Mesophyll conductance constrains photosynthesis in three common sclerophyllous species in Central Chile

Abstract

Background

Quillaja saponaria Mol., Cryptocarya alba Mol. Looser, and Lithraea caustica Molina Hook et Arn., are common sclerophyllous species in Mediterranean Central Chile. Mesophyll conductance, g_m, may strongly limit photosynthesis in these semiarid environments.

Results

Simultaneous measurements of gas exchange and chlorophyll fluorescence were carried out in 45 nursery plants from these species to determine diffusional and biochemical limitations to photosynthesis. Values of stomatal conductance, g_s, were greater than those of mesophyll conductance, g_m, while their ratio (g_m/g_s) was not influenced by species being on average 0.47. Relative limitations posed by mesophyll conductance to photosynthesis, L_m, (0.40 ± 0.02) were high compared to those imposed by stomata, L_s (0.07 ± 0.01). The average CO₂ concentration in the intercellular air spaces (C_i) was 32 μmol mol⁻¹ lower than in the atmosphere (C_a), while the average CO₂ concentration in the chloroplasts (C_c) was 131 μmol mol⁻¹ lower than C_i independent of species. Maximal rates of Rubisco carboxylation, V_cmax, and maximal electron transport rates driving regeneration of RuBP, J_max, ranged from 13 to 66 μmol CO₂ m⁻² s⁻¹ and from 33 to 148 μmol electrons m⁻² s⁻¹, respectively, and compare well to averages for C₃ plants.

Conclusions

Photosynthetic performance was in the series: Q. saponaria > C. alba ≥ L. caustica, which can be attributed first to mesophyll conductance limitations, probably mediated by leaf anatomical traits and then to species specific foliage N partitioning strategies.

Background

The biochemical model of leaf photosynthesis originally proposed by Farquhar et al. ([1980]) and later improved (von Caemmerer and Farquhar, [1981]; Sharkey, [1985]; Harley and Sharkey, [1991]) is widely used in ecophysiological research for describing CO₂ exchange processes and to scale carbon exchange from leaves to canopies (Baldocchi and Harley [1995]; McMurtrie et al. [1992]; Whitehead et al. [2004]). In this model, photosynthesis is considered to be limited by the maximal rate of ribulose-1,5-bis phosphate (RuBP) carboxylase-oxygenase (Rubisco) carboxylation, V_cmax, and by the maximal electron transport rate driving regeneration of RuBP, J_max, and by triose phosphate utilization (TPU) (Farquhar et al. [1980]; von Caemmerer [2000]; von Caemmerer and Farquhar [1981]). These parameters are fitted to the response of photosynthesis, A, to the CO₂ concentration in the intercellular air spaces, C_i, known as A/C_i curves. Values of C_i, are estimated from stomatal conductance to CO₂ transfer, g_s, and the ambient CO₂ concentration external to the leaf, C_a. This model has been used to explain the mechanisms of photosynthetic acclimation to elevated atmospheric CO₂ (Griffin et al. [2000]; Hogan et al. [1996]; Kellomaki and Wang [1996]; Murray et al. [2000]; Turnbull et al. [1998]), the influence of global warming on plant carbon budgets (Turnbull et al. [2002]), and for identifying factors limiting photosynthesis under water stressed conditions in Mediterranean plants (Grassi and Magnani [2005]; Gulias et al. [2002]; Niinemets et al. [2009b]), among others.

In most studies, the mesophyll conductance of CO₂ from the intercellular air spaces to the sites of Rubisco carboxylation in the chloroplasts, g_m, has been considered to be sufficiently large to be negligible (Farquhar et al. [1980]). However, there is a growing awareness that the significance of g_m in limiting photosynthesis can be similar to the limitation imposed by stomata (Flexas et al. [2008]; Harley et al. [1992]; Loreto et al. [1992]; von Caemmerer [2000]; Warren and Adams [2006]; Warren et al. [2003]). As a result, C_i is greater than the CO₂ concentration in the chloroplasts, C_c, and values of V_cmax and J_max are underestimated when fitted to estimates of C_i rather than C_c.

Arid and semiarid lands account for 41% of Chile’s continental territory, covering about 31 million ha (Benites et al. [1994]). Within this area, semiarid sclerophyllous (i.e. a woody plant with small coriaceous evergreen leaves dominant of the Mediterranean region) shrublands and forests extends from 32-36° S latitude (~345,000 ha) in Central Chile (Armesto et al. [2007]; CONAF [1999]); exhibiting high levels of endemism, and are therefore considered a priority for biodiversity conservation (Arroyo et al. [2004]). These ecosystems are subjected to high radiative and water stresses that limit their development and reproduction (Cabrera [2002]); and, additionally, have a long history of degradation by human activity, which may represent an important adaptive pressure (Galmes et al. [2007]). Therefore, understanding how environmental limitations are imposed on the carbon budget is relevant to accurately estimate carbon uptake and water use by sclerophyllous species.

