Alternate partial root-zone irrigation reduces bundle-sheath cell leakage to CO2 and enhances photosynthetic capacity in maize leaves.

The physiological basis for the advantage of alternate partial root-zone irrigation (PRI) over common deficit irrigation (DI) in improving crop water use efficiency (WUE) remains largely elusive. Here leaf gas exchange characteristics and photosynthetic CO(2)-response and light-response curves for maize (Zea mays L.) leaves exposed to PRI and DI were analysed under three N-fertilization rates, namely 75, 150, and 300 mg N kg(-1) soil. Measurements of net photosynthetic rate (A(n)) and stomatal conductance (g(s)) showed that, across the three N-fertilization rates, the intrinsic WUE was significantly higher in PRI than in DI leaves. Analysis of the CO(2)-response curve revealed that both carboxylation efficiency (CE) and the CO(2)-saturated photosynthetic rate (A(sat)) were significantly higher in PRI than in DI leaves across the three N-fertilization rates; whereas the N-fertilization rates did not influence the shape of the curves. The enhanced CE and A(sat) in the PRI leaves was accompanied by significant decreases in carbon isotope discrimination (Δ(13)C) and bundle-sheath cell leakiness to CO(2) (Φ). Analysis of the light-response curve indicated that, across the three N-fertilization rates, the quantum yield (α) and light-saturated gross photosynthetic rate (A(max)) were identical for the two irrigation treatments; whilst the convexity (κ) of the curve was significantly greater in PRI than in DI leaves, which coincided with the greater CE and A(sat) derived from the CO(2)-response curve at a photosynthetic photon flux density of 1500 μmol m(-2) s(-1). Collectively, the results suggest that, in comparison with the DI treatment, PRI improves photosynthetic capacity parameters CE, A(sat), and κ of maize leaves and that contributes to the greater intrinsic WUE in those plants.

Introduction

Shortage of irrigation water constrains crop production worldwide, and with the projected climate change its impacts will become more significant in the near future (WRI, 2005). To cope with the n class="Chemical">water shortage, it is necessary to develop water-saving agriculture countermeasures, thereby producing more crops per drop. Deficit irrigation (DI) and alternate partial root-zone irrigation (PRI) are water-saving irrigation strategies being widely used in arid and semi-arid regions (Jensen ). DI is a method that irrigates the entire root zone with an amount of water less than the potential evapotranspiration, and the mild stress has minimal effects on the yield (English and Raja, 1996). PRI is a further development of DI; it involves irrigating only part of the root zone, leaving the other part to dry to a pre-determined level before the next irrigation (Kang and Zhang, 2004). By alternately wetting and drying part of the root zone, PRI allows the induction of the abscisic acid (ABA)-based chemical signalling from the drying roots to regulate growth and water use of the shoots, thereby increasing water use efficiency (WUE; Loveys ). Accordingly, DI and PRI treatments are expected to trigger different water deficit stress mechanisms, consequently causing different plant physiological and growth responses (Dodd, 2007; Liu ).

Recently, two meta-analyses have been carried out to examine the relative advantage of PRI over DI in terms of improving crop WUE (Dodd, 2009; Sadras, 2009); although the conclusions of the two analyses are somewhat different, both authors have indicated that n class="Chemical">PRI is superior to DI in enhancing crop WUE across several crop species. During recent years, in order to illustrate the mechanisms underlying the agronomic advantage of PRI over DI, a number of studies have been carried out to reveal the differences in the two types of irrigation in influencing ABA signalling (Dodd , 2009), root growth (Mingo ; Liu ), and crop N nutrition (Kirda ; Wang ; 2010a, b). Most of the studies showed that, under a similar degree of water saving, PRI plants possess significantly stronger ABA signalling, larger root systems, and greater N accumulation than do the DI plants (Mingo ; Dodd ; Wang ), and all of these responses might have contributed to the higher WUE in those plants.

