 species showed a gradual decline in net photosynthesis (AN) as water stress intensified, starting at mild water stress, except for the two Beta spp. (Fig. 1). Vc,max followed a different pattern, maintaining values similar to those in irrigated plants under mild-to-moderate water stress, depending on the species, and declining thereafter (Fig. 1). Both stomatal (gs) and mesophyll (gi) conductances to CO2 declined progressively as water stress intensified (Fig. 2). Remarkably, under irrigation gi was equal to or smaller than gs for all the species analysed, although the differences became smaller as water stress intensified. By 24 h after rewatering all parameters showed some recovery, although its extent largely depended on the species, from almost null (e.g. P. lentiscus) to almost complete (e.g. L. maritima).

Details are in the caption following the image

Net photosynthetic rate (AN, •) and maximum velocity of carboxylation (Vc,max, ○) under different irrigation treatments: control (CO), mild water stress (MiWS), moderate water stress (MoWS), severe water stress (SeWS) and rewatering (RW). Values are means ± SE of four to five replicates per species and treatment.

Details are in the caption following the image

Stomatal conductance (gs, •) and mesophyll conductance (gi, ○) under different irrigation treatments: control (CO), mild water stress (MiWS), moderate water stress (MoWS), severe water stress (SeWS) and rewatering (RW). Values are means ± SE of four to five replicates per species and treatment.

To see whether these data fitted the photosynthetic response pattern usually described for C3 plants (Flexas et al., 2002, 2004), the above parameters, as well as the ETR, were plotted against gs pooling all species together (Fig. 3). For the entire range of gs, a decline in gs resulted in a proportional decline in AN, and a strong relationship was found between both variables (Fig. 3a). The ETR plot presented larger scattering because of the large variability in maximum ETR values among species (Fig. 3b). The mesophyll conductance to CO2 (gi) was related linearly to gs when pooling all species together, although B. maritima ssp. marcosii appeared to follow a somewhat curvilinear pattern (Fig. 3c). Regarding Vc,max (Fig. 3d), the pattern resembled that of ETR, except that interspecific differences in the maximum values were not so large. None of the species analysed presented a decline in Vc,max until gs dropped below approx. 0.10–0.15 mol H2O m−2 s−1, and in both Limonium spp. even lower gs values were required before Vc,max declined.

Details are in the caption following the image

Relationship between stomatal conductance (gs) and (a) net photosynthetic rate (AN); (b) electron transport rate (ETR); (c) mesophyll conductance (gi); (d) maximum rate of carboxylation (Vc,max). Values from rewatering treatment are not included. Regression coefficients and significance of each relationship are shown. Values are means ± SE of four to five replicates per species and treatment. Symbols and species: •, Diplotaxis ibicensis; ○, Beta maritima ssp. marcosii; □, B. maritima ssp. maritima; ▪, Limonium magallufianum; ▵, Limonium gibertii; ▴, Phlomis italica; ▾, Lavatera maritima; ▿, Cistus albidus; ◆, Hypericum balearicum; ◊, Pistacia lentiscus.

Photosynthetic limitations during water-stress imposition

The responses described above relate qualitatively water stress-induced variations in some photosynthetic parameters to water stress-induced reductions in AN. A quantitative relationship can be obtained through a limitation analysis (Jones, 1985; Grassi & Magnani, 2005). The results are shown in Table 5. At mild water stress (as well as at moderate water stress in L. maritima and the two Limonium spp.), the biochemical limitations (BL) were negligible, and the sum of stomatal (SL) and mesophyll conductance (MCL) limitations accounted for the entire photosynthetic limitation. In some species, such as L. maritima and the two Limonium spp., SL was much more important than MCL at mild to moderate water stress. In other species, such as C. albidus, H. balearicum and P. lentiscus (the most sclerophyll species), MCL was much larger than SL. In the remaining species, both limitations were of similar magnitude. At moderate-to-severe water stress, SL was still the most important limitation on photosynthesis only in L. maritima. In most species, MCL was the most important limitation at severe water stress, although in some (D. ibicensis, B. maritima ssp. marcosii and ssp. maritima, L. magallufianum), BL was of similar magnitude. As shown previously (Grassi & Magnani, 2005), the evolution of these limitations with water stress was closely correlated with gs (Fig. 4a), and BL became detectable only when gs dropped below 0.05–010 mol H2O m−2 s−1, a situation where MCL was the most important limitation on photosynthesis.

