Stable Carbon Isotope Discrimination Is under Genetic Control in the C4 Species Maize with Several Genomic Regions Influencing Trait Expression

In plants with C₄ photosynthesis, physiological mechanisms underlying variation in stable carbon isotope discrimination (Δ13C) are largely unknown, and genetic components influencing Δ13C have not been described. We analyzed a maize (Zea mays) introgression library derived from two elite parents to investigate whether Δ13C is under genetic control in this C₄ species. High-density genotyping with the Illumina MaizeSNP50 Bead Chip was used for a detailed structural characterization of 89 introgression lines. Phenotypic analyses were conducted in the field and in the greenhouse for kernel Δ13C as well as plant developmental and photosynthesis-related traits. Highly heritable significant genetic variation for Δ13C was detected under field and greenhouse conditions. For several introgression library lines, Δ13C values consistently differed from the recurrent parent within and across the two phenotyping platforms. Δ13C was significantly associated with 22 out of 164 analyzed genomic regions, indicating a complex genetic architecture of Δ13C. The five genomic regions with the largest effects were located on chromosomes 1, 2, 6, 7, and 9 and explained 55% of the phenotypic variation for Δ13C. Plant development stage had no effect on Δ13C expression, as phenotypic as well as genotypic correlations between Δ13C, flowering time, and plant height were not significant. To our knowledge, this is the first study demonstrating Δ13C to be under polygenic control in the C₄ species maize.During photosynthesis, plants use light energy to convert atmospheric CO₂ and water into carbohydrates. For the element carbon, there are two stable isotopes, ¹²C and ¹³C. Due to the physical and chemical properties of photosynthetic CO₂ fixation, plants are depleted in ¹³C compared with atmospheric CO₂. In C₃ plants, this discrimination of stable carbon isotopes (Δ13C) has long been used to detect genetic differences of water use efficiency and has been shown to be an accurate predictor for yield under drought (Rebetzke et al., 2002). As Δ13C is linearly related to the ratio of intercellular to atmospheric CO₂ partial pressure (Farquhar et al., 1982), stomatal closure under drought stress is associated with reduced Δ13C. For C₄ plants, our knowledge about the mechanisms underlying Δ13C and about its association with water use efficiency is much more limited. Differences in Δ13C between genotypes of C₄ species have been reported, among others, for sorghum (Sorghum bicolor; Hubick et al., 1990) and maize (Zea mays; Monneveux et al., 2007). However, comprehensive studies analyzing the inheritance of Δ13C have not been performed to date.In C₃ plants, the important steps of CO₂ uptake include the diffusion of atmospheric CO₂ through the boundary layer and the stomata. Subsequently, CO₂ is taken up by the cell and enters the chloroplast, where carboxylation by Rubisco takes place. During photosynthetic carbon fixation, the strongest fractionation of carbon isotopes occurs during the carboxylation reaction of Rubisco (Roeske and O’Leary, 1984). A theoretical model of Δ13C in C₃ photosynthesis has been described by Farquhar et al. (1982), in which Δ13C depends linearly on the ratio of intercellular to ambient partial pressure of CO₂ (p_i p_a−1), and thus provides an indication of stomatal conductance and photosynthetic capacity. Additionally, the model includes the dependency of Δ13C on the fractionation of carbon isotopes during CO₂ diffusion in the air and on the enzymatic properties of the Rubisco enzyme.For rice (Oryza sativa), tomato (Solanum lycopersicum), and wheat (Triticum aestivum), it has been shown that genetic variation for Δ13C is quantitative, genotype-by-environment interaction is small, and the trait heritability is high (Condon and Richards, 1992; Rebetzke et al., 2002; Comstock et al., 2005; Impa et al., 2005). Quantitative trait loci (QTL) for Δ13C have been mapped (Handley et al., 1994; Price et al., 2002; Rebetzke et al., 2008), and in the model plant Arabidopsis (Arabidopsis thaliana), four genes have been identified that are associated with Δ13C. Two are involved in stomatal patterning and thus influence stomatal conductance (Masle et al., 2005; Nilson and Assmann, 2010), and one of them influences photosynthetic capacity as well (Masle et al., 2005). One gene plays a role in cuticular wax composition and is also associated with stomatal conductance (Lü et al., 2012), whereas the fourth gene encodes a cellulose synthase subunit, and mutations in this gene lead to decreased Δ13C. Presumably, this is the result of a decreased cell turgor due to a decreased