There is extensive work on comparisons of g_m between species and plant functional groups (Loreto et al. [1992]; De Lucia et al. [2003]; Hassiotou et al. [2009]; Niinemets et al. [2009b]). However, little is known about the extent to which g_m regulates the rate of photosynthesis in sclerophyllous species of central Chile. Consequently we chose to work with three native sclerophyllous species: Quillaja saponaria Mol., Cryptocarya alba Mol. Looser and Lithraea caustica Molina Hook et Arn., to determine stomatal, mesophyll and biochemical limitations to photosynthesis. Specifically, we assessed whether mesophyll conductance to CO₂ induces similar constrains to photosynthesis across different sclerophyllous species.

Methods

Plant material

Plant material consisted of 15 randomly selected plants from each of the following species: Q. saponaria, C. alba and L. caustica, from a nursery of the Faculty of Forestry and Nature Conservation at University of Chile, Santiago, Chile. Seeds were sown in the winter 2008, and transferred during spring 2008 and spring 2009 to 12 × 15 cm (0.25 L) and 20 × 30 cm (2 L) black polyethylene bags filled with an even mixture of composted plant residues, soil and sand. No fertilizers were applied. Plants developed under 46% shading (in order to mimic the light environment experienced by saplings in their natural habitat) and received weekly irrigation during winter, and daily irrigation to field capacity in other seasons. At the time of measurements, we selected fifteen homogenous plants of Q. saponaria, C. alba and L. caustica which exhibited an average (±1 SD) plant height of 82.9 ± 14.4 cm, 63.4 ± 18.9 cm, and 47.6 ± 15.3 cm; and a collar diameter of 7.5 ± 0.8 mm, 9.1 ± 1.4 mm, and 11.0 ± 2.6 mm, respectively. All gas exchange and chlorophyll fluorescence measurements were taken in a laboratory without thermal regulation, with open windows to allow circulation of fresh air. Measurements took 45 days; with only one plant measured per day, spread in the summer period from December 20, 2010 to March 15, 2011; six months after plants were transferred from small to large containers. Gas exchange measurements for each individual plant typically took all day between 10 am and 6 pm; and followed the exact same sequence of measurements for all plants (see below). We alternated species each day in order to avoid differences brought about by day to day meteorological changes.

Gas exchange measurements

Simultaneous measurements of gas exchange and chlorophyll a fluorescence were carried out on the 45 selected plants, using a portable photosynthesis system (CIRAS-2, PP Systems, MA, USA) equipped with an integrated chlorophyll fluorescence and gas exchange chamber (PLC6-U Auto Leaf Cuvette, PP System, MA, US). Plants were shifted from the nursery to the laboratory the day before measurements were undertaken.

For each plant, a fully-developed leaf in the upper third of the plant was chosen and placed inside a circular 2.5 cm² (18 mm in diameter) cuvette. Temperature in the cuvette (block) was maintained at 25ºC while leaf-to-air vapour pressure deficit (VPD) was maintained below 1 kPa. Each leaf was left to equilibrate inside the cuvette for 10 min at about 368 ± 3 μmol mol⁻¹ CO₂ concentration and saturating irradiance (2,000 μmol photons m⁻² s⁻¹), before measuring the response of net assimilation (A) to intercellular CO₂ concentration (C_i). External CO₂ concentration (C_a) was supplied with a CO₂ mixer across the sequence 25, 50, 75, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1200 and 1500 μmol mol⁻¹, with saturating irradiance, Q (400–700 nm), maintained at 2,000 μmol m⁻² s⁻¹. Measurements were recorded after values of A, C_i and g_s became stable but with a minimum waiting time of 3 min at each step within the sequence. At each value of C_i, measurements of fluorescence for the light-adapted leaf were made simultaneously to the gas exchange measurements. Values of F and F_m’ (the steady and maximal fluorescence respectively), were used to calculate photochemical efficiency of photosystem II, Φ_PSII. The response of net assimilation to irradiance (A/Q curves) was measured immediately after each A/C_i curve ended, across the Q sequence: 0, 50, 100, 150, 200, 300, 400, 500, 800, 1000, 1200, 1400, 1600, 1800 and 2000 μmol m⁻² s⁻¹, with ambient CO₂ concentration maintained at 368 ± 3 μmol mol⁻¹ using a CO₂ mixer. The A/C_i and A/Q curves were measured on the same foliage sample.

Models fitted

The biochemical model of leaf photosynthesis by Farquhar et al. ([1980]) describes the rate of photosynthesis (A) as:

$$A=min\left\{A_{\mathrm{c}},A_{\mathrm{q}}\right\}-R_{\mathrm{d}}$$

(1)

where A_c and A_q are the photosynthetic rates limited by Rubisco carboxylation and by electron transport rate respectively, and min {} indicates the minimum of these two rates. R_d is the rate of daytime respiration resulting from processes other than photorespiration. The photosynthetic rate limited by Rubisco carboxylation (A_c) is given by:

$$A_{\mathrm{c}}=V_{\mathrm{c}max}\frac{C_{\mathrm{i}}-{\Gamma }^{*}}{C_{\mathrm{i}}+K_{\mathrm{c}}\left(1+O_{\mathrm{i}}/K_{\mathrm{o}}\right)}$$

(2)

where V_cmax is the maximum rate of Rubisco carboxylation under saturating RuBP and CO₂, C_i and O_i are the intercellular CO₂ and O₂ concentrations, Γ^* is the CO₂ compensation concentration in the absence of day respiration, and K_c and K_o are the Michaelis constants for CO₂ and O₂, respectively.

