Introduction

Maize (Zea mays L.) is an important crop with the C4 pathway of photosynthesis, and is widely planted in arid and semi-arid regions around the world. Water shortage is one of the major problems limiting maize productivity in these regions (Yandigeri et al. 2012; Liu et al. 2013). Water-use efficiency (WUE) is defined as the ratio of plant dry matter produced to total water used over the same period (plant WUE) or, defined in physiological terms (transpiration efficiency at the leaf level), as the ratio between the rates of carbon fixation and transpiration (Zhao et al. 2004). Quick assessment approaches to measuring WUE accurately in maize under water stress may be considered as theoretical guides to water saving in agriculture. However, to obtain accurate WUE value for a specific plant cultivar or to compare WUE among various cultivars is time- and labor-consuming in the field, because accurate measurements of crop transpiration and total dry matter production in the field are difficult (Wright et al. 1994). A series of studies attempting to find a sensitive, quick and exact WUE assessment approach showed that 13C was fractionated by leaves during photosynthesis, and carbon isotope composition (δ 13C) in leaves was positively correlated with WUE in many C3 species when water was limited, such as cotton (Saranga et al. 1998), upland rice (Zhao et al. 2004) and peanut (Wright et al. 1988). However, most of these studies emphasized C3 plants and few dealt with C4 plants.

Isotope composition is usually expressed as the molar ratio of heavier to lighter isotopes (R), and R is generally converted to δ (stable isotope composition) and Δ13C (carbon isotope discrimination),

$$\delta (\permille) = [R({\text{sample}})/R({\text{standard}}) - 1]\; \times \;1000$$
(1)
$$\Delta = \frac{{\delta_{a} - \delta_{P} }}{{1 + \delta_{P} /1000}}$$
(2)

where R(sample) and R(standard) are the molar ratios of the heavier and lighter isotope of the sample and the standard, respectively. δ13CP and δ13Ca are carbon isotope composition of plant and CO2 in atmosphere, respectively, with δ13Ca as a constant of −8.0 ‰ (Farquhar et al. 1989).

In C3 species, the Δ13C during photosynthetic CO2 fixation is defined as photosynthesis Δ13C, which is linearly related with the ratio of intercellular and atmospheric concentrations of CO2 (Ci/Ca) and can be expressed as (Monneveux et al. 2007):

$$\Delta ^{13} {\text{C}} = a + [b_{3} - a](C_{i} /C_{a} )$$
(3)

where a (4.4 ‰) is the discrimination during diffusion of CO2 in air, b 3 (29 ‰) is the discrimination associated with carboxylation by ribulose 1,5-bisphosphate carboxylase (Rubisco), and C i /C a is the ratio of intercellular and atmospheric concentrations of CO2. The Δ13C and transpiration efficiency are both related to C i /C a (Zhao et al. 2004), and WUE positively correlated with δ13C and negatively correlated with Δ13C for C3 species under water-limited conditions has been widely reported (Farquhar et al. 1989; Meinzer et al. 1992; Zhao et al. 2004).

In C4 species, photosynthesis Δ13C in leaf tissue is more complex, involving the diffusion of CO2 through stomata, dissolution and conversion of CO2 to HCO3 −, and fixation of CO2 catalyzed by phosphoenolpyruvate carboxylase (PEPC) in mesophyll and by Rubisco in the bundle sheath (Monneveux et al. 2007). Variation of photosynthesis Δ13C in C4 species depends on both C i /C a and bundle sheath leakiness (φ) during CO2 assimilation (Farquhar 1983; Monneveux et al. 2007). φ is defined as the proportion of CO2 released by the decarboxylation of C4 compounds in the bundle sheath which is not fixed by Rubisco that subsequently leaks back to the mesophyll (Farquhar 1983; Williams et al. 2001). φ may be regarded as a measure of the amount of “overcycling” that occurs in the mesophyll cells in order to raise the partial pressures of CO2 in the bundle sheath cells (Farquhar 1983), and φ is also used as a measure of the extent to which PEPC exceeds Rubisco (Monneveux et al. 2007). The Δ13C in C4 species can be expressed as follows (Farquhar 1983; Henderson et al. 1992):

$$\Delta ^{13} {\text{C}} = a + [b_{4} + \varphi (b_{3} - s) - a](C_{i} /C_{a} )$$
(4)

where b 4 (−5.7 ‰) is the isotopic discrimination during dissolution of HCO3 − and fixation by PEPC, and s (1.8 ‰) is the discrimination during leakiness of CO2 from the bundle sheath to the mesophyll, with other symbols as mentioned above.

