Effects of elevated CO2 on grain yield and quality of wheat: results from a 3-year free-air CO2 enrichment experiment

Authors


  • Editor
    L. de Kok

P. Högy, Institute for Landscape and Plant Ecology, Universität Hohenheim (320), Oekologiezentrum 2, August von Hartmann Str. 3, D-70599 Stuttgart, Germany.
E-mail: pethoegy@uni-hohenheim.de

Abstract

Spring wheat (Triticum aestivum L. cv. TRISO) was grown for three consecutive seasons in a free-air carbon dioxide (CO2) enrichment (FACE) field experiment in order to examine the effects on crop yield and grain quality. CO2 enrichment promoted aboveground biomass (+11.8%) and grain yield (+10.4%). However, adverse effects were predominantly observed on wholegrain quality characteristics. Although the thousand-grain weight remained unchanged, size distribution was significantly shifted towards smaller grains, which may directly relate to lower market value. Total grain protein concentration decreased significantly by 7.4% under elevated CO2, and protein and amino acid composition were altered. Corresponding to the decline in grain protein concentration, CO2 enrichment resulted in an overall decrease in amino acid concentrations, with greater reductions in non-essential than essential amino acids. Minerals such as potassium, molybdenum and lead increased, while manganese, iron, cadmium and silicon decreased, suggesting that adjustments of agricultural practices may be required to retain current grain quality standards. The concentration of fructose and fructan, as well as amounts per area of total and individual non-structural carbohydrates, except for starch, significantly increased in the grain. The same holds true for the amount of lipids. With regard to mixing and rheological properties of the flour, a significant increase in gluten resistance under elevated CO2 was observed. CO2 enrichment obviously affected grain quality characteristics that are important for consumer nutrition and health, and for industrial processing and marketing, which have to date received little attention.

Introduction

Since the industrial revolution, the global atmospheric carbon dioxide (CO2) concentration has risen by approximately 40%, and is currently 387 μl·l−1, the highest for at least 650,000 years (IPCC 2007). According to recent models, CO2 is predicted to reach 550 μl·l−1 by the middle of this century (Meehl et al. 2007). Since CO2 is the major greenhouse gas in the atmosphere, there is no doubt that this increase will affect global climate in the future.

Besides plant growth, elevated CO2 concentration is known to directly alter carbon (C) and nitrogen (N) metabolism of important agricultural C3 species such as wheat, resulting in changes in chemical composition of vegetative plant parts (Cotrufo et al. 1998; Loladze 2002). Therefore, the redistribution and availability of metabolites for developing grain may also be affected, with consequences for both grain yield and quality. In experiments using free-air CO2 enrichment (FACE) technology under realistic field conditions, aboveground biomass and yield increased by 10–16% (550 versus 380 μl·l−1 CO2) when given ample nutrients and water (e.g. Kimball et al. 2002; Long et al. 2006).

Although wheat is one of the world’s major food crops, little or no information is available from FACE experiments regarding the effects of elevated CO2 on grain quality parameters such as nutritional dietary value and industrial processing. CO2-induced responses on important aspects of grain quality of wheat, taking into account different exposure techniques, were recently reviewed (Högy & Fangmeier 2008). Thousand-grain weight (TGW) often increased under elevated CO2, indicating positive effects on grain quality in terms of milling quality and hence economic value (Li et al. 2001; Högy & Fangmeier 2008). However, wheat was unaffected in other FACE studies (Kimball et al. 2001; Manderscheid et al. 2004; Högy et al. 2009), and even shifts towards a smaller grain size were found in a previous FACE study (Högy et al. 2009), resulting in a lower market value.

