Combining carbon-13 and oxygen-18 to unravel triticale grain yield and physiological response to water stress

Abstract

Water availability in semi-arid regions is increasingly becoming threatened by erratic rains and frequent droughts leading to over-reliance on irrigation to meet food demand. Improving crop water use efficiency (WUE) has become a priority but direct measurements remain a challenge. There is therefore a need to identify reliable proxies and screening traits for WUE. Carbon isotope discrimination (Δ13C) offers potential as a proxy for WUE, but its application is hindered by environmental factors and thus varies greatly among different studies. A two-year study was carried out with four moisture levels, ranging from well-watered (430–450mm) to severe stress (SS) (220–250mm), combined with four commercial triticale genotypes grown under field conditions in a hot, arid, steppe climate of Limpopo in South Africa. The study tested the use of Δ13C as a proxy of intrinsic WUE and grain yield of triticale. Second, δ13C and δ18O in combination with measured gas exchanges were used to test the functionality of the dual isotope model to interpret causes of variation in carbon isotope composition. Third, grain filling carbon assimilate sources were inferred from measured flag leaf and grain Δ13C.The results showed that moisture level significantly influenced grain yield, intrinsic WUE and Δ13C in triticale. Well-watered conditions increased grain yield, which ranged from 3.5 to 0.8tha−1 and 4.9–1.8tha−1 in 2013 and 2014 respectively. Δ13C was also high under well-watered conditions and decreased with decreasing moisture level while WUEintrinsic increased with decreasing moisture level. The relationship between Δ13C and grain yield was positive (P<0.01), but only significant under water stressed conditions, indicating dependence of the relationship on moisture level. The relationship between Δ13C and WUEintrinsic did not depend on the moisture level but showed a negative relationship when data for all moisture levels was combined. δ13C showed a negative relationship with photosynthetic rate (A), while the relationship between stomatal conductance (gs) and δ18O varied with season. Hence, the dual isotope model could only predict that variation observed in Δ13C and thus intrinsic water use efficiency was due to a concomitant decrease in both A and gs when transpiration was not limited by evaporative demand. Flag leaf Δ13C measured under SS at GS71 in the 2014 growing season, was significantly higher (2.2–3.6‰) than grain Δ13C, also measured under SS, suggesting minimal contribution of flag leaf photosynthesis to grain filling. No genotypic differences were observed in Δ13C, grain yield and WUEintrinsic, indicating a probable lack of diversity in the studied genotypes.The results of this study show that carbon isotope discrimination could be useful as a predictor of triticale grain yield in drought prone areas. Δ13C also offers potential as a proxy for WUEintrinsic and breeding for lower Δ13C values could result in varieties with higher WUEintrinsic in triticale. Flag leaf photosynthesis and pre-anthesis assimilates contribute much less carbon to grain filling under water stress than previously thought. Lastly, our results show that the dual isotope model is operational, but is not all encompassing but depends on evaporative demand.