Therefore, the correlation between Δ13C and C i /C a in C4 species in theory can be either negative or positive, depending on whether φ is greater or less than (a − b 4)/(b 3 − s) (Monneveux et al. 2007), which lead to an uncertain relationship between δ 13C and WUE.

Rubisco discriminates more against 13C than PEPC, and an increase in φ results in higher photosynthesis Δ13C because b3 is expressed more as shown in Eq. (4) (von Caemmerer et al. 2014). The variation in φ has been widely studied in different biochemical subtypes of C4 plants, and these studies clarified the effects of physiological characteristics on φ, showing that φ depended on the tightness and thickness of the walls of bundle sheath cells (Farquhar 1983; Henderson et al. 1992). However, some studies indicated that φ was significantly influenced by environmental factors such as water stress and atmospheric CO2 elevation (Williams et al. 2001; Fravolini et al. 2002), but different or conflicting results were found. For example, Bowman et al. (1989) reported that φ varied diurnally and increased with increasing water stress. Williams et al. (2001) found that φ was about 0.2 in Sorghum bicolor (L.) Moench. cultivated under different irradiances, temperatures and partial pressures of CO2. Fravolini et al. (2002) pointed out that φ was higher in Eragrostis lehmanniana grown with limited water, while in Aristida glabrata there was low leakiness under well-watered conditions. Therefore, there are no definitive conclusions about the response of φ to environmental change and there is little information on φ in maize.

Since the slope of the relationship between photosynthesis Δ13C and C i /C a in C4 species (Eq. 4) is lower than in C3 species (Eq. 3), variations in cultivar and environment in C i /C a lead to much smaller changes in Δ13C than in C3 species (Monneveux et al. 2007). And leaf δ 13C is usually more deleted in 13C compared with estimated made from measurements of photosynthesis δ 13C (von Caemmerer et al. 2014). Therefore, there is limited information regarding the relationship between δ 13C and WUE in C4 crops. Williams et al. (2001) pointed out that the metabolic mechanism of CO2 in C4 photosynthesis involved coordinated functioning between mesophyll and bundle sheath cells and two carboxylation pathways. Therefore, we hypothesize that the regulation of activities of PEPC and Rubisco in maize under water stress results in variation of φ, and then leads to changes in the relationship between WUE and leaf δ 13C. However, the responses of leaf photosynthesis to water stress significantly affects the fixation of CO2, distribution of assimilates, and also leaf sizes and shapes. Therefore, specific leaf area (SLA, leaf area per unit dry mass) varies with water conditions during plant growth. Wright et al. (1988, 1994) reported that SLA was closely and negatively correlated with WUE in peanut (C3 species), and suggested that breeders could use SLA to screen for high WUE among peanut germplasm within particular environments. However, they also indicated that the mechanisms of the relationship between SLA and WUE were unknown and required further investigations. At present, few studies have focused on how water stress influences the relationship between SLA and WUE in C4 species.

The effects of water stress on leaf δ 13C, φ, SLA and WUE in maize are not well documented. Especially, it is still poorly understood how and whether leaf δ 13C and SLA could be used to assess the influence of water stress on WUE of maize. Therefore, the main objectives of this study were to (1) clarify the mechanisms for the responses of leaf δ 13C, φ, SLA and WUE to water stress, and (2) analyze the relationships between leaf δ 13C, SLA and WUE.

Materials and methods

Experimental design

The experiment was conducted from June 9 to September 24, 2011,and repeated again from June 5 to September 20, 2012, in the Agro-ecological Experimental Station of the Chinese Academy of Sciences in Fengqiu, Henan Province, located in the Huang-Huai-Hai Plain of China (35°01′N, 114°24′E). The soil, classified as sandy loam, was collected from the surface layer (at a depth of 0–30 cm). Field moisture capacity and saturated volumetric water content of the soil were 0.275 cm3 cm−3 and 0.415 cm3 cm−3 respectively.