Predominantly, wheat grains are important sources of carbohydrates and proteins, and used as flour for production of bread and other products (Pomeranz 1987). Data from chamber-based experiments indicate that 80% of the N present in vegetative plant parts is distributed to grains after anthesis (Conroy & Hocking 1993; Fangmeier et al. 1997). Accordingly, grain proteins generally also decrease in FACE studies (Kimball et al. 2002; Högy & Fangmeier 2008; Taub et al. 2008; Wieser et al. 2008), associated with a deterioration of important grain quality parameters for end-use purposes. Besides quantity, protein composition in terms of glutenins and gliadins is important for bread-making properties as these compounds form a viscoelastic gluten complex when in contact with water (MacRitchie et al. 1990; Weegels et al. 1996). Högy et al. (2009) reported lower gluten protein content mainly due to a decrease in gliadins, while both glutens and their fractions were reduced in comparable FACE studies (Wieser et al. 2008), resulting in reduced grain quality properties as a result of the CO2 enrichment. Moreover, Wieser et al. (2008) reported that the glutenin–gliadin ratio increased under elevated CO2, thereby affecting traits fundamental to producing high-quality bread. While the polymeric glutenin fraction is the major protein factor responsible for elasticity, and thus dough strength, monomeric gliadins confer viscosity and extensibility. These more specific quality attributes required by the baking industry are associated with rheological and mixing properties of dough. Within related processing characteristics, Kimball et al. (2001) found a 6% increase in optimum mixing time for bread dough, while only dough resistance was decreased in another FACE experiment (Högy et al. 2009). Experimental evidence for these changes is still poor.

Other minor grain constituents such as lipids and non-starch polysaccharides also play a role in determining wheat flour quality. Moreover, essential amino acids and minerals affect the nutritional value, and starch composition determines the utilisation properties of wheat grains. Currently, no FACE study exists with regard to grain quality traits in terms of lipids and non-starch carbohydrates or starch composition. CO2-induced effects on amino acids and minerals in wheat grains were recently published by Högy et al. (2009), which is currently the only report addressing this topic under FACE conditions. Grains from this study had concentrations of all amino acids reduced by 0.2–8.3%, although effects were only significant for glycine (Gly) and valine (Val); a negative trend was also observed for glutamine/glutamic acid (Glx). These data indicate that CO2 enrichment may affect grain quality in terms of nutritional and processing value. The composition of proteinogenic amino acids was also found to change in the study of Högy et al. (2009), but no clear response pattern of different types of amino acids was observed under CO2 enrichment. In the same FACE experiment, concentrations of macro- and micro-elements other than N remained unaffected under elevated CO2.

Until today, knowledge of CO2-induced impacts on the chemical composition of wheat grains from FACE experiments, except for protein concentration, is scarce. To identify future effects of CO2 enrichment on crop production and grain quality parameters of interest for different markets and utilisation requirements, wheat was grown for three consecutive seasons under field conditions in a FACE system. The aim of this paper is to summarize these results and identify general response patterns for the quality of wheat under rising levels of CO2.

Materials and methods

Plant cultivation

In a 3-year study from 2004 to 2006, spring wheat (Triticum aestivum cv. TRISO) and 13 associated weed species were exposed in a chamberless mini-FACE system. The field site contains 15 circular field plots (2-m diameter) and is located at Heidfeldhof, a farm managed by the Plant Breeding Station of the University of Hohenheim (9°11′ E, 48°42′ N, 395 m asl), south of Stuttgart (Germany). In each of the 3 years, the plots with local soil (slightly stagnic luvisol) were manually dug to a depth of 30 cm at the end of March. Wheat as the main crop was sown in lines at a spacing of 15 cm, seed distance of 1.5 cm and depth of 2 cm to a density of 200 plants·m−2. Based on agronomic practice recommended for spring wheat, the plots were fertilised with 140 kg·N·ha−1, 30 kg·phosphorus (P)·ha−1 and 60 kg·potassium (K)·ha−1 at 50/25/25% at different development stages of the plants (EC 13–29, EC 30–39 and EC 40–59; Tottman & Broad 1987). In addition, on EC 30–39, a trace element fertiliser was applied containing 0.03 kg·boron (B)·ha−1, 0.45 kg·magnesium (Mg)·ha−1, 0.36 kg·sulphur (S) ha−1 and 0.03 kg·manganese (Mn)·ha−1. Soil moisture was measured using time-domain reflectometry (TDR) sensors (IMKO, Ettlingen, Germany). Plots were manually irrigated, if necessary, with the same amount of water applied to all treatments. Irrigation volumes were noted and taken into account for calculation of total water supply, including precipitation amounts.