water transport capacity of the xylem (Liang et al., 2010).For C₄ plants, our knowledge about the genetic mechanisms and physiological processes underlying Δ13C is much more limited, due to the more complex mechanism of CO₂ fixation. The first carboxylation step in C₄ plants takes place in mesophyll cells, in which CO₂ is fixed by phosphoenolpyruvate carboxylase (PEPC). During this reaction, combined with the fractionation of carbon isotopes during HCO₃⁻ formation, carbon is actually enriched in ¹³C (Farquhar, 1983). The C₄ organic acid formed by PEPC is transported to the bundle sheath cells, where CO₂ is released to be fixed by Rubisco in the second step. However, a fraction of CO₂ released in the bundle sheath can diffuse back to the mesophyll cells. The proportion of carbon fixed by PEPC that subsequently leaks out of the bundle sheath cells is termed leakiness (ϕ) and reduces the opportunity of Rubisco to discriminate against ¹³C in C₄ plants. According to the theoretical model by Farquhar (1983), Δ13C and p_i p_a−1 are also linearly related in C₄ plants, but the regression slope is determined by ϕ. Consequently, there can be a positive or a negative correlation of Δ13C and p_i p_a−1 depending on ϕ (Hubick et al., 1990). Regarding the entire fixation process, discrimination against ¹³C in C₄ plants is not as strong as in C₃ plants, and so far there have been few studies reporting a genetic variation of Δ13C in C₄ plants. In sorghum, small but significant differences in Δ13C have been observed among 12 cultivars (Hubick et al., 1990), and similar to C₃ plants, Δ13C has been shown to be correlated with transpiration efficiency (Henderson et al., 1998). Additionally, it has been shown for maize and sugarcane (Saccharum officinarum) that stress conditions lead to an increase in Δ13C (Bowman et al., 1989; Meinzer et al., 1994; Ranjith et al., 1995; Buchmann et al., 1996). Experiments under drought and under well-watered conditions showed higher values for Δ13C in drought-tolerant maize hybrids than in susceptible checks (Monneveux et al., 2007).The use of Δ13C as an indirect trait in breeding for drought tolerance in C₄ species would be highly beneficial, given a stable trait expression and high heritability similar to that in C₃ plants. To assess whether Δ13C can also be used in C₄ plants as an indirect selection trait for drought-tolerant lines, it needs to be shown that Δ13C is under genetic control, although the physiology and molecular mechanisms of Δ13C are not yet fully understood. In this study, we used an introgression library (IL; Eshed and Zamir, 1994) derived from two elite parents to analyze the genetic variation in Δ13C under well-watered conditions. ILs have been successfully used in genetics to identify QTL for various qualitatively and quantitatively inherited traits. An IL is a defined set of nearly isogenic inbred lines derived from repeated backcrosses with one of the parents (recurrent parent RP]) and marker-assisted selection for single fragments (Supplemental Fig. S1). Ideally, each IL line carries a single chromosome fragment of a donor parent (DP) in the genetic background of an RP. Taken together, the different segments cover the whole donor genome, allowing estimation of the effects of single donor fragments in an otherwise identical genetic background (Eshed and Zamir, 1994). The RP of the IL under investigation originates from southeastern Europe and is an elite inbred line of the maize dent pool. As DP, we chose an unrelated maize line representative of the European flint pool. The IL (IL_01–IL_89) was genotyped using the Illumina MaizeSNP50 Bead Chip (Ganal et al., 2011) carrying 56,110 single-nucleotide polymorphism (SNP) markers.Kernel Δ13C of 77 IL lines was measured in the field and in the greenhouse (Δ13C is genetically controlled in the C₄ species maize. Our specific goals were (1) to characterize the genetic architecture of Δ13C (i.e. to determine the number of genomic regions associated with Δ13C), (2) to localize genomic regions influencing Δ13C, and (3) to assess the extent to which genotypic variation in Δ13C might be the result of differences in plant development.

Table I.

Overview of the experiments and experimental designs

Phenotyping Environment	No. of IL Lines Tested	Experimental Design	Recorded Traitsa
Field	89 IL lines RP	RCB with three blocks; 20 plants per experimental unit	Δ13C, Fflow, Mflow, PH, SenL, SPAD, Fv/Fm
Greenhouse	77 IL lines RP DP	RCB with three blocks; one plant per experimental unit	Δ13C, Fflow, Mflow, PH, SenL, SPAD, Fv/Fm
Growth chamber	89 IL lines RP DP	RCB with three blocks; 25 plants per experimental unit	GV

Open in a separate window^aUnits are as follows: Δ13C, ‰; Fflow, d; Mflow, d; PH, cm; SenL, %; GV, %.