In maize (n class="Species">Zea mays L.), PRI has shown a great potential to save water and increase crop WUE (Kang ; Kirda ; Li ; Hu ). Both Kirda and Li have reported that, compared with DI treatment, PRI could also improve fertilization-N use efficiency in maize plants. Most recently, studies on C3 plants including potato and tomato have demonstrated that PRI could improve plant N nutrition, which might lead to a higher photosynthetic capacity of the leaves (Shahnazari ; Wang ). Even though C4 plants such as maize possess greater photosynthetic N use efficiency in relation to C3 plants and usually have a lower leaf N content threshold for the maximal photosynthetic rate, the response of photosynthesis to leaf N content is much stronger than in C3 plants (Sinclair and Horie, 1989). Thus, a slight increase of N content in maize leaves may significantly enhance carbon assimilation rates. Moreover, it has often been presumed that C4 photosynthesis is more drought resistant than that of C3 plants based on the fact that C4 plants predominate in hot, arid regions which are prone to frequent drought. In addition, a relatively lower stomatal conductance combined with the CO2-concentrating mechanism leads to a greater WUE in C4 compared with C3 plants. However, recent evidence indicates that C4 photosynthesis is equally or even more sensitive to drought stress than its C3 counterpart (Ghannoum, 2009, and references therein).

In C4 plants, the leakiness of the bundle-sheath cells to n class="Chemical">CO2 (Φ), defined as the fraction of CO2 originally fixed by phosphoenolpyruvate carboxylase (PEPC) in the mesophyll that subsequently leaks out of the bundle-sheath cells, quantifies the efficiency of the CO2-concentrating mechanism (Farquhar, 1983). Therefore, it has been frequently used as an indicator for evaluating the C4 photosynthetic efficiency (Williams ; Ubierna ). Studies in maize, sorghum (Sorghum bicolor L.), and other C4 grasses have shown that severe drought stress could significantly increase Φ (Bowman ; Williams , and references therein). However, until now it has remained unknown whether Φ is influenced by water-saving irrigation regimes where the plants are exposed to moderate drought stress.

Here data are presented on gas exchange, photosynthetic CO2–response and light–response curves, n class="Chemical">carbon isotope discrimination (Δ13C), and bundle-sheath cell leakiness to CO2 (Φ) of maize leaves subjected to PRI and DI treatments. In addition, three N-fertilization rates were included in the study in order to clarify the interactive effect between the irrigation regimes and plant N levels. The purpose was to explore the mechanisms by which the photosynthetic capacity of maize leaves is influenced by different water-saving irrigation strategies, and to examine if such effects could account for the superiority of PRI compared with DI in improving WUE.

Materials and methods

Plant material and growth conditions

The experiment was conducted from 1 April to 24 June 2010 in a climate-controlled greenhouse at the experimental farm of the Faculty of Life Sciences, University of Copenhagen, Taastrup, Denmark. At the fifth leaf stage, n class="Species">maize (Z. mays L.) seedlings were transplanted into 19.6 l pots (25 cm diameter and 40 cm deep). The pots were evenly divided into two vertical compartments by plastic sheets, such that water exchange between the two compartments was prevented. The pots were filled with 20.2 kg of naturally dried soil with a bulk density of 1.14 g cm−3. The soil was classified as sandy loam, having a pH of 6.7, total C 12.9 g kg−1, total N 1.4 g kg−1, NH4+-N 0.7 mg kg−1, and NO3−-N 19.1 mg kg−1. The soil was sieved by it passing through a 2 mm mesh and it had a volumetric soil water content (SWC) (%, vol.) of 30.0% and 5.0% at water holding capacity and permanent wilting point, respectively. The average SWC of the soil was monitored by a time domain reflectometer (TDR, TRASE, Soil Moisture Equipment Corp., Goleta, CA, USA) with probes (33 cm in length) installed in the middle of each soil compartment. The climate conditions in the greenhouse were set at: 26/20±2 °C day/night air temperature, 15 h photoperiod, and >500 μmol m−2 s−1 photosynthetic photon flux density (PPFD) supplied by sunlight plus metal–halide lamps. The concentration of CO2 in the greenhouse remained almost stable throughout the experiment, approximately equal to the concentration in the outside air (i.e. 380 μl l−1).