CO2 exposure

The CO2 exposure involved three treatments with five replicates each: (i) CONTROL plots with ambient air CO2 concentration without technical equipment, (ii) AMBIENT plots with ambient air CO2 concentration and the frame system, and (iii) FACE plots with elevated CO2 concentration and the frame system. The exposure started before seed emergence and continued until final harvest. To achieve a target gas concentration of 150 μl·l−1 above AMBIENT (24-h average), pure CO2 was released directly into the plant canopy through an injection system and regulated according to wind direction and velocity to maintain the set concentration. The CO2 concentrations achieved at canopy height were monitored in the centre of each field plot. Additional information on the performance of the system is given elsewhere (Erbs & Fangmeier 2006; Erbs et al. 2009; Högy et al. 2009).

Biomass harvest

The final harvests took place when wheat plants reached maturity (DC 89; Tottman & Broad 1987) in 2004–2006. The biomass of the central square metre of each plot was cut at ground level and the plants were differentiated into stems, leaves and ears. While leaves and stems were dried at 60 °C in a drying oven, the ears were kept at room temperature to mimic normal storage conditions. All biomass fractions were weighed. Grains were removed from the ears by threshing to determine grain yield and harvest indices. TGW (Condator “E”, Pfeuffer, Winnenden, Germany) and grain size classes (>2.8; 2.8–2.5; 2.5–2.2; <2.2 mm) were determined using a Sortimat (Type K, Pfeuffer).

Sample preparation and chemical grain quality analyses

Wheat grains were ground into wholemeal flour using a laboratory mixer mill (MM 301, Retsch, Haan, Germany) equipped with zircon oxide vessels. The samples were stored in a refrigerator until analyses of minerals, non-structural carbohydrates, lipids and amino acids. For analyses of protein types as well as mixing and rheological properties, the grains were milled to refined flour (Type 550) using a Brabender Quadrumat Junior mill.

Concentrations of N and S were analysed in wholemeal flour according to ISO 10694 (1995) using an elemental analyser (Vario EL, Elementar Analysensysteme, Hanau, Germany). The total protein concentration was calculated by multiplying the concentration of N by a conversion factor of 5.7 for wheat grain. Minerals were analysed after high-pressure digestion with nitric acid (UltraClave III, MLS, Leutkirch, Germany) using (i) inductively-coupled plasma optical emission spectrometry (ICP-OES, Vista Pro Radial, Varian, Palo Alto, CA, USA) for aluminium (Al), calcium (Ca), iron (Fe), K, Mg, sodium (Na), P and silicon (Si); (ii) inductively-coupled plasma mass spectrometry (ICP-MS, Elan 6000, Perkin Elmer Sciex, Durham, NC, USA) for B, cadmium (Cd), chromium (Cr), copper (Cu), lead (Pb), Mn, molybdenum (Mo), nickel (Ni) and zinc (Zn); and (iii) hydride generation atomic absorption spectrometry (Hy-AAS) for Se (selenium) according to the VDLUFA methods VII 2.2.2.5 (2003) and VII 2.2.2.6 (2003). Total and individual non-structural carbohydrates were extracted using pure hot water and analysed by high-performance anion-exchange chromatography (HPAEC) coupled with pulsed amperometric detection (PAD). Starch and fructan were degraded to carbohydrate monomers using amyloglucosidase and fructanase, respectively. The modular system (Bischoff, Leonberg, Germany) was equipped with a RCX-10 column (Hamilton, Reno, NV, USA) and isocratic separation was performed with 100 mm NaOH as eluent. Amylose and amylopectin were measured using a Megazyme Amylose/Amylopectin Assay kit (Megazyme International, Co. Wicklow, Ireland). The samples were analysed for concentration of crude fat, a presumptive estimate of total lipids, by semicontinuous extraction with petroleum ether in a Soxhlet apparatus according to Method 920.39 of the Association of Official Analytical Chemists (AOAC 1995). Total amino acids (protein hydrolysates) were analysed according to European Commission Directive 98/64/EC (1998). A modified Osborne fractionation followed by reverse-phase high-performance liquid chromatography (RP-HPLC) was used to determine protein types in wheat flour (Wieser et al. 1998). Mixing and rheological properties were examined in samples from 2004 to 2005 using micro-scale methods described by Kieffer et al. (1998).

