Comparisons of WUE in twelve genotypes of winter wheat and the relationship between δ¹³C and WUE

Experimental site

The experiment was carried out at the Shunyi Experimental Station (20 m above sea level, 40°N and 116°E), located in the northern part of the Huang (Yellow River)-Huai (Huai River)-Hai (Hai River) Plain of China. The area is in a warm temperate semi-humid continental monsoon climate zone with most of annual rainfall falling during the summer season. In this region, the mean rainfall is 625 mm, of which approximately 75% falls between July and August, and the annual average temperature is 11.5 °C, with average annual sunshine of 2,750 h. The experiment was carried out in the 2015–2016 and 2016–2017 winter wheat growing seasons. Each experimental plot was 10.0 m ×2.4 m, and the plot consisted of 12 lines separated by 20 cm. The levels of organic matter, total nitrogen, rapidly available phosphorus, rapidly available potassium, and pH in a 0–20 cm soil layer were 14.4 g kg⁻¹, 1.1 g kg⁻¹, 24.5 mg kg⁻¹, 10 mg kg⁻¹, and 7.7, respectively; 112.5 kg ha⁻¹ of nitrogen, 75 kg ha⁻¹ of phosphorus pentoxide, and 75 kg ha⁻¹ of potassium oxide were present, and the previously planted crop was field peas (Pisumsativum spp. arvense). The winter wheat was planted at a density of 6.75 × 10⁶ plants ha⁻¹ on October 5, 2015 and October 3, 2016. Winter wheat was harvested on June 12, 2016 and June 11, 2017. Eight near isogenic lines (NILs) along with their donor parent Jing 411 and recurrent parent Jinmai 47 and Shijiazhuang 8 and Yumai 18 were adopted as materials in this experiment. Twelve winter wheat genotypes have different drought-resistant and water-saving characteristics.

In the 2015–2016 and 2016–2017 winter wheat growing seasons, the precipitation totals were 173.4 and 119.6 mm, respectively (Fig. 1), which were less than the average seasonal precipitation (242.6 mm ) by 69.2 and 123 mm, respectively. In the 2015–2016 winter wheat growing season, precipitation occurred mainly from April to June, accounting for 62.9% of the total annual precipitation. By contrast, during the 2016–2017 winter wheat growing season, precipitation occurred mainly from October to November, and the precipitation during this period accounted for 64.4% of the total annual precipitation.

Experimental design

Twelve winter wheat genotypes, which differ in population water use efficiency, were selected as materials, based on the previous studies regarding evaluation of water consumption traits. Rain-fed and supplemental irrigation conditions were set as two water treatments. For rain-fed treatments, no irrigation was conducted after sowing in the winter wheat growing season. For supplemental irrigation treatments, 60 mm water was applied to each plot at the jointing and flowering stages by sprinkling water with a flow meter to record the irrigated water amount. Before sowing, each plot of the two treatments was watered uniformly to field capacity to ensure that the seeds would sprout and grow into intact seedlings. Each experimental growing plot was 10m in length and 2m in width. The plots were arranged in randomized blocks with three replications. Soil moisture content was measured at the sowing stage and final harvest stage, grain yield was determined after harvest, and leaf samples were taken in early, middle, and late grain filling stages for measurement of δ¹³C.

Measurements

Comparison of WUE in 12 genotypes of winter wheat

There were significant differences in WUE in different wheat genotypes under the same water conditions and water condition accounting for 14.51% of the total variance together with two-factor and three-factor interaction between genotype and water condition had significant effect on WUE (Table 1). The method of “Minimum Distance Between Clusters” was used to generate Figs. 2 and 3. Under rain-fed condition, the materials were divided into four groups with different WUE (Fig. 2). There were five genotypes in group I, four genotypes in group II, two genotypes in group III, and just one genotype in group IV. The average value of WUE in group I was significantly higher than that in groups II, III, and IV. The genotypes in group I exhibited a higher grain yield and lower water consumption. The genotypes in group II exhibited moderate grain yield and moderate water consumption. The genotypes in groups III and IV exhibited lower grain yield and higher water consumption. Under supplemental irrigation conditions, the materials were divided into three groups with different WUE (Fig. 3).