N-fertilization and irrigation treatments

Three N-fertilization rates, namely low N (N1, 75 mg N kg−1 soil), medium N (N2, 150 mg N kg−1 soil), and high N (N3, 300 mg N kg−1 soil), were included in the experiment. The N fertilizer supplied as NH4NO3 was mixed thoroughly with the soil before filling the pots. In addition, P and K were also applied as n class="Chemical">KH2PO4 (380 mg kg−1 soil) and K2SO4 (130 mg kg−1 soil) to the soil to meet the nutrient requirement for plant growth. The maize plants were well watered in the first 10 d after transplanting. Thereafter, the plants were exposed to two deficit irrigation regimes: (i) PRI in which one soil compartment was watered daily to SWC 28% (vol.) while the other was allowed to dry to ∼10 d, then the irrigation was shifted to the other compartment; and (ii) DI in which the same amount of water as used for the PRI plants was used to irrigate the whole pot evenly. The experiment was a complete factorial design comprising nine treatments and each treatment had four replicates. The irrigation water was tap water with negligible concentrations of nutrients. The irrigation treatments lasted for 8 weeks, during which each soil compartment of the PRI plants had experienced six drying/wetting cycles. The average SWC in the pots from 29 to 51 d after onset of treatment (DAT) was calculated to reveal the degree of soil water deficits under the two irrigation regimes.

Leaf gas exchange, chlorophyll content index, water potential, and stable carbon isotope signatures

Diurnal gas exchange [net photosynthetic rate (An) and stomatal conductance (gs)] measurements were made on five sunny days (i.e. 29, 30, 31, 44, and 51 DAT) with a Li-6200 portable photosynthesis system (LiCor Inc., Lincoln, NE, USA) on the first fully expanded leaf, which is the fourth leaf counted from the top of the shoot. The intercellular to ambient CO2 concentration ratio (Ci/Ca) obtained from the gas exchange measurements were used in the calculations of the n class="Disease">bundle-sheath cell leakiness to CO2 (Φ).

The chlorophyll content index (CCI) was taken using a CCM-200 (Opti-Science, Tyngsboro, MA, USA) 44 DAT from the uppermost fully expanded leaf. The CCI values are closely correlated with the Chla and Chlb and total Chl contents in the leaves and, therefore, have been used as a good indicator for photosynthetic capacity of the leaves (Richardson ). All CCI readings were taken midway between the stalk and the tip of the leaf; values of each leaf were the mean of five readings around the same position. Midday leaf n class="Chemical">water potential (Ψl) was measured 45 DAT with a pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, CA, USA) on the leaves just below those used for gas exchange measurements from 10:30 h to 13:00 h.

The plants were harvested on 24 June (51 DAT). The biomass of leaves used for δ13C measurements were determined after drying in an oven at 70 °C for 72 h. Dry samples of leaves were ground to a fine powder for n class="Chemical">13C analysis using the Dumas dry combustion method in a system consisting of an ANCA-SL Elemental Analyser coupled to a 20-20 Tracer Mass Spectrometer (Europa Scientific Ltd, Creve, UK).

The 13C composition (δ13C) of leaf dry biomass was calculated as:where Rsample and Rstandard are the 13C/12C ratios of the leaf sample and Pee Dee Belemnite (PDB) standard, respectively.

Carbon isotope discrimination (Δ13C) of the leaf samples was calculated using Equation 2:where the subscripts a and p refer to air and the plant, respectively (Farquhar ). The δa value for the ambient atmosphere was taken as –7.7‰.

Leakiness of the bundle-sheath cells to CO2 (Φ)

Φ was estimated using the equations derived by Farquhar for C4 photosynthesis. The instantaneous values of Ci/Ca from the 5 d leaf gas exchange measurements were combined with time-integrated values of Δ13C. Using this approach, Φ was estimated as:where a (4.4‰) is the fractionation occurring during diffusion of n class="Chemical">CO2 into the leaf; b4 (–5.7‰) is the combined fractionation due to PEPC (2.2‰) and the activity of carbonic anhydrase in the mesophyll, b3 (30‰) is the fractionation by Rubisco, and s (1.8‰) is the fractionation associated with leakage of CO2 from the bundle sheath to the mesophyll (von Caemmerer ).

Determination of the photosynthetic CO2–response curve

The photosynthetic CO2–response curve was determined using a Ciras-II Portable Photosynthesis System (Ciras-II, PP Systems, UK) with a PPFD level of 1500 μmol m−2 s−1 from 40 to 48 DAT on the same leaves used for gas exchange measurements. Measurements were taken at n class="Chemical">CO2 levels of 350, 250, 150, 100, 50, 500, 700, 1000, and 1500 μl l−1. A values were plotted against the respective intercellular CO2 concentrations (Ci) to produce a response curve. The model of C4 photosynthesis developed by von Caemmerer and Furbank (1999) was used to simulate the CO2 response curves. This model was expressed as:where An is the net photosynthetic rate and x is Ci. Using this equation, the CO2-saturated photosynthetic rate (Asat) was calculated as a+c and the carboxylation efficiency (CE) as the slope at An=0 (calculated as b[a+c]).