Statistical analysis

Statistical analyses were carried out in five replicates per treatment and year for aboveground biomass, grain yield and grain quality data using SPSS PC+ (version 15, SPSS). Data from AMBIENT and FACE treatments were used for the analyses, while data from the CONTROL treatment was excluded from further evaluation since differences from AMBIENT treatments were not significant. The mean values from individual replicates of ambient (ca) and elevated (ce) CO2 treatments were used as basic input data. The response ratios (ce/ca) of the variables at elevated (ce) versus ambient (ca) CO2 were calculated (Long et al. 2004) and the CO2 effect was computed as relative percentage change [(r − 1) × 100] in response to CO2 enrichment. Data from the experimental years were pooled to form one dataset and were analysed to determine the overall effects of CO2 enrichment independent of the exposure year. The results were expressed as average relative changes (±standard error). For each variable an analysis of variance (anova) was performed, and the results for significant treatment effects are presented as level of probability (P). The expression ‘trend’ denotes differences at 0.05 < P ≤ 0.1, while significant differences due to CO2 were resolved at P ≤ 0.05.

Results

Environmental conditions and CO2 concentrations achieved

In 2006, the growing season mean temperature was about 1.8 °C higher than in previous years, resulting in a shorter growing period (Table 1). Accordingly, the number of growing degree-days was also slightly increased in 2006. Total water supply was highest in 2005, but no differences in soil moisture occurred between individual years. The average set values for CO2 enrichment were more or less achieved in all 3 years.

Table 1.   Climatic conditions and atmospheric CO2 concentrations (24 h mean; average ± standard deviation) determined at the experimental site during the growing period 2004–2006, from wheat emergence until harvest. Temperature data are based on daily averages, irrigation as daily sums. CO2 treatments: AMB: ambient CO2 concentration, FACE: ambient + 150 μl·l−1 CO2 concentration.
yeargrowing periodduration daysmean temperature (°C)growing degree days >5 °Cwater supply (mm)mean soil moisture (vol%)CO2 concentration 24 h mean (μl·l−1)CO2 enrichment (μl·l−1)
AMBFACE
200430 March–04 August12713.4111027824399 ± 10549 ± 12150
200524 March–01 August13013.5112042526403 ± 10572 ± 13169
200630 March–27 July11815.3120238525409 ± 6537 ± 11128

Aboveground biomass fractions and grain yield components

When averaged for all 3 years, total aboveground biomass production was 1011 g·dry·weight·(DW)·m−2 under ambient CO2 concentrations. Aboveground biomass was significantly increased by 11.8% (P = 0.001) under CO2 enrichment, resulting from a higher number of tillers per plant. In the AMBIENT treatment, average stem, ear and leaf biomass were 374, 535 and 96 g·DW·m−2, respectively. In contrast to stem and ear biomass, there was no significant increase in leaf biomass at final harvest (Fig. 1). CO2 enrichment significantly promoted grain yield, which resulted from a higher number of grains per area. The 3-year averages for grain yield were 391 g·DW·m−2 in the AMBIENT treatment and 432 g·DW·m−2 in the FACE treatment. The number of ears per area and the number of grains per ear were not significantly increased in the FACE treatment. Averages from 2004–2006 were 420 and 449·ears·m−2, 26.3 and 28.2·grains·ear−1 and 35.7 and 35.1 g for TGW in ambient and elevated CO2, respectively. The average TGW was not significantly affected by CO2. However, there was a clear shift towards smaller grains according to the results obtained from different grain size classes. While grain yield in size class I (>2.8 mm) decreased by 13.0% (P = 0.001), it increased in size class III (2.5–2.2 mm) by 27.0% (P = 0.042), while size class II (2.8–2.5 mm) showed only a trend (+5.3%; P = 0.055), and no significant CO2 effect was observed in size class IV (<2.2 mm) despite an increase of 10.7%. The harvest index (HI) was not significantly affected by elevated CO2.