Both water consumption and grain yield in winter wheat genotypes were significantly different regardless of rain-fed or irrigation treatment (Tables 2–3). Water condition had the largest effect on grain yield and water consumption compared with other factors and interactions. The average grain yield difference between two water conditions was 1,592 kg ha⁻¹ in 2015–2016, and 1,563 kg ha⁻¹ in 2016–2017 growing seasons, respectively. However, genotypes differed largely in response to soil moisture alteration (Tables 4–5). The largest increase in grain yield occurred in genotype ‘908032’ with 1,925 kg ha⁻¹, and the smallest in genotype ‘908188’ with 1,267 kg ha⁻¹ in 2015–2016. The largest increase in grain yield was in genotype ‘908120’ with 2,494 kg ha⁻¹, and the smallest in genotype ‘908274’ with 835 kg ha⁻¹ in 2016–2017 growing seasons. This implied different responses to supplemental irrigation applied at jointing and anthesis stages, which brought about ultimate different yield responses, among wheat genotypes. The variation of the WUE among 12 genotypes was basically consistent with the grain yield. Under irrigation treatment, the average WUE for wheat genotypes was higher than that under rain-fed conditions by 1.5 kg ha⁻¹ mm⁻¹ in 2015–2016 growing season, and by 1.2 kg ha⁻¹ mm⁻¹ in 2016–2017 growing season, respectively. Similarly, genotypes differed largely in WUE increase under irrigation treatment, compared with rain-fed condition. The largest increase was in genotype ‘908032’ at 3.07 kg ha⁻¹ mm⁻¹, and the smallest increase was in genotype ‘908188’ at 0.13 kg ha⁻¹ mm⁻¹ in 2015–2016 growing season. The largest increase was in genotype ‘908188’ at 2.43 kg ha⁻¹ mm⁻¹ and the smallest increase was in genotype ‘Shijiazhuang 8’ at 0.3 kg ha⁻¹ mm⁻¹ in 2016–2017 growing season. Also, with the same amount of water consumption, genotypes differed significantly in grain yield. These results comprehensively reflected the differences in response to supplemental irrigation during critical water demand stages among genotypes. Moreover, under rain-fed and supplemental irrigation conditions, the grain yields were significantly and linearly positively correlated with the WUE (R² = 0.8231 − 0.9375).

Stable carbon isotope composition (δ¹³C) in flag leaves

The 12 winter wheat genotypes exhibited significant differences in δ¹³C values of flag leaves during early, middle, and late grain filling stages, both under rain-fed and supplemental irrigation conditions (Tables 6–7). For the three measurement time points from early to late grain filling stage, δ¹³C of some genotypes did not change substantially, while some genotypes increased significantly (P < 0.001) under both water conditions (Tables 6–7). The order of 12 winter wheat genotypes in term of δ¹³C values remained nearly the same at three measurements under the two water treatments, indicating the stability of δ¹³C in characterizing plants. The δ¹³C values of flag leaves were significantly higher under rain-fed conditions than that under supplemental irrigation conditions (Tables 6–7).

ANOVA of δ¹³C showed that the factors investigation year (Y), genotype (G), stage (S), water condition (W), and block (B) all had significant effects on δ¹³C (Table 8). Variance in δ¹³C from Y was the largest, covering 43.94% of the total variance. G ranked the second to Y, occupying 25.64%. Proportion of variance from S, W, and B were rather smaller than that of Y and G. This indicated that Y was the most influential variable for δ¹³C, and G had important effects on δ¹³C. Moreover, multiple-factor interactions among Y, G, W, S, and B had significant effects on δ¹³C, except two-factor interactions between S × G, S × W, and S × Y, though the proportions of variance were rather smaller, compared with that from Y and G.