The photosynthetic rate limited by RuBP regeneration driven by electron transport (A_q) is given by:

$$A_{\mathrm{q}}=\frac{J\left(C_{\mathrm{i}}-{\Gamma }^{*}\right)}{4\left(C_{\mathrm{i}}+2{\Gamma }^{*}\right)}$$

(3)

where J is the rate of electron transport at a given irradiance Q.

Prioul and Chartier ([1977]) described the response of A to Q by a non-rectangular hyperbola as:

$$A=\frac{\mathit{\alpha Q}+A_{\mathrm{sat}}-\sqrt{{\left(\mathit{\alpha Q}+{\mathrm{A}}_{\mathrm{sat}}\right)}^2-4\mathit{\alpha \theta Q}{A}_{\mathrm{sat}}}}{2\theta }-{\mathrm{R}}_{\mathrm{dark}}$$

(4)

where θ is the convexity of the non-rectangular hyperbola, α is the initial slope of the A/Q curve (often referred as ‘apparent maximum quantum efficiency’), A_sat is the light-saturated photosynthetic capacity and R_dark the respiration rate at zero irradiance. Individual A/Q response curves were fitted using Eq. 4 in order to estimate α, A_sat and R_dark. The light saturation point, Q_sat, was calculated as the irradiance for which 90% of A_sat was achieved.

Values of the rate of mitochondrial respiration in the light, R_d^*, and the chloroplastic CO₂ compensation point, Γ^*, were estimated for each sample using the Laisk method (von Caemmerer [2000]). Briefly, the A/C_i response was measured at three levels of low irradiance (Q = 50, 200 and 400 μmol m⁻² s⁻¹) for six increasing values of C_a from 25 to 400 μmol CO₂ mol⁻¹. Linear relationships between A and C_i were fitted and the point of intersection of the three lines was taken as R_d^* (y-axis) and Γ^* (x-axis) (Figure 1).

Figure 1

The response of net photosynthesis ( A ) to intercellular CO ₂ concentration ( C _i ) at three different irradiances ( Q = 400, 200 and 50 μmol m⁻² s⁻¹, 400–700 nm) for a representative foliage sample. Linear relationships between A and C_i were fitted for each Q level and their intersections averaged to yield a point which when projected to the A axis was taken as the rate of day respiration (R_d^*) and when projected to the C_i axis was taken as the intercellular CO₂ compensation concentration in the absence of day respiration (C_i^*). The mitochondrial CO₂ compensation concentration (Γ^*) was calculated as Γ^* = C_i^* + R_d^* / g_m (von Caemmerer [2000]), where g_m is mesophyll conductance to CO₂ transfer. Method originally proposed by Laisk, A. (von Caemmerer [2000]) to determine C_i^* and R_d. This figure is equivalent to the one presented by De Lucia et al. ([2003]) but with data drawn from this study.

The photosynthesis model described by Farquhar et al. ([1980]) (Eq. 1) was fitted to the A/C_i and A/C_c curves by non-linear least squares regression (SigmaPlot, Version 12.1, SPSS, Chicago, IL). Values of V_cmax and R_d were estimated from the lower part of the A/C_i or A/C_c curve (C_i or C_c < 220 μmol mol⁻¹) (Eq. 1 and 2), and these values were then used to estimate J_max over the entire range of measured C_i or C_c (Eq. 1,2 and 3). Michaelis-Menten constants of Rubisco for CO₂ and O₂, K_c and K_o, used in the fitting (25 °C) were 404.9 μmol mol⁻¹ and 278.4 mmol mol⁻¹, respectively as described by Bernacchi et al. ([2001]).

Calculations of mesophyll conductance

Mesophyll conductance, g_m, was estimated using the “constant J” method (Harley et al. [1992]; Loreto et al. [1992]) where J is the rate of electron transport (Figure 2b). This method uses data in the RuBP-regeneration limited portion of the A/C_i curve, where rates of electron transport are constant. Within this region, further increases in photosynthesis with increasing C_i are due to suppression of photorespiration as the rate of carboxylation progressively substitutes the rate of oxygenation. Thus, photosynthesis is related to C_c and the relative CO₂/O₂ specificity of Rubisco, normally described by the chloroplastic CO₂ compensation point, Γ^*. The constant J method is sensitive to errors in both the rate of mitochondrial respiration in the light, R_d^*, and Γ^* and the approach assumes that both J and g_m are constant across the range of C_a concentrations used for the measurements (Harley et al. [1992]; Pons et al. [2009]). The relationship of g_m with Γ^*, R_d^* and intercellular CO₂ compensation point in the absence of day respiration, C_i^* is given by Γ^* = C_i^* + R_d^*/g_m (Peisker and Apel [2001]; von Caemmerer [2000]).