Figure 1.

 Impacts of elevated CO2 concentration on biomass fractions and yield components of wheat. Presented are average relative changes (±standard error) due to CO2 enrichment against ambient CO2 concentration, with each of five replicates per treatment for the years 2004–2006. Results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Chemical quality characteristics of grain

The concentrations of the macro-element K and trace element Pb were significantly increased, while the micro-element Fe and trace element Cd decreased under CO2 enrichment (Fig. 2). Moreover, a positive trend was found for the micro-element Mo, while concentrations of the macro-element Mg and trace element Si tended to decrease. No significant increases were observed for the macro-element Na and the micro-element Cr. Concomitantly, several macro-elements such as Ca, P and S and micro-elements such as Zn, Se, Cu, Mn and Ni were not significantly lowered in the high-CO2 treatment. Also, concentrations of the trace elements Al and B were not significantly affected and decreased by 11.7% and 26.3%, respectively. The C/N ratio was significantly increased by 9.2% due to elevated CO2 but C concentration was unaffected (−0.3%, P = ns).

Figure 2.

 Effects of CO2 enrichment on concentrations of macro-elements, micro-elements and trace elements (% DW) in wheat grain. Presented are average relative changes (±standard error) due to elevated CO2 for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

With regard to the composition of total non-structural carbohydrates, the concentrations of fructose on a DW basis were significantly increased in the high-CO2 treatment (Fig. 3); there was also a positive trend for fructan. The concentration of glucose was unaffected, while non-significant increases were observed for all other carbohydrates such as sucrose, raffinose and maltose. In contrast, starch, as the main component of the wheat grain, was not significantly decreased, resulting in a lower concentration of total non-structural carbohydrates. Nevertheless, the composition of starch remained unchanged under CO2 enrichment since amylase decreased by 0.4% (P = ns), while amylopectin increased by 0.1% (P = ns). Apart from starch, there were significant increases in the amount per area of total non-structural and individual carbohydrates of 10.2–15.8% in the high-CO2 treatment (Fig. 4). Although the lipid concentration increased slightly, by 3.6% (P = ns), the amount of lipid per area was significantly increased by 13.7% (P = 0.009) in elevated CO2.

Figure 3.

 CO2-induced effects on concentrations of non-structural carbohydrates (% DW) in wheat grain. Presented are average relative changes and standard errors due to CO2 enrichment against ambient CO2 concentration for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Figure 4.

 Impact of elevated CO2 concentration on the amount of non-structural carbohydrates per area (g·DW·m−2) in wheat grain. Presented are average relative changes and standard errors due to CO2 enrichment against ambient CO2 concentration for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

The total protein concentrations were significantly decreased in the FACE treatment (Fig. 5), while total protein yield was increased by 1.0% (P = ns). Concomitantly, grain protein composition changed under CO2 enrichment, as gluten significantly decreased. This was mainly caused by a significant decrease in gliadins rather than glutenins, resulting in a slightly increased glutenin–gliadin ratio (+3.7%, P = ns).

Figure 5.

 Effects of CO2 elevation on concentrations and composition of wheat grain proteins (% DW). Presented are average relative changes (± standard error) due to elevated CO2 concentrations for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

CO2 enrichment resulted in an overall decrease in the concentrations of proteinogenic amino acids per unit flour weight (Fig. 6). Most of the amino acids decreased significantly in elevated CO2, while only a trend was observed for alanine (Ala). The reduction ranged from 7.1% (serine, Ser) to 10.7% (Glx) and was more pronounced for non-essential amino acids. The concentrations of amino acids considered essential for children were reduced by 7.4% (cysteine, Cys) and 8.9% (tyrosine, Tyr) in the FACE treatment, while semi-essential amino acids decreased by 5.1% (histidine, His) and 5.6% (arginine, Arg). For essential amino acids, reductions due to elevated CO2 ranged from 5.3% for tryptophan (Trp) to 8.7% for isoleucine (Ile).