Determination of the photosynthetic light–response curve

During the same days on which the photosynthetic CO2–response curves were measured, the photosynthetic light–response curves were developed by measuring An at different PPFD levels using the Ciras-II Portable Photosynthesis System at a n class="Chemical">CO2 concentration of 400 μl l−1. Measurements were taken at PPFD levels of 1000, 1500, 2000, 700, 500, 300, 200, 100, 50, and 0 μmol m−2 s−1. The photosynthetic light–response curve was modelled by a non-rectangular hyperbola. Photosynthetic parameters derived from the light–response curves were determined according to the method described by Richardson . The model was expressed as:where An is the rate of net photosynthesis (μmol CO2 m−2 s−1); Q is the PPFD (μmol m-2 s-1); Amax the irradiance-saturated rate of gross photosynthesis (μmol CO2 m−2s−1); Rd is the dark respiration rate (μmol m-2 s-1); α is the maximum apparent quantum yield of CO2 (mol CO2 mol−1 photons); and κ is a dimensionless convexity term (0 < κ < 1)].

Statistical analysis

The non-linear regression procedure of the statistical programme SPSS 13.0 for Windows (SPSS Inc., Chicago, IL, USA) was applied to fit photosynthetic CO2–response and light–response curves for individual plants. Linear regressions and analysis of covariance (ANCOVA; SPSS) were used to test the effects of irrigation regimes on An and gs across the three N-fertilization rates. Two-way analyses of variance (ANOVA; SPSS) were made to determine the effects of the irrigation regimes and N-fertilization rates on measured SWC, CCI, Ψl, Ci/Ca, Δn class="Chemical">13C, Φ, and the parameters derived from the CO2–response and light–response curves.

Results

Relationship between stomatal conductance and photosynthetic rate

Across the three N-fertilization rates, there were significant linear relationships between An and gs for both the DI and PRI plants (Fig. 1). The slopes of the two regression lines, namely the intrinsic WUE, were significantly greater in the n class="Chemical">PRI than in the DI leaves (P < 0.05) (Fig. 1).

Fig. 1.

Net photosynthetic rate (An) as a function of stomatal conductance (gs) for maize leaves exposed to different irrigation and N-fertilization treatments. The data were collected on five sunny days during the last 4 weeks of the experiment. The two regression lines are made for the DI and PRI plants, respectively, across the three N-fertilization rates. The slopes of the two regression lines, i.e. the intrinsic WUEs, are separated by analysis of covariance (ANCOVA) showing that the intrinsic WUE for PRI is significantly higher than that for DI (P=0.003) (n=50).

Net photosynthetic rate (An) as a function of stomatal conductance (gs) for maize leaves exposed to different irrigation and N-fertilization treatments. The data were collected on five sunny days during the last 4 weeks of the experiment. The two regression lines are made for the DI and n class="Chemical">PRI plants, respectively, across the three N-fertilization rates. The slopes of the two regression lines, i.e. the intrinsic WUEs, are separated by analysis of covariance (ANCOVA) showing that the intrinsic WUE for PRI is significantly higher than that for DI (P=0.003) (n=50).

Soil water content, chlorophyll content index, leaf water potential, carbon isotope discrimination, intercellular to ambient CO2 concentration ratio, and the bundle-sheath cell leakage to CO2

Figure 2a and Table 1 show that, across the three N-fertilization rates, SWC was identical for DI and PRI pots, whereas SWC was significantly affected by N-fertilization rates, and there was a significant interactive effect between irrigation regimes and N-fertilization rates on SWC (Table 1). In DI pots, SWC was the highest under N1, intermediate under N2, and the lowest under N3; whereas in n class="Chemical">PRI, SWC was similar across the three N-fertilization rates. Comparing the two irrigation regimes, SWC of DI was significantly higher than PRI under N1, and the opposite was true under N3; while under N2, SWC was similar for the two irrigation regimes.

Fig. 2.