Relationship between the δ¹³C value, grain yield, and WUE

The reliability of using δ¹³C to characterize WUE and grain yield was examined by analyzing the correlation between δ¹³C values, grain yield, and WUE. There was a linear positive significant correlation between the δ¹³C values, the grain yield, and WUE during different grain-filling periods under rain-fed and supplemental irrigation conditions (Figs. 4–7). The δ¹³C values increased during early grain-filling stage to the late grain-filling stage (Tables 6–7). However, the changes in the δ¹³C values did not cause changes in the linear relationships between δ¹³C values, the grain yield, and WUE (Figs. 4–7). Whether under rain-fed conditions or supplemental irrigation conditions, the genotypes with high δ¹³C values exhibited higher grain yield and WUE. On the contrary, the grain yield and WUE in genotypes with lower δ¹³C values were relatively lower. This indicated that the δ¹³C value measurement in winter wheat can be performed at any time during the grain-filling period without affecting its correlation with WUE and grain yield.

Effects of different factors on WUE and δ¹³C

The results indicated that there were significant differences in grain yield and WUE among winter wheat genotypes. Grain yield and WUE were higher under supplemental irrigation conditions in comparison with those under rain-fed conditions. These were consistent with previous researches (Li et al., 2014). The lower WUE of different genotypes under rain-fed conditions was mainly due to the fact that less water supplied during the experimental years and a severe degree of drought stress subsequently affected plant growth. Midgley & Moll (1993) conducted research on the cyclic drought stress in perennial shrubs in South Africa and pointed out that WUE increased with moderate water content of the soil, and when the soil water content was too low, stomatal conductance became conservative and eventually stabilized, and WUE began to decline again.

Previous studies have also shown that there are significant differences in δ¹³C values in C₃, C₄, and CAM plants (Mook, Bommerson & Staverma, 1974; O’Leary, 1984; Farquhar, Ehleringer & Hubick, 1989). Due to the synergistic effect of the absorption and diffusion phases in the enzymatic reaction during CO₂ formation from carboxyl groups after isotope fractionation, the same photosynthetic pathway plants also exhibit the phenomenon of high degree of differentiation. The 12 winter wheat genotypes belonged to C₃ plants, and the δ¹³C values in flag leaves in winter wheat genotypes exhibited significant differences for 3 measurements. The above theories sufficiently explain the results of the present study. Our study also showed that δ¹³C values slowly increased during the three measurements, whether under rain-fed conditions or supplemental irrigation at the jointing and anthesis stages. After plants develop into grain filling stage, strong photosynthesis for rapid accumulation of assimilates began to recede gradually. Instead, distribution and conversion of assimilates to grain turn to predominate this stage. The weakened photosynthesis leads to increased Ci/Ca, which in turn result in reduction of CO₂ entering into mesophyll cells. At this time, the ¹³C could not be fractionated in a timely fashion by the carboxylase of the plants, and then organic matter was directly synthesized with high δ¹³C values (Van de Water, Leavitt & Betancourt, 1994; Feng & Epstein, 1995).

As shown by previous studies, water stress induce stomatal closure, then cause decrease in stomatal conductance and Ci, which in turn leads to an increase in the δ¹³C values of fixed carbon in plant photosynthesis. Conversely, a decrease in δ¹³C values comes out under more water and high nutrient conditions. The δ¹³C values for the different genotypes of wheat were higher under rain-fed conditions than that under corresponding irrigation conditions. The differences in the δ¹³C values among the genotypes in this experiment were greater than that between different measurement periods and different water treatments (Tables 6–8), which indirectly indicates that the δ¹³C values in flag leaves of winter wheat have relative stability. Dramatically, the factor investigation year had the largest effect on δ¹³C, differing from its influence on yield, water consumption, and WUE. δ¹³C was affected by intrinsic photosynthetic capacity, driven by the amount of carboxylating enzyme (Bort et al., 2014; Condon et al., 2002). Precipitation in the wheat grain filling stage of 2016–2017 growing season was larger than that in 2015–2016 growing season (Fig. 1). Plant photosynthesis sensitively responds to water status, as well known. Thus we would rather ascribe the substantial influence of year on δ¹³C to the precipitation difference in grain filling stage between the two investigation years.