Figure 6.

 Impact of CO2 exposure on concentrations of amino acids (% DW) in wheat grain. Presented are average relative changes (±standard error) due to CO2 enrichment against ambient CO2 concentration for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Correspondingly, if amino acid concentrations were calculated on a per protein basis, there was a significant reduction only in the concentration of semi-essential His in the FACE treatment (Fig. 7). The concentrations of Glx and Pro tended to decrease in the high-CO2 treatment, while there was a trend that Cys, an essential amino acid for children, and the essential amino acid Trp increased. No significant impacts of CO2 enrichment were found for concentrations of all other amino acids on a protein basis, among which only Gly, Tyr and Phe were slightly decreased.

Figure 7.

 Effects of elevated CO2 on concentrations of amino acids (% Protein) in wheat grain. Presented are average relative changes and standard errors caused to CO2 enrichment against ambient CO2 concentration for each of five replicates per treatment for the years 2004–2006. The results of the anova are denoted by asterisks (trend: (*)0.1 ≥ P > 0.05; significant: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

With regard to processing properties, the only significant CO2-induced impact was the increase in gluten resistance (11.7%, P = 0.041). All other rheological properties, such as resistance of dough (−14.5%), extensibility of dough (+7.8%) and gluten (−7.4%) showed no significant response to CO2 enrichment. No significant CO2 effects were found for the mixing properties of dough, such as water absorption (−1.8%), dough development time (+1.7%) and degree of dough softening (−0.6%).

Discussion

CO2-induced impacts on aboveground biomass fractions and yield components

In our FACE study, the impacts of CO2 enrichment on wheat growth, yield components and quality parameters of mature grain with regard to nutritive value and processing properties were investigated. Kimball et al. (1995) reported that shoot biomass of wheat was increased by an average of 8.4% in 2 years of FACE experiments. This agrees with the present study, where elevated CO2 clearly acted as a C ‘fertiliser’ and significantly increased the total aboveground biomass production by 11.8% (data not shown), mainly resulting from higher biomass allocation towards stems and ears. In previous studies, Fangmeier et al. (1996) reported CO2-induced impacts on both stems and ears in open-top chambers (OTC), while only stems tended to increase in FACE (Högy et al. 2009). In accordance with Högy et al. (2009), biomass allocation towards leaves was not affected in the high-CO2 treatment.

Grain yield increased significantly by 10.4% in elevated CO2 (550 versus 380 μl·l−1), which was associated with production of more grain per unit ground area. The gain in grain yield reported here is in agreement with previous FACE studies, where increases of 10–16% were observed under comparable CO2 conditions (Kimball et al. 2001, 2002; Kimball 2004, 2006; Long et al. 2006; Schimel 2006). In agreement with a previous single-year FACE study (Högy et al. 2009), no significant CO2 effects on other grain yield components, such as grain number per unit ground area and grain number per ear, were found, and also HI was unaffected in the high-CO2 treatment. The latter is in contrast to Kimball et al. (1995), who observed a small but significant increase in HI of 4.4% in comparable FACE experiments. The present result on unchanged TGW supports previous findings (Kimball et al. 2001; Manderscheid et al. 2004; Högy et al. 2009), but an increase in TGW of 7% was found by Li et al. (2001) in elevated CO2. Concomitantly, the grain size pattern was significantly shifted towards smaller grains in the present FACE study, which may be directly related to a lower market value. As larger grains (>2.8 mm), which represented 42.4% (AMBIENT) and 37.0% (FACE) of total grain, decreased by 13.0% and smaller grains (<2.8 mm) increased by 9.4% in the grain size distribution, the net effect on TGW determined at maturity was small and below statistical significance. This is in agreement with previous chamber-based experiments (Högy 2002) and supports findings from a previous single-year data evaluation in FACE (Högy et al. 2009).