Average volumetric soil water content in the pots (SWC) (%, vol.), chlorophyll content index (CCI), leaf water potential (Ψl), the intercellular to ambient CO2 concentration ratio (Ci/Ca), carbon isotope discrimination (Δ13C), and the bundle-sheath leakiness to CO2 (Φ) of maize leaves exposed to different irrigation and N-fertilization treatments. Values are the means±SE (n=4 or 92 for SWC). Statistical comparisons (two-way ANOVA) between the irrigation and N-fertilization treatments and their interactions are presented in Table 1.

Table 1.

Output of the two-way analysis of variance (ANOVA) for soil water content (SWC) (%, vol.), chlorophyll content index (CCI), leaf water potential (Ψl), the intercellular to ambient CO2 concentration ratio (Ci/Ca), carbon isotope discrimination (Δ13C), and the bundle-sheath cell leakiness to CO2 (Φ) of maize leaves as affected by the irrigation (DI and PRI) regimes and N-fertilization (N) rates (data are presented in Fig. 2)

Factor	SWC (%, vol.)	CCI	Ψl (MPa)	Ci/Ca	Δ13C (‰)	Φ
Irrigation	NS	*	NS	NS	*	*
N	***	**	*	NS	NS	NS
Irrigation×N	***	NS	NS	NS	NS	NS

*, **, and *** indicate the significance level at P < 0.05, P < 0.01, and P < 0.001, respectively; NS denotes non-significance.

Output of the two-way analysis of variance (ANOVA) for soil water content (SWC) (%, vol.), n class="Chemical">chlorophyll content index (CCI), leaf water potential (Ψl), the intercellular to ambient CO2 concentration ratio (Ci/Ca), carbon isotope discrimination (Δ13C), and the bundle-sheath cell leakiness to CO2 (Φ) of maize leaves as affected by the irrigation (DI and PRI) regimes and N-fertilization (N) rates (data are presented in Fig. 2)

Figure 2b shows that both N-fertilization rates and irrigation regimes had significant effects on CCI; across the two irrigation regimes, N2 and N3 leaves had significantly higher CCI than the N1 leaves; whereas, when analysed across the three N-fertilization rates, it was found that PRI plants had significantly higher CCI than DI plants. However, there were no significant interactions between N-fertilization rates and irrigation regimes on CCI in n class="Species">maize leaves (Table 1).

Across the N-fertilization rates, there was no significant difference for leaf water potential (Ψl) under n class="Chemical">PRI and DI (Fig. 2c, Table 1); whereas across the irrigation regimes, Ψl was significantly higher in N2 than in N1 and N3 (Fig. 2c).

Irrigation regimes had a significant effect on Δ13C and Φ in n class="Species">maize leaves (Table 1), which is especially true under N1 and N2 treatments (Fig. 2e, f). When analysed across the N levels, it was found that both Δ13C and Φ were significantly less in PRI than in DI leaves (Fig. 2e, f, Table 1). However, Ci/Ca was influenced by neither the N-fertilization rates nor the irrigation regimes (Fig. 2d, Table 1). As both Ci/Ca and Φ could affect Δ13C, Δ13C was plotted separately against Ci/Ca and Φ (Fig. 3). It can be seen that only Φ was significantly correlated with Δ13C (Fig. 3b) and there was no clear relationship between Ci/Ca and Δ13C (Fig. 3a), indicating that Φ was the major factor influencing Δ13C in maize leaves (Fig. 3).

Fig. 3.

Relationships between carbon isotope discrimination (Δ13C) and the ratio of intercellular to ambient CO2 concentration (Ci/Ca) and the bundle-sheath leakage to CO2 (Φ) of maize leaves exposed to different irrigation and N-fertilization treatments. Values are the means±SE (n=4).

Photosynthetic CO2–response curve

The model of C4 photosynthesis developed by von Caemmerer and Furbank (1999) was used to simulate the photosynthetic CO2–response curve for n class="Species">maize leaves (Equation 4, Fig. 4). The initial slope of the curve indicates the CE, whereas the plateau denotes the Asat. It was found that both CE and Asat were significantly affected by the irrigation regimes; PRI leaves had significantly higher CE and Asat values than the DI leaves when analysed across the N-fertilization rate; whilst N-fertilization rates did not affect the two parameters (Table 2, Fig. 4).

Fig. 4.