Correlations between δ¹³C and WUE

The negative relationship between carbon isotope discrimination and WUE at leaf level for wheat genotypes was first proposed in 1984 (Farquhar & Richards, 1984). In the current study, the δ¹³C values in leaf material showed a significant linear positive correlation with WUE at field population level under the experimental conditions (Figs. 4–7). The correlation between δ¹³C values during the three determination periods at the grain-filling stages with grain yield and WUE reached an extremely significant level, and different water conditions had no effect on the correlation between the δ¹³C values with grain yield and WUE, which is consistent with a large number of previous studies (Wright, Rao & Farquhar, 1994; Condon, Richards & Rebetzke, 2004). The data implied that comparisons might be carried out at any time during the grain-filling stage under various moisture conditions to indicate WUE in winter wheat with δ¹³C values as an indicator.

The theoretical basis of the linear positive correlation between the δ¹³C values and leaf WUE can be obtained from the calculation formula of δ¹³C values and WUE derived from Farquhar, O’Leary & Berry (1982) and Condon, Richards & Rebetzke (2004). Their researches indicate that plant δ¹³C values and WUE have the same influencing factor of Ci/Ca. Therefore, WUE can be deduced by measuring the plant δ¹³C values. However, there are also reports that the δ¹³C values are negatively correlated or not correlated with yield and WUE (Delucia & Schlesinger, 1991; Fengjun et al., 2006; Chamekh et al., 2016). This might be ascribed to the differences in water conditions and growth environments. It can be seen that the degree of water stress may have an effect on the correlation between δ¹³C values with grain yield and WUE, although the specific reasons require further study. Therefore, it is necessary to define water conditions clearly for using δ¹³C value as surrogate trait in screening water-saving crop varieties.

In addition to the effect of water conditions discussed above, the correlation between δ¹³C and WUE seems to be related with crop species and temporal-spatial scale of WUE. Transpiration efficiency (TE), different from field WUE in scale, is biomass accumulation divided by water consumed through transpiration of individual plant in a period of time. In legume plants, TE was reported to negatively correlate or have no correlation with δ¹³C. A research conducted in the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) found TE was negatively associated with △¹³C in groundnut leaves at the final harvest after drought stress imposition, but fairly weak (R² = 0.13) (Krishnamurthy et al., 2007). In lentil (Lens culinaris Medikus), chickpea (Cicer arietinum L.), and lupin (Lupinus angustifolius L.) 3 grain legumes, leaf δ¹³C and TE in the vegetative phase were not significantly correlated under well-watered conditions (Turner et al., 2007). Also, no correlation between TE and leaf △¹³C in groundnut isogenic materials with large variation in TE was found regardless of the water regime (Devi et al., 2011). In wheat, however, positive correlations between △¹³C and field WUE or yield are apt to be reported more by the researches. Leaf △¹³C at anthesis positively correlated with grain yield for spring wheat in the post-anthesis water stress conditions of northwest Mexico. Both leaf △¹³C and grain △¹³C were strongly and positively correlated to grain yield under rain-fed conditions of northern China (Xu et al., 2007). These are in agreement with our results in winter wheat. Therefore, to elucidate and then efficiently use the relations between δ¹³C and WUE in screening water saving varieties, crop species and WUE scale need to be fully considered.

In general, though the accumulating studies proved high stability of δ¹³C characterizing plants, the relationship between δ¹³C value and WUE were determined by diverse environmental factors, crops species, along with WUE scale. Therefore, in order to improve the reliability and scientific accuracy of the δ¹³C value as an indicator to assist with breeding water-saving genotypes, further studies need to be performed to elucidate some definite relations under specific environmental conditions.