Effect of CO2 enrichment on chemical quality characteristics of grain

The beneficial effects on crop biomass and yield of growth in elevated CO2 were counteracted by the mainly negative effect on wholegrain chemical quality. The average total protein concentration per DW was 15.5% (AMBIENT) and 14.2% (FACE), which is quite high for spring wheat produced with a conventional N supply (140 kg·ha−1). An explanation for the rather high protein concentrations may be the comparably low-grain yield obtained in the present study, which in turn resulted from a low planting density (200·plants·m−2). Expected grain yield for the cultivar used in the present study under normal agricultural practice is about 54.7·dt·ha−1 (2004–2006; Amann & Ott 2006), whereas only 39.1·dt·ha−1 were achieved in the AMBIENT treatment.

Based on the modified physiology and biochemistry of wheat plants under CO2 enrichment, the concentration of total protein in grain was significantly decreased by 7.4% in the FACE treatment. The reduction in grain protein due to elevated CO2 is consistent with previous reports (Kimball et al. 2001; Taub et al. 2008; Wieser et al. 2008; Högy et al. 2009), resulting in potentially far-reaching consequences for the nutritional value and use by the processing industry. The lowered grain protein concentration is probably not caused by dilution due to increases in non-protein components (Gifford et al. 2000). The observed phenomenon is considered to be a result of the limited N supply from vegetative plant parts rather than enhanced C accumulation during grain filling under CO2 enrichment. Our findings suggest that current rates of N fertiliser are probably inadequate to maintain existing grain quality standards. Unfortunately, results from chamber-based experiments suggest that the CO2-induced reduction in protein may not easily be overcome by additional N supply since this may simply result in additional biomass and yield production (Fangmeier et al. 1999; Weigel & Manderscheid 2005). Currently, the mechanisms by which elevated CO2 decreases proteins are not well understood. Protein yield was unaffected in the high-CO2 treatment since the increase in yield and decrease in protein concentration more or less balanced each other, which supports earlier findings under FACE conditions (Högy et al. 2009).

Among the grain proteins, the N- and glutamine-rich gliadin fraction was significantly decreased under CO2 enrichment, thereby lowering the gluten concentration that is fundamental in determining physical properties of dough formation and product quality (Weegels et al. 1996; Cornish et al. 2006). Our findings support a previous report by Högy et al. (2009), while Wieser et al. (2008) found a CO2-induced decrease in gluten both due to glutenins and gliadins in a comparable FACE study. In agreement with Högy et al. (2009), in the present FACE experiment, the glutenin–gliadin ratio was unaffected by elevated CO2, while increases were found in another cultivar by Wieser et al. (2008), which may result in different dough properties. However, the accumulation of gluten storage proteins is constrained by N sources to the developing grain (Martre et al. 2003), which may be limited under CO2 enrichment.

Along with the lowered protein concentration and the altered composition of protein fractions in wheat grains, concentrations of total amino acids changed. As gluten proteins are rich in glutamine and Pro, these amino acids were most significantly decreased by 10.7% and 9.7%, respectively, in the high-CO2 treatment. Nearly all other amino acids were also significantly decreased per unit flour weight by 5.1–8.9% in the high-CO2 treatment, and there was a similar trend for Ala. This may have detrimental effects with regard to nutritive value. These results are in agreement with a previous FACE study, where concentrations of all amino acids were lowered, although often not significantly (Högy et al. 2009). The average CO2-induced decreases of different types of amino acid were 5.4% (semi-essential), 7.4% (essential), 8.1% (essential for children) and 8.6% (non-essential). Nevertheless, there is no clear indication whether essential amino acids are more or less affected than non-essential amino acids under CO2 enrichment in FACE studies. In OTC studies, elevated CO2 caused a shift towards relatively more essential amino acids (Manderscheid et al. 1995; Högy et al. 1998). The formation of acryl amide during baking processes may decrease under CO2 enrichment since amino acids such as asparagine, and marginally glutamine and aspartic acid, were found to be lowered. The composition of proteinogenic amino acids was altered due to elevated CO2, confirming previous results from a single-year data evaluation (Högy et al. 2009), and supports the change in protein composition of grain as stated above. Currently, no database on such effects and no information about the physiological processes involved are available on this topic from which to draw final conclusions.