Figure 2

Graphic description of the constant J method to determine transfer conductance. (a) The rate of net photosynthesis (A; open circles) and photochemical efficiency of photosystem II (Φ_PSII; solid triangles) as a function of the intercellular CO₂ concentration (C_i) for a representative foliage sample. Solid lines represent a least-squares fit to the A/C_i and Φ_PSII/C_i response. Open squares are observed values used to estimate mesophyll conductance (g_m). These values are within the portion of the A/C_i response where Φ_PSII indicated that electron transport rate was constant. (b) The variance of estimated electron transport rates, J, for different values of g_m. Values of J were estimated for each of the four A values indicated as open squares in Figure 2a using the equation given by Harley et al. ([1992]). The g_m that minimized the variance of J estimates for this foliage sample was 0.054 mol m⁻² s⁻¹ bar⁻¹. This figure is equivalent to the one presented by Harley et al. ([1992]) and De Lucia et al. ([2003]) but with data drawn from this study.

The photochemical efficiency of photosystem II (Φ_PSII) was estimated from chlorophyll fluorescence measurements as (F_m’ – F) / F_m’, where F and F_m’ are the steady and maximal fluorescence in the light-adapted sample respectively (Schreiber et al. [1994]). The values of Φ_PSII are directly proportional to the rate of electron transport through photosystem II, and therefore can be used to determine the portion of the A/C_i curve where the rate of electron transport is constant (Genty et al. [1989]) (Figure 2a). Optimal values for g_m were resolved iteratively from three or more measurements of photosynthesis at high values of C_i that correspond with constant rates of electron transport (Singsaas et al. [2004]; Warren [2006]). This was done using the Generalized Reduced Gradient Nonlinear Solving Method for nonlinear optimization included in the Solver Tools of Microsoft Excel with measurements of A, C_i, Γ^*, R_d^* to resolve the value of g_m that best explained changes in photosynthesis with changes in C_i indicated by minimum variance in J (Harley et al. [1992]; Singsaas et al. [2004]; Warren [2006]) (Figure 2b).

Stomatal and mesophyll limitations to photosynthesis

A/C_i response and values for g_s and g_m, were used to partition stomatal, mesophyll and biochemical limitations to photosynthesis. Values of stomatal conductance to CO₂ transfer were calculated by g_s = A/(C_a-C_i) dividing values by 1.64 (Jones [1992]). Relative stomatal limitations were calculated using the method of Farquhar and Sharkey ([1982]) as L_s = 1 – A_a-gs / A_i-gs where A_a-gs and A_i-gs are the actual value of A and the value estimated when g_s is infinite, respectively. Similarly, following Bernacchi et al. ([2002]), relative limitation to photosynthesis imposed by g_m was calculated as: L_m = 1 – A_a-gm / A_i-gm where A_a-gm and A_i-gm are the actual value of A and the value estimated when g_m is infinite, respectively. The CO₂ concentration in the chloroplasts, C_c, was calculated from C_c = C_i – A/g_m.

Foliage surface area and nutrient concentrations

Following the completion of A/C_i and A/Q curves, the leaf was carefully removed from the cuvette, scanned to determine leaf area by optical methods, dried at 70°C to constant mass and then both measurements used to calculate the foliage area to mass ratio, M. Foliage samples were finely ground, Dumas combusted and N concentrations determined by thermo conductivity (LECO, TruSpec CN, US) at the Laboratory of Soil and Foliar Analysis at Universidad Católica de Chile, Santiago, Chile. Nitrogen foliage concentrations are expressed on a hemi-surface area basis (N_a) and photosynthetic nitrogen-use efficiency, E_N, defined as A_sat/N_a.

Data analysis

All statistical analyses were undertaken at the plant level using The R System (R Core Team [2013]). Variables were tested for normality and homogeneity of variance and transformations were made as necessary to meet the underlying statistical assumptions of the models used. All values are presented as means ± 1 standard error (n = 15) unless stated otherwise. A one way analysis of variance was used to compare diffusional and biochemical limitations to photosynthesis between sclerophyllous species. Tukey’s least significant difference test was used to distinguish among individual means where applicable with a confidence level of P < 0.05. Differences in slopes and intercepts in the linear relationships between photosynthetic parameters and foliar nitrogen among sclerophyllous species were tested for significance by analysis of covariance.

Results and discussion

Overall photosynthetic performance

The Farquhar et al. ([1980]) model was fitted to the A/C_i curves while the Prioul and Chartier ([1977]) model to the A/Q curves measured for each leaf sample (n = 45). There was excellent correspondence between modelled and observed data independent of the sclerophyllous species (Figure 3).

Figure 3

The response of net assimilation to either (a) photosynthetically active radiation ( A / Q curve) or (b) internal CO ₂ concentration (A / C _i curve) for Q. saponaria , C. alba and L. caustica. The Prioul and Chartier ([1977]) model was fitted to the A/Q curve while the Farquhar et al. ([1980]) model to the A/C_i curve.