Table 5. Limitations of AN, expressed as percentage, under different irrigation treatments: mild water stress (MiWS), moderate water stress (MoWS) and severe water stress (SeWS)
Treatment Limitation
Total (TL) Stomatal (SL) Mesophyll conductance (MCL) Biochemical (BL)
Diplotaxis ibicensis
MiWS 17 7 9 1
MoWS 61 29 21 11
SeWS 78 47 16 15
Beta maritima ssp. marcosii
MiWS 0 0 0 0
MoWS 40 19 10 11
SeWS 99 18 42 39
B. maritima ssp. maritima
MiWS 0 0 0 0
MoWS 32 13 13 6
SeWS 98 21 34 43
Lavatera maritima
MiWS 24 18 6 0
MoWS 57 47 9 1
SeWS 87 70 7 10
Phlomis italica
MiWS 31 7 24 0
MoWS 68 29 28 11
SeWS 96 22 52 22
Cistus albidus
MiWS 15 7 8 0
MoWS 49 15 22 12
SeWS 95 15 52 28
Hypericum balearicum
MiWS 24 4 20 0
MoWS 45 10 24 11
SeWS 89 33 38 18
Pistacia lentiscus
MiWS 17 5 12 0
MoWS 41 10 21 10
SeWS 94 12 56 26
Limonium magallufianum
MiWS 30 20 9 1
MoWS 60 40 19 1
SeWS 89 28 33 28
Limonium gibertii
MiWS 12 6 4 2
MoWS 47 34 12 1
SeWS 81 33 30 18
Details are in the caption following the image

Relationship between limitations of net photosynthetic rate (AN) and stomatal conductance (gs) considering all 10 species studied. Values obtained from (a) mild, moderate and severe water-stress treatments; (b) rewatering treatment. The regression coefficient and significance of the relationships between total limitations and stomatal conductance are shown. BL, biochemical limitation; MCL, mesophyll limitation; SL, stomatal limitation; TL, total limitations.

Limitations on photosynthesis recovery after a water-stress period

In the present study, we analysed the recovery of photosynthesis 24 h after rewatering severely water-stressed plants, in which gs and AN were strongly depressed. The extent of recovery of photosynthesis was species-dependent, ranging from < 10% of control values in P. lentiscus to almost 70% in L. maritima (Table 6). In general, and with the exception of L. maritima, herbs showed the largest recovery (49–64%), semideciduous an intermediate recovery (21–42%), and evergreens the lowest recovery (10–29%).

Table 6. Limitations of AN (expressed as percentage) 24 h after refilling water in pots at saturation point
Species Limitation
Total (TL) Stomatal (SL) Mesophyll conductance (MCL) Biochemical (BL)
Diplotaxis ibicensis 47 13 26 8
Beta maritima ssp. marcosii 36 6 26 4
B. maritima ssp. maritima 52 8 36 7
Lavatera maritima 31 22 8 0
Phlomis italica 58 19 26 13
Cistus albidus 78 15 46 18
Hypericum balearicum 71 31 31 9
Pistacia lentiscus 91 13 45 32
Limonium magallufianum 70 17 49 5
Limonium gibertii 78 24 52 2

Regarding the mechanisms limiting photosynthetic recovery after severe water stress, the different extents in recovery of AN were accompanied by different extents in recovery of either gs, gi or Vc,max (1, 2). However, the limitation analysis revealed that MCL was, by far, the strongest limitation on photosynthesis recovery in all species analysed, with the exception of L. maritima, the species showing the largest recovery. The recovery of biochemical limitations after severe water stress was generally large. Only in P. lentiscus BL still accounted for 32%, but even so, it contributed only to one-third of the total limitation. Remarkably, the relationship between photosynthetic limitations and gs during recovery was not the same as during water-stress imposition (Fig. 4b). While there was still a highly significant relationship between total limitation and gs (AN and gs maintained their coregulation), MCL was the most important limitation at any given gs, while SL and BL were of similar magnitude throughout the entire range. That limited recovery of gi was the most important limitation on photosynthetic recovery in these species was further highlighted by comparing the relationships between total photosynthetic limitation and partial limitations after rewatering, pooling all species together. The relationship between TL and SL was nonsignificant (Fig. 5a), and that between TL and BL was only marginally significant (Fig. 5b). However, the relationship between TL and MCL was highly significant (Fig. 5c).

Details are in the caption following the image

Relationship between total limitation of photosynthesis (TL) 24 h after rewatering plants and (a) stomatal limitation (SL); (b) biochemical limitation (BL); (c) mesophyll limitation (MCL). Regression coefficients and the significance of each relationship are shown. Symbols and species as in Fig. 3.


The present results show that the 10 Mediterranean plants analysed follow the pattern of photosynthesis response to progressive water stress usually described in C3 plants (Flexas et al., 2004). Although small differences have been observed between species, they all follow roughly this general pattern, consisting of an early phase of water stress-induced AN decline associated with decreases in gs and gi, followed by a second phase in which Vc,max and ETR decrease to some extent (Flexas et al., 2004). This pattern therefore seems very robust and independent of any possible particular adaptation to Mediterranean conditions. Moreover, it is independent of growth forms and leaf types, as well as of water relations, as it was followed by both isohydric and anisohydric species (Galmés et al., 2006).