Photosynthetic CO2–response curves of maize leaves exposed to different irrigation and N-fertilization treatments (the measurements were made at a PPFD of 1500 μmol m−2 s−1). The two regression curves (based on Equation 4) are made for the DI and PRI plants, respectively, across the three N-fertilization rates. For DI leaves, the carboxylation efficiency (CE) and the CO2-saturated photosynthetic rate (Asat) were 0.69±0.03 μmol m−2 s−1 and 23.9±1.20 μmol m−2 s−1, respectively; both were significantly less than those for PRI leaves (i.e. 0.98±0.09 and 31.7±2.55, respectively). Statistical comparisons (two-way ANOVA) of the parameters between the irrigation and N-fertilization treatments and their interactions are shown in Table 2.

Table 2.

Output of the two-way analysis of variance (ANOVA) for PEPC carboxylation efficiency (CE, mol m−2 s−1) and the CO2-saturated photosynthetic rate (Asat, μmol m−2 s−1) derived from the photosynthetic CO2–response curve (Equation 4, Fig. 4) and for the maximum apparent quantum yield of CO2 (α, mol CO2 mol−1 photons), the irradiance-saturated rate of gross photosynthesis (Amax, μmol m−2 s−1), the dark respiration rate (Rd, μmol m−2 s−1), and the dimensionless convexity term (κ) derived from the photosynthetic light–response curve (Equation 5, Fig.5) of maize leaves as affected by the irrigation (DI and PRI) and N-fertilization (N) treatments

Factor	CE	Asat	α	Amax	Rd	κ
Irrigation	*	*	NS	NS	NS	*
N	NS	NS	NS	NS	NS	NS
Irrigation×N	NS	NS	NS	NS	NS	NS

indicates the significance level at P < 0.05; NS denotes non-significance.

Output of the two-way analysis of variance (ANOVA) for PEPC carboxylation efficiency (CE, mol m−2 s−1) and the n class="Chemical">CO2-saturated photosynthetic rate (Asat, μmol m−2 s−1) derived from the photosynthetic CO2–response curve (Equation 4, Fig. 4) and for the maximum apparent quantum yield of CO2 (α, mol CO2 mol−1 photons), the irradiance-saturated rate of gross photosynthesis (Amax, μmol m−2 s−1), the dark respiration rate (Rd, μmol m−2 s−1), and the dimensionless convexity term (κ) derived from the photosynthetic light–response curve (Equation 5, Fig.5) of maize leaves as affected by the irrigation (DI and PRI) and N-fertilization (N) treatments

Photosynthetic CO2–response curves of n class="Species">maize leaves exposed to different irrigation and N-fertilization treatments (the measurements were made at a PPFD of 1500 μmol m−2 s−1). The two regression curves (based on Equation 4) are made for the DI and PRI plants, respectively, across the three N-fertilization rates. For DI leaves, the carboxylation efficiency (CE) and the CO2-saturated photosynthetic rate (Asat) were 0.69±0.03 μmol m−2 s−1 and 23.9±1.20 μmol m−2 s−1, respectively; both were significantly less than those for PRI leaves (i.e. 0.98±0.09 and 31.7±2.55, respectively). Statistical comparisons (two-way ANOVA) of the parameters between the irrigation and N-fertilization treatments and their interactions are shown in Table 2.

Photosynthetic light–response curve

The light–response curves of maize leaves under different N-fertilization rates and irrigation regimes are shown in Fig. 5. The light–response curves show important photosynthetic characteristics including α, Amax, κ, and Rd. N-fertilization increased the Amax of n class="Species">maize, but not significantly; whereas the irrigation regimes hardly affected Amax in maize leaves. α, the efficiency of light utilization in photosynthesis, was enhanced with the increase of N levels, although not significantly. Similarly, there were no significant differences for α among irrigation regimes. Interestingly, across the N-fertilization rate, PRI leaves had significantly higher κ that DI leaves (P = 0.046).

Fig. 5.

Photosynthetic light–response curves of maize leaves exposed to different irrigation and N-fertilization treatments (the measurements were made at a CO2 concentration of 400 μl l−1). The two regression curves (based on Equation 5) are made for the DI and PRI plants, separately, across the three N-fertilization rates. For DI leaves, the maximum apparent quantum yield of CO2 (α, mol CO2 mol−1 photons), the irradiance-saturated rate of gross photosynthesis (Amax, μmol m−2 s−1), the dark respiration rate (Rd, μmol m−2 s−1), and the dimensionless convexity term (κ) were 0.037±0.001, 43.1±1.05, 1.35±0.11, and 0.73±0.04, respectively; while for the PRI leaves, the values were 0.035±0.001, 40.2±1.20, 1.87±0.22, and 0.90±0.03, respectively. Here, only the convexity (κ) of the curve for PRI leaves was significantly greater than for the DI leaves. Statistical comparisons (two-way ANOVA) of the parameters between the irrigation and N-fertilization treatments and their interactions are shown in Table 2.