Light response curves (Figure 3a) were very similar among species up to an irradiance level of about 500 μmol photons m⁻² s⁻¹, changing drastically thereafter with clearly distinct maximums of photosynthetic rates in the series: Q. saponaria > C. alba > L. caustica which reflects that the three species exhibited a similar initial pseudo-linear slope of the A/Q curve (the apparent maximum quantum efficiency: α) averaging 0.033 ± 0.007 mol CO₂ mol⁻¹ quanta. The rate of photosynthesis at saturating irradiance (2000 μmol m⁻² s⁻¹) and ambient CO₂ concentration (A_sat), on the other hand, was significantly greater in Q. saponaria (14.2 ± 0.8 μmol CO₂ m⁻² s⁻¹) compared to C. alba (9.5 ± 0.4 μmol CO₂ m⁻² s⁻¹) and L. caustica (7.8 ± 0.7 μmol m⁻² s⁻¹) (Figure 3a; Table 1). The light saturation point, Q_sat, which followed a similar pattern as A_sat, was consistently higher in Q. saponaria (1168 ± 57 μmol photons m⁻² s⁻¹) compared to C. alba (964 ± 56 μmol photons m⁻² s⁻¹) and L. caustica (987 ± 65 μmol photons m⁻² s⁻¹). The rate of respiration in the absence of light at ambient CO₂ concentration (R_dark) was 0.77 ± 0.10 μmol CO₂ m⁻² s⁻¹ independent of the species. The light compensation point, or the irradiance value at which the rate of photosynthesis equals zero, did not differ across species, being 38 ± 4 μmol photons m⁻² s⁻¹. The rate of transpiration at saturating irradiance (2000 μmol m⁻² s⁻¹) and ambient CO₂ concentration (E) was 2831 ± 309 μmol H₂O m⁻² s⁻¹ while the instantaneous water use efficiency (E/ A_sat) was 337 ± 41 mol H₂O mol⁻¹ CO₂, unaffected by species (Table 1).

Table 1 Comparison of leaf and photosynthetic parameters for three common sclerophyll species in Central Chile

The response of photosynthetic rates (A) to internal CO₂ concentration (C_i) (Figure 3b), changed drastically among species for the whole range of C_i values in the series: Q. saponaria > C. alba > L. caustica. Consequently the initial slope of the A/C_i curves (dA/dC_i; often referred to as ‘carboxylation efficiency’) was greater in Q. saponaria (0.070) compared with C. alba (0.047) and L. caustica (0.038). A similar trend was observed for the rate of photosynthesis near saturating C_i (800 μmol mol⁻¹), A_max, being 12.03, 17.01 and 23.81 μmol CO₂ m⁻² s⁻¹ for L. caustica, C. alba and Q. saponaria, respectively (Table 1).

Mesophyll conductance

Mesophyll conductance (g_m) was similar between L. caustica (0.060 mol CO₂ m⁻² s⁻¹ bar⁻¹) and C. alba (0.065 mol CO₂ m⁻² s⁻¹ bar⁻¹), but collectively significantly lower than Q. saponaria (0.097 mol CO₂ m⁻² s⁻¹ bar⁻¹). Although unsignificantly, g_s tended to be greater for Q. saponaria (0.250 mol CO₂ m⁻² s⁻¹) compared to C. alba (0.227 mol CO₂ m⁻² s⁻¹) and L. caustica (0.221 mol CO₂ m⁻² s⁻¹). The ratio g_m/g_s was on average 0.47 ± 0.08, independent of species (Table 1). Relative limitations posed by mesophyll conductance to photosynthesis, L_m, (0.40 ± 0.02) were high compared to those imposed by stomata, L_s (0.07 ± 0.01) and not significantly influenced by species. The average CO₂ concentration in the intercellular air spaces (C_i) was 31.7 μmol mol⁻¹ lower than in the atmosphere (C_a), while the average CO₂ concentration in the chloroplasts (C_c) was 130.6 μmol mol⁻¹ lower than in C_i , independent of the species (Table 1).

The g_m values found in this study were generally lower than those reported for other plant functional groups. Flexas et al. ([2008]) argues that g_m is associated with leaf forms and plant functional groups, rather than evolutive trends: herbaceous plants exhibit generally the largest g_m values (~0.4 mol CO₂ m⁻² s⁻¹ bar⁻¹), perennial herbs and woody deciduous angiosperms display intermediate values (~0.2 mol CO₂ m⁻² s⁻¹ bar⁻¹), while woody evergreen plants exhibit g_m values slightly above and below 0.1 mol CO₂ m⁻² s⁻¹ bar⁻¹ in angiosperms and gymnosperms, respectively. The mean value across all angiosperm sclerophyllous species studied was 0.073 mol CO₂ m⁻² s⁻¹ bar⁻¹, similar to those found by Niinemets et al. ([2009b]) for Australian sclerophyllous species (0.087 mol CO₂ m⁻² s⁻¹ bar⁻¹).

There seems to be a consistent pattern for sclerophyllous species to exhibit both low g_m and low g_m/g_s values (Gulias et al. [2002]; Niinemets et al. [2009b]; Tomas et al. [2013]; Warren [2004]). Consequently, differences between C_a and C_i are relatively small compared to the difference between C_i and C_c. This drawdown of CO₂ implies that relative limitations imposed by mesophyll conductance to photosynthesis are much greater than those posed by stomata. Warren ([2004]) found that mesophyll limitations (L_m ~ 0.19 to 0.38) were greater than stomatal limitations (L_s ~ 0.05 to 0.23) for Eucalyptus globulus which are similar to our study. Equivalent mesophyll conductance limitations were observed for other sclerophyllous species (L_m ~ 0.20-0.50) by Niinemets et al. ([2009b]). This seems to be explained by leaf anatomical traits such as thickness, density and the leaf mass to area ratio (Flexas et al. [2012]; Niinemets et al. [2009b]; Tomas et al. [2013]) and also by tissue and cell anatomical traits, such as chloroplast surface area exposed to the intercellular air spaces, cell wall thickness and palisade tissue path length, among others (Tomas et al. [2013]; Flexas et al. [2012]; Tosens et al. [2012]; Terashima et al. [2011]; Hassiotou et al. [2009]).