The balance of grain protein fractions is fundamental for producing high-quality bread products. There was only one significant effect of elevated CO2 on mixing and rheological properties defining the wheat grain quality for industrial processing: the resistance of gluten was significantly increased by 11.7% in the present study. In a previous FACE experiment, alterations in bread dough properties such as optimum mixing time (Kimball et al. 2001) and peak resistance (Högy et al. 2009) were reported. However, CO2-induced impacts on functional properties for industrial processing are often less pronounced than changes in grain proteins (Rudorff et al. 1996; Lawlor & Mitchell 2001). In previous FACE studies, bread loaf volume was up by 9% (Kimball et al. 2001; Högy et al. 2009) due to the strong correlation between protein concentration and bread loaf volume. In summary, processing properties may be affected by CO2 enrichment, but in an inconsistent fashion, depending on experimental conditions and cultivars studied.

Among the minerals, concentrations of the macro-element K and trace element Pb were significantly increased under CO2 enrichment; there was also a positive trend for the micro-element Mo. In contrast, concentrations of the micro-element Fe and trace element Cd were significantly decreased, and negative trends were found for the macro-element Mg and trace element Si. Although not significant, concentrations of other minerals, except for Na and Cr, were slightly decreased under higher CO2 levels. Our findings suggest both positive and negative implications for the nutritional value of wheat grain. Moreover, as Fe deficiency currently affects more than 3.5 billion people, this decrease is likely to aggravate worldwide malnutrition. A lack of CO2-induced impacts on minerals in grain was reported in the previous FACE field study of Högy et al. (2009). There are no data available from other FACE experiments. Currently, the impacts of CO2 enrichment on processes related to minerals are not well documented.

Analyses of total non-structural carbohydrates and their fractions revealed significant effects of CO2 enrichment only on fructose concentration, which was significantly increased. Additionally, there was a positive trend for fructan. Although carbohydrates others than starch constitute <3% in total, they are also important as they contribute to the sugar supply required by yeast in the bread-making process. In the present study, the starch concentration, as the main C pool in grains, was unaffected under CO2 enrichment, which is in accordance with our findings on C concentrations. The composition of starch in terms of the balance between amylose and amylopectin did not respond to elevated CO2. Currently, no other data from FACE experiments have been reported. In chamber-based experiments, the carbohydrate composition of grain was also often unaltered, and especially no CO2-induced impact on starch concentration was observed if plants were grown in field soil (Högy & Fangmeier 2008). As CO2 enrichment increases C availability for sink organs, total and individual non-structural carbohydrates per unit ground area were significantly increased in the FACE treatment, apart from starch.

Lipids are essential for the milling properties of flour and for bread-making quality (Stone & Savin 1999). The Soxhlet extraction methodology used in the present study for determination of total lipids is widely used in the food industry for regulatory purposes and nutritional labelling of food products, and recovers all simple, compound and derived lipids that are soluble in ether but sparingly soluble or insoluble in water, primarily triacylglycerols (oils) and phospholipids. The total lipid concentration remained unaffected in the high-CO2 treatment, and consequently the lipid yield per ground area was significantly increased.

Overall, the obvious CO2-induced changes will have implications for grain quality of wheat with regard to healthy food, industrial processing and market value. Experimental evidence for these impacts is still restricted to a few studies and is sometimes contradictory, suggesting that further multi-year FACE research with several cultivars at different locations is required to estimate the consequences for grain quality aspects in a future high-CO2 world.

Acknowledgements

The authors thank Dr M. Erbs, G. Gensheimer and Dr S. Weber for their valued participation in the wheat experiment and for management of the FACE site. H. Stelz from the Experimental Station of Plant Breeding (Heidfeldhof) at the Universität Hohenheim and Dr J.-J. Kim and A. Axthelm from the Hans-Dieter-Belitz Institute for Cereal Research (Garching) are thanked for excellent technical support. We also thank Dr W. Hermann from the Universität Hohenheim, Research Station for Crop Production and Crop Protection, for his support in measuring the grain size classes of wheat.

Corrigendum after online publication
The authors declare no conflicts of interest.

Ancillary