The maize variety was Zhengdan 958. Plants were grown in steel pots (0.36 m diameter × 0.60 m high), sealed at the bottom with rubber stoppers, which contained 60.00 kg of air-dry soil (equal to 56.15 kg of oven-dry soil). All pots were arranged randomly, and a mobile rainproof canopy was installed over the pots, and only operated when rainfall events occurred to prevent rain water entering the pots. The position of each pot was changed weekly to reduce edge effects. Before sowing, 5.08 g of urea (with 46 % N), 4.42 g of Ca(H2PO4)2 (with 46 % P2O5) and 10.17 g of K2SO4 (with 50 % K2O) were completely mixed with the soil. Additional chemical fertilizer (7.63 g of urea) was applied in each pot at the booting stage when slight N deficiency symptoms were observed. Six maize seeds per pot were sown. Seedlings were thinned to three plants per pot 12 days after planting, and all pots were well watered for 22 days from planting to ensure initially good and uniform seedling growth, when each plant had five leaves. And then the soil surface was covered by two layers of aluminum foil so that almost all water loss could be attributed to plant transpiration. Small slits were opened simultaneously in the aluminum foil for maize growth. After that, the pots were subjected to three water treatments including: W1, with soil moisture kept at field capacity; W2, with soil moisture kept at 75 % field capacity; and W3, with soil moisture kept at 50 % field capacity. Field capacity was measured as the gravimetric water content of saturated soil that was allowed to drain for 6 h in a filter funnel jars with soil equilibrated in darkness (15 days at 25 °C), periodically adjusting soil moisture based on weight loss. The experimental pots were weighted every 3 days and then water was added according to its actual loss. There were 54 pots: three water regimes × six growth stages × three replications. The growth stages of seedling, jointing, booting, tasseling, filling and maturity were observed 12, 44, 57, 75, 89, 107 days after transplanting, respectively.

Measurements

Maize was harvested at seedling, jointing, booting, tasseling, filling and maturity, respectively. At each harvest time, dry matter accumulation, WUE, gas exchange, PEPC and Rubisco activities, stable carbon isotopic component of leaf and stem, leaf area (LA) and SLA were measured, respectively.

Transpiration water loss was measured gravimetrically by weighing the pots every 3 days. Aboveground dry matter of maize in each pot was cut at ground level and roots were washed carefully with tap water. Dry weights of the biomass were determined after drying at 80 °C to a constant weight. WUE was determined by dividing dry weight of the whole plant by the cumulative amount of water transpired during the specific growing stage.

The activities of PEPC and Rubisco were determined on the recently fully-expanded top leaves. PEPC activity was determined spectrophotometrically at 340 nm using an UV-1601 spectrophotometer by coupling the reaction to the oxidation of NADH from malate dehydrogenase (Blanke and Ebert 1992). Rubisco activity was also measured spectrophotometrically at 340 nm using UV-1601 spectrophotometer according to Camp et al. (1982).

Leaf gas exchanges were measured during each growth stage. Photosynthesis rate (A), leaf stomatal conductance (G), intercellular CO2 concentration (C i ), atmospheric CO2 concentration (C a ) and transpiration rate (T) were determined using recently fully-expanded top leaves, by a Portable Photosynthesis System (LI-6400, LI-COR Inc., USA), with a 2 cm × 3 cm natural light leaf chamber. Measurements of leaf gas exchange were conducted only under cloudless or nearly cloudless conditions, and approximately 25–30 s was required to obtain a steady-state reading for each data measurement. During the measurement processes of all growth stages, CO2 in the leaf chamber was set to track the ambient concentration. The light intensities, leaf temperatures and relative humidity changed little and their variation ranges were 1600–1800 µmol quanta m−2, 26–30 °C and 78–83 %, respectively. Therefore, leaf gas exchange of all pots was only influenced by water stress and growth stages in this study.

Both recently fully-expanded leaves and stems were collected at the six growth stages for stable carbon isotope analysis. All plant materials were oven-dried at 80 °C for 48 h and finely ground in a ball mill to powder. The stable carbon isotope composition was determined with an Automated Nitrogen and Carbon Analyzer-Mass spectrometer (Thermo Finnigan America, MAT251). The standard in this study was Pee Dee Belemnite. The analytical precision was 0.1 ‰ and the accuracy was better than 0.1 %.

LA and SLA were measured during the five growth stages except for maturity. LA and SLA were estimated using the following calibration equation based on the measurements of length and maximal width of each leaf,

$$LA = 0.75\; \times \;\sum {L_{i} \; \times \;B_{i} }$$
(5)
$$SLA = LA/m$$
(6)

where L is leaf length (cm), B is maximal leaf width (cm), the subscript of i represents the number of leaves per plant, m is the biomass dry weight of leaves (g).