C.albaThere are only a few studies on leaf anatomical traits for the sclerophyllous species considered in this study. Q. saponaria generally exhibits lower density palisade parenchyma and spongy mesophyll compared to C. alba and L. caustica (Gotor [2008]); and additionally L. caustica exhibits substantially thicker cell walls compared to the other sclerophyllous species (Montenegro [1984]). We also found L.caustica leaves to be thicker and denser, as shown by its lower foliage area to mass ratio (M) (7.0 m² kg⁻¹), compared to Q. saponaria (8.7 m² kg⁻¹) and C. alba (10 m² kg⁻¹). Tomas et al. ([2013]) found that for sclerophyllous species g_m was strongly constrained by cell wall thickness (e.g. L. caustica). Such differences in leaf, tissue and cell anatomical traits may at least partially explain why values of g_m and photosynthetic rates are in the series: Q. saponaria > C. alba ≥ L. caustica.

In our study, values of A_sat were strongly and positively correlated with g_m (A_sat = 1.92 + 100.49 g_m, r² = 0.61, P < 0.001). Intercepts (F_2,38 = 8.07, P < 0.01) but not slopes (F_2,38 = 1.6, P = 0.21 ) of this linear relationship were influenced by species (Figure 4). Values of g_s and g_m were uncorrelated (r² = 0.03, P = 0.29). An increase in the rate of photosynthesis with increasing g_m is consistent with previous findings from a wide range of species (Flexas et al. [2004]; Grassi and Magnani [2005]; Loreto et al. [1992]; Niinemets et al. [2009b]; Singsaas et al. [2004]; Tomas et al. [2013]; von Caemmerer and Evans [1991]; Warren et al. [2003]). Using the variable J method and carbon isotopes to estimate g_m for 15 angiosperm species, Loreto et al. ([1992]) estimated that the slope of the relationship between g_m (mol m⁻² s⁻¹ bar-1) and the rate of photosynthesis at saturating irradiance, A_sat (μmol m⁻² s⁻¹) was 0.025 (i.e. A_sat = 40 g_m). However, the magnitude of this slope in our study was 3.45 times greater. Grassi and Magnani ([2005]) estimated that the slope of the relationship between g_m and A_sat was 0.0132 (i.e. A_sat = 75.7 g_m) for oak trees, being our slope 1.8 times greater. This confirms that mesophyll conductance strongly limited the photosynthetic rates in the sclerophyllous species included in our study, compared to other plant groups; although they compare well with other sclerophyllous species.

Figure 4

The relationship between Asat and g _m . A_sat stands for the rate of photosynthesis at saturating irradiance and ambient CO₂ concentration; and, g_m for mesophyll conductance. The A_sat/g_m linear relationship was highly significant (A_sat = 2.4 + 109 g_m, r² = 0.6, P < 0.001; without intercept: A_sat = 138 g_m). Intercepts (F_2,38 = 8.07, P < 0.01) but not slopes (F_2,38 = 1.6, P = 0.21) of the A_sat/g_m linear relationships differed between species. Dashed-lines are A_sat = 40 g_m, r² = 0.89, determined for 15 angiosperms species by Loreto et al. ([1992]).

The ‘constant J method’ used to estimate g_m is sensitive to errors in both the rate of mitochondrial respiration in the light, R_d^*, and the chloroplastic CO₂ compensation concentration in the absence of mitochondrial respiration, Γ^*. Values of Γ^* in our study were very similar across sclerophyllous species (F_2,41 = 0.26, P = 0.78), with a mean value of 50.7 μmol mol⁻¹ (Table 1). In contrast, R_d^*, was significantly lower (F_2,41 = 9.87, P < 0.001) in L. caustica (1.0 μmol CO₂ m⁻² s⁻¹) compared to C. alba (1.7 μmol CO₂ m⁻² s⁻¹) and Q. saponaria (1.5 μmol CO₂ m⁻² s⁻¹) (Table 1).

A recent paper by Gu and Sun ([2014]) suggests that the so-called Laisk method to estimate Γ^* and R_d^* seems to be invalid, and hence an alternative estimate for g_m in our study would reduce the uncertainty related to our results. We tested whether our results for g_m withhold when using values found in the literature for Γ^* and R_d^*. We used values for tobacco from Brooks and Farquhar ([1985]) (Set 1: R_d = 0.8 μmol m⁻² s⁻¹, Γ^* = 36.9 μmol mol⁻¹) and for spinach from Bernacchi et al. ([2001]) (Set 2: R_d = 1 μmol m⁻² s⁻¹, Γ^* = 42.8 μmol mol⁻¹), which have been commonly used to calculate g_m for evergreen species (Gallé et al. [2011]; Hassiotou et al. [2009]; Niinemets et al. [2009a]; Niinemets et al. [2009b]). We did not find significant differences within each species when comparing the three estimates of g_m i.e. using our estimates using the Laisk method, Set 1 and Set 2 (Q. saponaria: P = 0.48, C. alba: P = 0.38, L. caustica: P = 0.12). Tomas et al. ([2013a], [b]) performed a similar analysis for deciduous, semideciduous and evergreen trees and herbs plants, obtaining similar values of g_m independent of the values chosen for Γ^*y R_d^*. This shows that our results hold independent of having used the Laisk method to estimate Γ^*y R_d^*.