Photosynthetic light–response curves of maize leaves exposed to different irrigation and N-fertilization treatments (the measurements were made at a n class="Chemical">CO2 concentration of 400 μl l−1). The two regression curves (based on Equation 5) are made for the DI and PRI plants, separately, across the three N-fertilization rates. For DI leaves, the maximum apparent quantum yield of CO2 (α, mol CO2 mol−1 photons), the irradiance-saturated rate of gross photosynthesis (Amax, μmol m−2 s−1), the dark respiration rate (Rd, μmol m−2 s−1), and the dimensionless convexity term (κ) were 0.037±0.001, 43.1±1.05, 1.35±0.11, and 0.73±0.04, respectively; while for the PRI leaves, the values were 0.035±0.001, 40.2±1.20, 1.87±0.22, and 0.90±0.03, respectively. Here, only the convexity (κ) of the curve for PRI leaves was significantly greater than for the DI leaves. Statistical comparisons (two-way ANOVA) of the parameters between the irrigation and N-fertilization treatments and their interactions are shown in Table 2.

Discussion

Based on literature surveys and meta-analyses, Sadras (2009) and Dodd (2009) concluded that, given a similar degree of water saving, n class="Chemical">PRI is superior to DI in terms of improving WUE in several crop species. Consistent with this, it was found here that the intrinsic WUE of maize leaves was significantly higher in PRI than in DI plants (Fig. 1). A higher intrinsic WUE could be achieved by either an increases in An or a decrease in gs. In the present study, both possibilities were probably involved, and the trend of changes in An or gs could have been modulated by either the irrigation regime or the N-fertilization treatments, or both. Under N1, while keeping a similar value of An, PRI was seemingly more efficient in reducing gs as compared with DI; whereas under N2, gs ranged from 0.3 mol m−2 s−1 to 0.4 mol m−2 s−1 for both PRI and DI plants, and the An values were generally higher in PRI than in DI. Under N3, the gs and An remained lower in both PRI and DI plants compared with N1 and N2 treatments. Nevertheless, under both N1 and N2 conditions, an improvement in the An/gs ratio, namely the intrinsic WUE, could be achieved. It was confirmed by comparing the slopes of the two regression lines in Fig. 1 that the intrinsic WUE was significantly higher in PRI than in DI plants. The lowered gs in the PRI plants under N1 might have been a result of fine-tuning of stomatal control by the ABA signalling under PRI, as has been suggested in many earlier studies (e.g. Liu , 2009). While most of those studies have exclusively highlighted the significance of a lowered gs induced by the xylem-borne ABA signalling in contributing to the improvement of WUE, the potential role of an increase in the photosynthetic capacity in enhancing crop WUE under PRI irrigation has received less attention. The present study was, therefore, designed to illustrate whether the photosynthetic capacity of maize leaves is improved by PRI in relation to its counterpart DI.