Estimation of φ

The values of φ were estimated using the data of gas exchange and photosynthesis Δ13C according to Eq. (2), which could be rearranged to yield an expression of φ (Farquhar 1983; Henderson et al. 1992),

$$\varphi = \left[ {\Delta ^{13} {\text{C}}-a + \, \left( {a-b_{4} } \right)C_{i} /C_{a} } \right]/\left[ {\left( {b_{3} -s} \right)C_{i} /C_{a} } \right]$$
(7)

where the symbols are as mentioned above.

Since a, b 4, b 3 and s were relatively constant, then φ could be estimated from Δ13C and C i /C a . In this study, we used leaf Δ13C instead of photosynthesis Δ13C to calculate φ, and leaf Δ13C was calculated by measured value of leaf δ 13C.

Statistical analysis

The regression analyses of δ 13C between leaf and stem were conducted to determine whether fractionation occurred when photosynthetic carbohydrates transported from leaf to stem under water stress. The statistical model used included sources of variation due to replication, water regime and growth stage. All statistical analyses were performed using SAS V.9.0 (SAS Institute Inc., Cary NC, USA).

Results

Dry matter and WUE

Water stress resulted in marked decrease in biomass accumulation (Table 1). The extent of the decrease was dependent on growth stage and degree of water stress. The differences of dry matter accumulation among the three water regimes varied with growth stages, which were significant from tasseling to maturity, but not from seedling to booting. WUE increased with water stress during different growth stages (Table 1). WUE in the W1 treatment was relatively stable and changed little during the whole growth cycle. WUE in W2 increased gradually from jointing to tasseling, and then decreased at maturity. Moreover, WUE in W3 was greatest among the three water regimes and with a range of 7.69–9.93 g kg−1 during the whole growth cycle.

Table 1 Dry matter accumulation, WUE, Δ13C in leaf, φ, leaf area, and specific leaf area during different growth stages of maize under three water regimes
Full size table

PEPC and Rubisco activities

During each growth stage, the Rubisco and PEPC activities decreased significantly with water stress (Fig. 1A). The ratio between the activities of Rubisco and PEPC increased markedly with water stress (Fig. 1B), indicating that PEPC was more sensitive to water deficiency than Rubisco.

Fig. 1
figure 1

Activities of Rubisco and PEPC (A), and the ratio between the activity of Rubisco and PEPC (B) during different growth stages under three water regimes. W1, W2 and W3 represent field capacity of 100, 75, and 50 %, respectively; Rubisco and PEPC represent ribulose 1, 5-bisphosphate carboxylase and phosphoenolpyruvate carboxylase, respectively. The error bars refer to the standard deviation. Different letters on each error bar are significantly different at P = 0.05 according to multiple comparisons by Tukey test

Full size image

Gas exchange

In the W1 treatment, A increased from seedling to booting, then changed little until the filling stage, and decreased markedly at maturity (Table 2). A similar trend was observed in T during the whole growth cycle; C a fluctuated within 368.8–380.4 µmol mol−1 and remained stable in the different growth stages (P > 0.05); the variation in C i /C a was consistent with C i . The same trends also existed in W2 and W3 treatments. The ratio of photosynthesis to transpiration rates was defined as transpiration efficiency (A/T), which represented the instantaneous state at the leaf level, and increased with water stress.

Table 2 Photosynthetic rate (A), transpiration rate (T), transpiration efficiency as A/T, ratio of intercellular and atmosphere concentration of CO2 (Ci/Ca), and stomatal conductance (G) measured during different growth stages under three water regimes
Full size table

A, G, C i , T and C i /C a showed a general trend of decreasing with water stress (Table 2). At each growth stage, the decrease of G with water stress indicated that leaf stomata closed and stomatal resistance increased when maize grew in drought. However, the variation of C i /C a under the three water regimes was less marked (Table 2). This suggested that photosynthesis was not affected by stomatal limitation, but mainly affected by non-stomatal limitation, such as the activities of Rubisco and PEPC (Fig. 1) (Hubbard et al. 2001). A decreased with water stress; however, transpiration efficiency (A/T) increased because the decrease in T was greater than in A with less soil water (Table 2).