Biochemical limitations to photosynthesis

Values of V_cmax (range 13–66, mean 36 μmol m⁻² s⁻¹) and J_max (range 33–148, mean 82 μmol m⁻² s⁻¹) in our study were within the range compiled for 109 C₃ plant species by Wullschleger ([1993]) (V_cmax: range 6–194, mean 64 μmol m⁻² s⁻¹; J_max: range 17–372, mean 134 μmol m⁻² s⁻¹). Wullschleger ([1993]) also provided specific values of V_cmax (range 11–119, mean 47 μmol m⁻² s⁻¹) and J_max (range 29–237, mean 104 μmol m⁻² s⁻¹) for temperate hardwoods and sclerophyllous species (V_cmax: range 35–71, mean 53 μmol m⁻² s⁻¹; J_max: range 94–167, mean 122 μmol m⁻² s⁻¹); which are similar to the ones found in our study. Pena-Rojas et al. ([2004]) observed values of 29 μmol CO₂ m⁻² s⁻¹ and 59 μmol electrons m⁻² s⁻¹ for V_cmax and J_max, respectively for Quercus ilex. Niinemets et al. ([2009b]) observed values of 37 μmol CO₂ m⁻² s⁻¹ and 89 μmol electrons m⁻² s⁻¹ for V_cmax and J_max, respectively for different Australian sclerophyllous species. This emphasizes the fact that photosynthetic capacity in our study was more limited by g_m than by biochemical limitations, as our values of V_cmax and J_max were greater than those reported by Pena-Rojas et al. ([2004]) and Niinemets et al. ([2009b]) and comparable to averages for C₃ hardwoods.

The maximal rate of Rubisco carboxylation (V_cmax) calculated on a C_i basis, were significantly greater in Q. saponaria (49.0 μmol CO₂ m⁻² s⁻¹) compared to C. alba (32.6 μmol CO₂ m⁻² s⁻¹) and L. caustica (26.8 μmol CO₂ m⁻² s⁻¹) (Table 1). Corresponding values of V_cmax on a C_c basis were 98.3, 64.4 and 45.5 μmol CO₂ m⁻² s⁻¹, respectively. Therefore, values of V_cmax calculated on a C_c basis were 101%, 97% and 70% greater than those on a C_i basis for Q. saponaria, C. alba and L. caustica, respectively.

Similar to V_cmax, the maximal rate of electron transport driving regeneration of RuBP (J_max) calculated on a C_i basis, were significantly greater in Q. saponaria (111.4 μmol electrons m⁻² s⁻¹) compared to C. alba (79.6 μmol electrons m⁻² s⁻¹) and L. caustica (56.3 μmol electrons m⁻² s⁻¹) (Table 1). Corresponding values of J_max on a C_c basis were 126.4, 89.9 and 60.4 μmol electrons m⁻² s⁻¹, respectively. Therefore, values of J_max were 13%, 13% and 7% greater on a C_c than a C_i basis for Q. saponaria, C. alba and L. caustica, respectively.

Although there were significant differences in values of J_max and V_cmax among species, the J_max/V_cmax ratio was constant both on a C_i (2.29 ± 0.05) and C_c (1.38 ± 0.04) basis across species (Table 1). Similarly, the relationship between J_max and V_cmax was highly significant both on a C_i (J_max = 2.267 V_cmax, r² = 0.82, P < 0.001) and C_c (J_max = 1.288 V_cmax, r² = 0.81, P < 0.001) basis (Figure 5).

Figure 5

Relationship between the maximum rate of electron transport driving regeneration of RuBP, J _max , and the maximum rate of Rubisco carboxylation, V _cmax , calculated on the basis of chloroplastic CO ₂ concentration, C _c (solid line) and intercellular CO ₂ concentration, C _i (dotted line). On a C_c basis: J_max = 1.288 V_cmax, r² = 0.81, P < 0.001; on a C_i basis: J_max = 2.267 V_cmax, r² = 0.82, P < 0.001

Most values of V_cmax and J_max reported in the literature are calculated from A/C_i response curves, rather than A/C_c curves, with the implicit assumption that mesophyll conductance is infinitely large. When this assumption is invalid, values of V_cmax and J_max are underestimated (Ethier and Livingston [2004]; Ethier et al. [2006]; Grassi and Magnani [2005]; Harley et al. [1992]; Long and Bernacchi [2003]; Loreto et al. [1992]; Manter and Kerrigan [2004]; von Caemmerer [2000]; Niinemets et al. [2009a]). To prove this assumption Grassi and Magnani ([2005]) showed that the relationship between V_cmax,ci and V_cmax,cc values result in a slope of 1.62 (r² = 0.94), showing that the C_i calculation underestimated the real photosynthetic capacity of the leaf. This relation is similar to the one found in our study, which resulted in a slope of 1.28 (r² = 0.76).