In C4 plants, photosynthetic efficiency can be evaluated by several parameters relating to the efficiencies of n class="Chemical">CO2 transportation, concentration, and fixation. It is widely accepted that measurement of the carbon isotope discrimination (Δ13C) of leaf samples can reflect the environmental influences on the efficiency of photosynthesis in plants (Farquhar ). For instance, measurements of Δ13C in C4 plants have been shown to vary in response to soil water availability (Buchmann ; Saliendra ). Models relating C4 photosynthesis to Δ13C suggest that changes in Δ13C are largely the result of increases in Φ (Farquhar ). Here, the result in maize leaves showed that PRI decreased Δ13C compared with DI and this decrease might have been a result of lowered Φ as exemplified by the significant correlation between the two variables (Table 1, Fig. 3b). In the present experiment, Φ estimated from Δ13C and Ci/Ca (Equation 3) was similar to those reported for other C4 plants (Bowman ; Saliendra ; Meinzer ). Most interestingly, Φ was significantly higher in the DI (0.38) than in PRI plants (0.31); and these values of Φ implied that close to 40% of the CO2 fixed by PEPC and transported to bundle-sheath cells was subsequently leaked back to the mesophyll under DI, while for PRI this values was ∼30%. Bowman have shown that Φ values could reach 0.55 in water-stressed maize leaves. However, the significantly lowered Φ in the PRI leaves as compared with the DI leaves observed here was not due to a better water status of the plants as both SWC and Ψl were identical for the two irrigation regimes (Table 1). Earlier studies have indicated that the magnitude of Φ in C4 plants was determined by both the physical conductance of bundle-sheath cell walls and the balance between Rubisco and PEPC activity (Farquhar ). Particularly under drought stress the increased Φ may be caused by a reduced activity of C3, relative to C4, cycle enzymes; namely a lowered Rubisco to PEPC activity ratio (von Caemmerer and Furbank, 1999). In addition, Ranjith found that the Rubisco to PEPC activity ratio decreases linearly with declining leaf N content. Therefore, a lowered leaf N content could be associated with an increase in Φ. In the present study, although the leaf N content was not determined, based on the fact that a significant portion of leaf N is present in the chlorophyll, the significantly higher CCI value for PRI than for DI plants may indirectly confirm that the former had higher N contents in the leaves, and this might have led to a higher Rubisco to PEPC activity ratio and thereby a lowered Φ in the PRI plants. However, when analysed across all the combinations of the irrigation and N-fertilization treatments, there was no clear link between CCI and Φ of maize leaves (Fig. 2b, f), and Φ did not respond to N-fertilization rates (Table 1). The reasons for the disassociation between the N-fertilization rates/or leaf N status (CCI) and photosynthetic capacity parameters of maize leaves observed here remain unknown. It is speculated that most of the plants used here might have had been grown under luxury N conditions as the N requirement for achieving Amax is much less in C4 than in C3 plants (Sinclair and Horie, 1989). This was unlike the case in the study by Ranjith where the plants were exposed to N stress. Therefore, rather than leaf N content, other unknown factors might have exerted a role in maintaining a small Φ in the PRI maize leaves.

It is believed that a small Φ will be beneficial for achieving a high level of photosynthesis in C4 plants (Farquhar ). von Caemmerer and Furbank (1999) found that increasing Φ could cause a decline in both CE and Asat derived from the CO2–response curve. In good agreement with this, here it was found that, across the three N-fertilization rates, both CE and Asat were significantly less in DI than in n class="Chemical">PRI plants (Fig. 4, Table 2). Similarly finding has been observed by Tahi in tomato plants where the Rubisco carboxylation efficiency was less affected by PRI than by DI treatments in relation to the fully irrigated controls. In addition, increased Φ may result in a decline in the light use efficiency of C4 plants, because CO2 that leaks from the bundle-sheath cells is either lost or re-fixed by PEPC in the mesophyll, thus increasing the energy costs for CO2 fixation (Farquhar, 1983; Hatch ; Watling ). However, analysis of the photosynthetic light–response curve of maize leaves indicated that there was no difference in the quantum yield (α) between the PRI and DI plants (Table 2, Fig. 5). von Caemmerer suggested that the relationship between Φ and α is non-linear, so that a relatively large increase in Φ actually has a rather small impact on α. In addition, the irrigation regimes and N-fertilization rates had no significant effects on most of the photosynthetic parameters derived from the light–response curve of maize leaves, except for the convexity (κ) of the curve, which was significantly higher in the PRI than in DI plants (Table 2). Earlier work by Ögren (1993) had pointed out that κ determines the photosynthetic efficiency in the intermediate light range under natural growing conditions. A greater κ means that An increases much more rapidly to Amax with increasing PPFD. Consistent with this, in the present study, it was observed that at a PPFD of 1500 μmol m−2 s−1, the photosynthetic efficiency, as exemplified by the higher CE and Asat derived from the CO2–response curve, was significantly enhanced under the PRI treatment. As has been discussed previously, the lower Φ could have contributed to the higher CE and Asat in the PRI leaves; however, more investigations are necessary to explore the biochemical and physiological mechanisms underlying the increase in κ under PRI treatment.

In conclusion, the results suggest that, in comparison with the DI treatment, the PRI irrigation enhances WUE of n class="Species">maize leaves, which was associated with an improved photosynthesis capacity as indicated by a lowered Φ and consequently a higher CE and Asat under natural light conditions.