Carbon isotope composition and bundle sheath leakiness

All maize was irrigated fully at seedling stage, and the average values of leaf and stem δ 13C were −13.04 and −12.91 ‰, respectively. During the following growth stages, δ 13C in both leaf and stem increased with water stress (Fig. 2). Differences in δ 13C between leaf and stem were observed at each growth stage, and the relationship between leaf and stem δ 13C under three water regimes was shown in Fig. 3. In the well-watered treatment (W1), leaf δ 13C showed no significant difference from stem δ 13C, and the regression line was close to a 1:1 line (Fig. 3), indicating no 13C fractionation when photosynthetic carbohydrates were transported from leaf to stem. However, differences in δ 13C between leaf and stem became greater with water stress.

Fig. 2
figure 2

δ 13C in maize leaf (A) and stem (B) and their relationship under various water treatments. δ 13C represents carbon isotope composition. W1, W2 and W3 represent field capacity of 100, 75, and 50 %, respectively. Dashed line was y = x

Full size image
Fig. 3
figure 3

Relationship between plant WUE and leaf δ 13C (A) and SLA (B) under three water regimes. Other symbols are the same as mentioned above. δ13C, represents carbon isotope composition; WUE, represents water use efficiency; SLA, represents specific leaf area

Full size image

During the whole growth cycle, φ reached highest in seedling. From jointing to mature, φ decreased with water stress, and showed an opposite trend to the variation in δ 13C (Table 1).

LA and SLA

LA increased rapidly from seedling to booting, then varied slightly from booting to maturity, and then decreased with water stress during each growth stage (Table 1). SLA varied with leaf dry weight and leaf area during different growth stages, which was greatest in booting, and decreased in the following growth stages. SLA decreased with water stress, and its variation trend under the three water treatments was opposite to that of WUE.

Relationship between WUE and leaf δ 13C and SLA

Leaf WUE in the photosynthetic process is presented as transpiration efficiency. The leaf is the most important organ for photosynthesis, and so should be the most appropriate one for sampling to determine the relationship between WUE and δ 13C (Zhao et al. 2004). δ 13C and WUE changed little in seedlings without water stress (Fig. 4). From jointing to maturity, leaf δ 13C was significantly positively correlated with WUE. The data of the whole growth cycle under the three water regimes showed a significant positive correlation between WUE and δ 13C. There was a similar trend and significant positive correlation between leaf δ 13C and SLA.

Fig. 4
figure 4

Relationship between Δ13C and C i /C a and φ under three water regimes. C i /C a , represents ratio of CO2 concentration between intercellular and atmosphere; Δ13C, represents carbon isotope discrimination; φ, represents bundle sheath leakiness

Full size image

Discussion

Effects of water stress on PEPC, Rubisco, φ, Ci/Ca and leaf δ 13C

Compared with the control pots (W1), the activities of PEPC and Rubisco both decreased in W2 and W3 treatments (Fig. 1), and A was lower (Table 2), resulting in a marked decrease in dry matter accumulation during different growth stages (Table 1). However, the functional affinity for CO2 and the carboxylation efficiency of PEPC are much higher than those of Rubisco (Henderson et al. 1992). The photosynthetic assimilates are finally formed in bundle sheath cells through the C3 pathway catalyzed by Rubisco. The ratio between the activities of Rubisco and PEPC increased significantly with water stress, implying that PEPC was more sensitive to water stress. At same time, CO2 in the bundle sheath leaked back to the mesophyll cells (φ) decreased with water stress (Table 1) and left more CO2 in bundle sheath cells to be fixed by Rubisco, and assimilated more 13C, and resulted in increased δ 13C in leaves.