Therefore, in all the species, regardless of growth form and leaf type, there was a shift from limitations mostly caused by to CO2 diffusion (SL plus MCL) at mild-to-moderate water stress, to a combination of diffusion and biochemical limitations (BL) at severe water stress, as suggested by previous studies (Tenhunen et al., 1985; Harley et al., 1986; Harley et al., 1987a, 1987b; Gulías et al., 2002; Lawlor & Cornic, 2002; Flexas et al., 2004; Peña-Rojas et al., 2004). In contrast to these studies, the present data highlight the importance of gi as a limiting factor for photosynthesis in Mediterranean plants, as suggested by Niinemets et al. (2005), particularly under water-stress conditions (Roupsard et al., 1996). Limitation by gi has been suggested as a possible cause of the observed discrepancies between measured water-use efficiency and that estimated with current gas-exchange models in Mediterranean ecosystems (Reichstein et al., 2002). In all the plants studied here, gi was ≤ gs. A gi smaller than gs has been described in woody plants (Miyazawa & Terashima, 2001; Hanba et al., 2002; Centritto et al., 2003; De Lucia et al., 2003; Peña-Rojas et al., 2004; Warren et al., 2004; Warren & Adams, 2006) – although not in all cases (Epron et al., 1995) – and it is rarely observed in herbaceous plants (Loreto et al., 1992; De Lucia et al., 2003; Warren et al., 2006). This has been interpreted in terms of the leaf mesophyll anatomy effects on gi (Syvertsen et al., 1995; Hanba et al., 1999). However, the present data suggest that gi may be more limiting for photosynthesis than gs in different Mediterranean plants, regardless of their growth form and leaf anatomy. This is consistent with a predominant role of metabolic rather than structural determinants of gi, such as aquaporins (Flexas et al., 2006b). On the other hand, whether the relationship between gi and gs is linear or curvilinear is an unresolved question (Flexas et al., 2004; Warren et al., 2006), which is important for understanding gi effects on photosynthetic nitrogen and water-use efficiency under water or salinity stress (Warren et al., 2006). The present results, along with those of Centritto et al. (2003), suggest that linear relationships may be more common, but a curvilinear relationship may be found in some species, such as B. maritima ssp. marcosii or Vitis vinifera (Flexas et al., 2002). The implications of these differences remain to be established.

In contrast to photosynthetic limitations during water-stress development, which have been studied intensively over the past 30 yr, photosynthetic limitations during recovery after a water-stress period have received much less attention. Usually photosynthesis recovery after a mild water stress (whenever gs is maintained above 0.15 mol H2O m−2 s−1) is rapid (1 d after rewatering) and almost complete (Flexas et al., 2006a). In contrast, after severe water stress the recovery of photosynthesis is progressive and slow (lasting from days to weeks) and sometimes incomplete (De Souza et al., 2004; Miyashita et al., 2005; Flexas et al., 2006a). In the latter case, it would be interesting to know which are the factors limiting recovery in the short term. However, with the exception of early studies by Kirschbaum (1987, 1988), which did not take into account mesophyll limitations, a detailed photosynthetic limitations analysis, including SL, MCL and BL, has not yet been performed.

The present results show that, with the exception of L. maritima, herbs showed the largest recovery, semideciduous species an intermediate recovery, and evergreens the least recovery. This may reflect different adaptations to water-stress periods under Mediterranean conditions. For instance, herbs may experience short water-stress periods during the favourable season, and therefore a capacity for rapid recovery may be important to ensure their carbon-balance requirements before ending their life cycle in late spring. In contrast, evergreens suffer less from short, dry periods during the favourable season because of their large root system (Rambal, 1984; Canadell et al., 1996), but may have to endure a long water-stress period in summer, during which they may rely on more permanent physiological changes precluding rapid recovery (Mittler et al., 2001).

The limitation analysis performed for recovery data revealed that, contrary to what is usually assumed (Flexas et al., 2004), the recovery of biochemical limitations after severe water stress was generally large. This result contrasts with recent results of Ennahli & Earl (2005), who showed in cotton that, after severe water stress, recovery 24 h after rewatering was mostly caused by biochemical limitations, while stomatal and mesophyll limitations were almost totally absent. In the 10 species studied here, the main photosynthetic limitation during photosynthesis recovery after a severe stress appears to be mesophyll conductance. To the best of our knowledge, this is the first report showing that limited recovery of gi is the most important factor limiting photosynthesis recovery after a severe water stress. This finding highlights the role of gi in controlling photosynthesis, and indicates the need for a better understanding of the physiological and molecular mechanisms underlying the regulation of gi.


The authors are very grateful to Dr M. Ribas-Carbó for help during experiments. Drs Hans Lambers, Martin A.J. Parry, Fernando Valladares, Javier Gulías and Alfred J. Keys are acknowledged for their helpful comments on a previous version of the manuscript. This work was partly funded by Projects REN2001-3506-CO2-O2 and BFU2005-03102/BFI (Plan Nacional, Spain).