There is mounting evidence that g_m changes with chloroplastic CO₂ concentration and irradiance, although some of that variation can be an artifact due to the mathematical methods employed (Gu and Sun [2014]). Also Tholen et al. ([2012]) suggests there are limitations in the precise estimate of g_m when taking several A values from varying CO₂ concentrations and rates of photorespiration to estimate g_m. Taking aside artefactual responses, it seems that g_m initially increases, then peaks to decline thereafter with increasing C_c; while g_m seems to increase with irradiance for the same level of C_c (Flexas et al. [2007]). We are aware that converting A/C_i into A/C_c curves assuming a constant g_m can be invalid; but still useful for comparing our results with previous studies. For instance, the parameters and results found in our study were similar for the photosynthesis model stated by Farquhar et al. ([1980]) with model parameters in normal scenarios of V_cmax ~ 50 μmol CO₂ m⁻² s⁻¹, J_max ~ 125 μmol electrons m⁻² s⁻¹, and dark respiration rates ~ 1.25 μmol CO₂ m⁻² s⁻¹. These values were approximate to an ‘average’ C₃ leaf based on C_c-derived estimates and in the range of the results for the three sclerophyllous species of our study; therefore emphasizing that our species were more limited by g_m that biochemical parameters. We are also aware that using ‘standard’ Rubisco kinetics from tobacco to calculate g_m, V_cmax and J_max may lead to biases (Walker et al. [2013]) to parameterize photosynthesis in the sclerophyllous species considered in this study. This emphasizes the need for future work to estimate Rubisco kinetics for sclerophyllous species in central Chile.

Foliar nitrogen and photosynthetic parameters

Photosynthetic rates are known to be closely related to foliar nitrogen concentrations (Field and Mooney [1986]; Walcroft et al. [1997]; Grassi et al. [2002]; Ripullone et al. [2003]). This is explained by the high proportion of total nitrogen partitioned to the carboxylating enzyme Rubisco (Sage and Pearcy [1987]; Evans [1989]; Warren and Adams [2002]; Takashima et al. [2004]) and also by the strong coupling effect among capacities V_cmax and J_max (von Caemmerer and Farquhar [1981]; Wullschleger [1993]; Sharkey [1985]).

At the plant level, observed N_a ranged almost sixfold from 54 to 339 mmol N m⁻² being significantly greater in C. alba (195.5 mmol m⁻²) compared to Q. saponaria (165.8 mmol m⁻²) and L. caustica (120.2 mmol m⁻²) (Table 1). Values of g_m were uncorrelated to N_a (r² = 0.001, P = 0.40); but as expected values of V_cmax and J_max on a C_i basis significantly increased with N_a (F_1,38 > 5.95, P < 0.019). Intercepts (F_2,38 > 31.3, P < 0.001) but not slopes (F_2,38 < 1.42, P > 0.25) of the V_cmax/N_a and J_max/N_a linear relationships were significantly different between species (V_cmax = a + 0.075 N_a, with a = 17.728 for L. caustica, a = 32.489 for C. alba, a = 39.204 for Q. saponaria, r² = 0.61, P < 0.001; J_max = a + 0.149 N_a, with a = 38.349 for L. caustica, a = 85.976 for C. alba, a = 101.296 for Q. saponaria, r² = 0.62, P < 0.001). It is worth noting that L.caustica exhibited the lowest intercept followed by C. alba and Q. saponaria consistently for both the V_cmax/N_a and the J_max/N_a relationships which can be attributed to reasons other than foliage N (i.e. foliage N was accounted for in the slope of these relationships). We then may speculate that intercept differences among species can be attributed to distinct leaf N investment strategies. Consistently we found significant differences in Nitrogen use efficiency (E_n), being greatest in Q. saponaria (95.9 μmol CO₂ mol N s⁻¹), compared to L. caustica (69.8 μmol CO₂ mol N s⁻¹) and C. alba (56.4 μmol CO₂ mol N s⁻¹) (Table 1). Such differences in E_n, may suggest that Q. saponaria, that exhibits a greater photosynthetic capacity, may invest proportionally more N to Rubisco compared to L. caustica and C. alba that exhibit poorer photosynthetic performance.

Conclusions

In conclusion, values of stomatal conductance, g_s, were greater than those of mesophyll conductance, g_m, while their ratio (g_m/g_s) was not influenced by species, being on average 0.47. Therefore, the relative limitations imposed by g_m were high (L_m ~ 0.40, C_i-C_c ~ 131 μmol mol⁻¹) compared to those imposed by g_s (L_s ~ 0.07, C_a-C_i ~ 32 μmol mol⁻¹). Consequently photosynthetic rates in our study were mainly limited by g_m as biochemical limitations V_cmax and J_max compare well to averages for C₃ plants. Photosynthetic performance was in the series: Q. saponaria > C. alba ≥ L. caustica which can be attributed first to mesophyll conductance limitations, probably mediated by leaf anatomical traits and then to species specific foliage N partitioning strategies.

Appendix

Abbreviations used throughout the text can be found in Table 2.

Table 2 Abbreviations