During the photosynthetic process of maize, atmospheric CO2 is concentrated through PEPC in mesophyll cells, and then released to bundle sheath cells at high concentrations. The CO2 in the bundle sheath can partly leak out to the mesophyll cells. In present study, leaf Δ13C was used to calculate φ by the theoretical model of Eq. (4), which was also conducted by a series of studies (Farquhar 1983; Henderson et al. 1992; Monneveux et al. 2007). However, it was should be photosynthesis Δ13C rather than leaf Δ13C when φ was calculated. Generally, leaf δ 13C is usually different with photosynthesis Δ13C due to some post-photosynthetic fractionations (von Caemmerer et al. 2014), respiration typically enriched more 13C than the bulk leaf signatures (von Caemmerer et al. 2014) and increased leaf δ 13C, but isotope effects of the biochemical pathways for synthesizing different organic compounds usually deleted more 13C compared with estimated made from measurements of photosynthesis δ 13C (von Caemmerer et al. 2014), which would decreased leaf δ 13C. Because φ was positively correlated with Δ13C and negatively correlated with δ 13C from Eqs. (2) and (7). The reduction of leaf δ 13C might overestimate the value of φ. The difference between leaf δ 13C and photosynthesis δ 13C was not clear. However, all pot experiments in present study were conducted using only a maize variety which avoided the distinction of biochemical subtype, and all plant materials at different growth stages used for stable carbon isotope analysis were recently fully-expanded, which might partially reduce the distinguish brought by respiration. Therefore, we considered that leaf Δ13C used for calculation of φ was acceptable when explored the effect of water stress on φ.

Leakiness (φ) varies with species and growing environment. Henderson et al. (1992) reported that φ was about 20 % through the measurement of several C4 species, while Farquhar et al. (1989) indicated that φ should be close to 37 %. Hattersley and Roksandic (1983) measured 31 C4 species of the Poaceae family and reported that φ was mainly affected by biochemical subtypes within the range of 37–53 %. In the present study, φ in maize varied with water stress and growth stages, in the range of 30.27–44.17 % (Table 1). Vogel (1993) showed that φ was positively correlated with the concentration difference of CO2 between bundle sheath and mesophyll cells, while negatively correlated with the diffusional resistance during leakiness. In the present study, the variation of φ was consistent with that of Rubisco/PEPC (the activity ratio between Rubisco and PEPC; Fig. 1), which increased with water stress. Therefore, the relationship between φ and Rubisco/PEPC implied that φ in maize varied under different water regimes, possibly due to the regulation of activities of PEPC and Rubisco. According to Eq. (4), the values of leaf Δ13C in maize are mainly affected by C i /C a and φ, and φ has a large effect on leaf Δ13C at high C i /C a , and less at low C i /C a . Another interesting calculation result from Eq. (7) is that the value of Δ13C should be a constant of 4.4 ‰ when φ is 33.6 %.

To determine the extent to which Δ13C was influenced by φ and C i /C a , Δ13C was regressed against these two parameters individually (Fig. 4). Leaf Δ13C showed an extremely significant positive correlation with φ under three water regimes (P < 0.01, Fig. 4). However, the negative correlation between leaf Δ13C and C i /C a was significant in W1 (P < 0.05) but not in W2 and W3. The decrease of φ and increase of Rubisco/PEPC finally resulted in more fixed 13C, and leaf Δ13C decreased due to the increase of leaf δ13C (Chaves et al. 2002).

The results of analysis of correlation between leaf Δ13C and φ, and C i /C a showed that the variation of leaf Δ13C was mainly affected by φ, consistent with the findings of Saliendra et al. (1996). However, Madhavan et al. (1991) reported that the main impact factor on leaf Δ13C was C i /C a rather than φ. Williams et al. (2001) pointed out that the attribution of C i /C a and φ to leaf Δ13C varied with water conditions, and φ contributed more in a drought environment. Their findings were inconsistent with our study, in which leaf Δ13C in maize was mainly affected by φ in all three water regimes. In fact, the reason might be the subtype of NADP-ME in maize, and that leaf Δ13C was mainly influenced by φ as for Fravolini et al. (2002), who found that φ was the main factor in the types of NADP-ME and PCK, while C i /C a played a major role in NAD-ME.

Mechanisms for the relationship between WUE and leaf δ 13C

A/T is usually used as the definition of instantaneous WUE at the leaf level. Meinzer et al. (1992) reported that there were two mechanisms in the increase of WUE. The first was that the degree of water consumed (T) reduced more than that of CO2 fixed (A), and as a result G was lower when stomata were closed. The second was that photosynthetic capacity enhanced and enabled the increase in A to exceed the increase in T. The effect of water stress on WUE in maize clearly belonged to the first mechanism, and this mechanism has been widely reported in many other species (Chaves et al. 2002), such as spruce (Sun et al. 1996) and cotton (Saranga et al. 1998).

The relationship between transpiration efficiency (WUE in the photosynthetic process at leaf level) and Δ13C in C4 species can be expressed as follows (Wright et al. 1988; Farquhar et al. 1989):

$$WUE = C_{a} \left( {b_{4} + \varphi b_{3} {-}\varDelta^{13} C} \right)/\left[ {1.6v\left( {b_{4} + \varphi b_{3} {-}a} \right)} \right]$$
(8)

where v is vapor pressure deficiency and the other symbols are as described above.

Based on Eqs. (4) and (8), WUE was positively related to δ13C and negatively to Δ13C—this has been confirmed in many greenhouse and field experiments, especially under conditions of water deficiency (Fotovat et al. 2007). The same relationship between δ 13C and WUE has also been observed in cotton (Saranga et al. 1998), upland rice (Zhao et al. 2004) and other crops (Wright et al. 1988). In the present study, WUE was significantly positively correlated with δ 13C, indicating that δ 13C in leaf could be used as an index for identifying maize WUE in drought conditions. However, some conflicting results have also been reported. For example, Dingkuhn et al. (1991) reported that δ 13C of upland rice was positively correlated with transpiration efficiency, but was not correlated with WUE in the field. Walley et al. (1999) studied wheat exposed to different tillage systems and pointed out that variations in WUE were not consistent with variations in δ13C. In general, these previous studies have been conducted under two conditions. The first was that crops were grown under well-watered conditions, or water source was rich around the crop, and the stomata opened and led to high G and C i /C a , with less 13C fixed by photosynthesis, and low δ13C in leaves. The second was that the studies were conducted in large fields, and the estimation of WUE depended on the calculation methods of water used by crops, with soil evaporation and deep seepage usually included in the crop water consumption.

In the present study, soil evaporation was prevented by the two layers of aluminum foil. The maize transpiration rate reduced more than photosynthetic rate with water stress (Table 2), and thus WUE increased from W1 to W3 treatments (Table 1). However, the leakiness of CO2 from bundle sheath to mesophyll cells decreased with water stress (Table 1), and left more CO2 in bundle sheath cells to be catalyzed by Rubisco due to the regulation of Rubisco/PEPC, which increased with water stress (Fig. 1). The decrease of φ caused more 13C to be assimilated, and then increased δ 13C in leaves. Therefore, WUE was positively correlated with leaf δ 13C under water stress.

Mechanisms for the relationship between WUE and SLA

At each growth stage, SLA decreased with the increase in water stress (Table 2), indicating that leaves of maize became thicker in response to water deficiency. Although respiration and other non-photosynthetic processes played a role during these physiological processes, we believe that the main reason for the variation of SLA was that leaf expansion and translocation of assimilates from the leaf were prevented by water stress, resulting in thicker leaves and lower SLA—this was confirmed by the differences of Δ13C between leaf and stem (Fig. 3). Compared with stem Δ13C, leaf Δ13C decreased with water stress, indicating that more 13C was enriched in leaves when maize suffered from drought, implying that more assimilates accumulated in leaves rather than being transported to stems under water stress, which resulted in thicker leaves and lower SLA.

Water stress affected leaf expansion and translocation of assimilates from leaves, with less 13C transported to stems and more 13C enriched in leaves. Leaf δ13C increased and Δ13C decreased simultaneously. In addition, leaves became thicker and SLA decreased. Therefore, SLA was negatively correlated with WUE (Fig. 4), because SLA was negatively correlated with leaf δ 13C.

Conclusion

Both photosynthetic rate and dry matter accumulation in maize decreased under water stress, due to the decreased activities of Rubisco and PEPC. Additionally, leaf stomata closing decreased the stomatal conductance, and also reduced transpiration rate. Because the transpiration rate was reduced more than the photosynthetic rate, the WUE increased with water stress. The leakiness of CO2 from bundle sheath to mesophyll cells decreased with water stress, and left more CO2 in bundle sheath cells to be catalyzed by Rubisco due to the regulation of Rubisco/PEPC. The decrease of φ caused more 13C to be assimilated, and then increased δ 13C in leaves. Therefore, WUE was positively correlated with leaf δ 13C under water stress. There was no fractionation between leaf and stem when maize was grown in the well-watered condition. However, the fractionation became greater with water stress, indicating that water stress affected leaf expansion and translocation of assimilate from leaves, resulting in thicker leaves and lower SLA. Therefore, WUE was negatively correlated with SLA under water stress. The relationship between WUE and δ 13C and SLA in maize implied that leaf δ 13C and SLA could effectively reflect the drought adaptation and high WUE under different water conditions.