The Effects of Inorganic Nitrogen form and CO2 Concentration on Wheat Yield and Nutrient Accumulation and Distribution

Inorganic N is available to plants from the soil as ammonium (NH4+) and nitrate (NO3-). We studied how wheat grown hydroponically to senescence in controlled environmental chambers is affected by N form (NH4+ vs. NO3−) and CO2 concentration (“subambient,” “ambient,” and “elevated”) in terms of biomass, yield, and nutrient accumulation and partitioning. Wheat supplied with NH4+ as a sole N source had the strongest response to CO2 concentration. Plants exposed to subambient and ambient CO2 concentrations typically had the greatest biomass and nutrient accumulation under both N forms. In general NH4+-supplied plants had higher concentrations of total N, P, K, S, Ca, Zn, Fe, and Cu, while NO3--supplied plants had higher concentrations of Mg, B, Mn, and NO3- - N. NH4+-supplied plants contained amounts of phytate similar to NO3−-supplied plants but had higher bioavailable Zn, which could have consequences for human health. NH4+-supplied plants allocated more nutrients and biomass to aboveground tissues whereas NO3+-supplied plants allocated more nutrients to the roots. The two inorganic nitrogen forms influenced plant growth and nutrient status so distinctly that they should be treated as separate nutrients. Moreover, plant growth and nutrient status varied in a non-linear manner with atmospheric CO2 concentration.


INTRODUCTION
Nitrogen (N) is the mineral element that most often limits plant growth and primary productivity in natural and agricultural systems. Plants usually acquire N from the soil in the forms of ammonium (NH + 4 ) and nitrate (NO − 3 ), and management of these forms is vital to agriculture. Wheat can utilize either form alone (Wang and Below, 1992), but mixed N nutrition (e.g., NH 4 NO 3 ) typically produces the best grain yields and quality in hydroponically grown (Gentry et al., 1989;Heberer and Below, 1989;Wang and Below, 1995) and field-grown plants (Bock, 1987;Camberato and Bock, 1990).
Ammonium and nitrate affect crops differently when either is supplied as the sole N source (Bloom, 1997). Ammonium requires less energy to assimilate into organic compounds (Bloom, 1997), but can prove toxic if it accumulates to high concentrations within plant tissues (Cox and Reisenauer, 1973;Goyal and Huffaker, 1984). Nitrate is generally the predominant form available in aerated, temperate agricultural soils (Haynes, 1986;Bloom, 1997), and may accumulate within plant tissues to high concentrations without toxicity (Goyal and Huffaker, 1984). In wheat, the N form supplied has been found to influence many physiological parameters profoundly including biomass (Wang and Below, 1995Bloom et al., 2002), leaf area (Bloom et al., 2002), tillering (Chen et al., 1998), seed mass (Wang and Below, 1995), protein content (Wang and Below, 1995), and mineral nutrient acquisition and distribution (Gashaw and Mugwira, 1981;Wang and Below, 1998), although such differences can vary among cultivars (Gashaw and Mugwira, 1981;Wang and Below, 1995).
The presence of NH + 4 , as either a sole N source or in mixed N nutrition, increased organic N concentration in shoots, roots, and grain and decreased partitioning of dry matter to the roots in wheat (Wang and Below, 1995). Decreased cation uptake has been found in wheat under NH + 4 nutrition (e.g., Gashaw and Mugwira, 1981;Wang and Below, 1998), although results varied among cultivars (Gashaw and Mugwira, 1981). For example, NH + 4 nutrition decreased whole plant and shoot accumulations of K, Cu, Ca, Mg, Fe, Mn, and Zn (Wang and Below, 1998). Nutrient allocation to plant tissues also varied between N forms. NH + 4 -fed plants distributed a smaller percentage of total P, K, Cu, and B to roots relative to NO + 3 -fed plants Below, 1995, 1998). Also, a greater percentage of reduced N was allocated to the shoots in NH + 4 -fed plants (Wang and Below, 1995). Elevated atmospheric concentrations of CO 2 alter growth and N dynamics of wheat and other C 3 plants. Under elevated CO 2 , wheat has lower protein and N concentrations (e.g., Thompson and Woodward, 1994;Bloom et al., 2002;Wu et al., 2004), and lower macro-and micronutrients concentrations (Manderscheid et al., 1995;Fangmeier et al., 1997Fangmeier et al., , 1999Wu et al., 2004;Högy and Fangmeier, 2008). Grain phytate concentrations are also thought to increase or remain the same under elevated CO 2 , and NO − 2 reductase activities under elevated CO 2 (Bloom et al., 2002).
The interaction between atmospheric CO 2 concentration and inorganic N form and how it influences plant growth and nutrient concentrations has not been examined in wheat or any other crop species grown to senescence. Here, we grew wheat hydroponically in controlled environment chambers and measured mineral nutrition, biomass, and nutrient allocation in response to three concentrations of atmospheric CO 2 (subambient, ambient, and elevated) and two forms of N nutrition (NH + 4 and NO − 3 ). We tested the following hypotheses: (1) plant nutrient concentrations and allocation patterns will respond differently to CO 2 enrichment under the two N forms, and (2) NO − 3 -fed plants will show a smaller biomass and yield enhancement in response to CO 2 enrichment than NH + 4 -fed plants as a result of CO 2 inhibition of shoot NO − 3 assimilation. Also, we observed both differences in the Zn concentration between plants grown on NH + 4 and NO − 3 and a strong dependence of Zn absorption on Zn and phytate concentration, indicating that phytate and bioavailable Zn are affected by N form and CO 2 . Therefore, we used the well supported Miller equation (Miller et al., 2007) to estimate how N and CO 2 might impact a hypothetical human population. Iron, another important micronutrient that forms complexes with phytate, was not analyzed because we observed no significant differences in iron concentrations between the N forms and because how best to estimate Fe absorption in humans is still uncertain (Welch and Graham, 2004).

EXPERIMENTAL
Wheat seeds (Triticum aestivum cv. Veery 10) were surface sterilized for one minute in 2.6% sodium hypochlorite solution and thoroughly rinsed with DDI water. The seeds were then rolled up in germination paper saturated with 10 mM CaSO 4 . The germination paper was placed in a 400 mL beaker with approximately 75 mL of 10 mM CaSO 4 solution, covered with a plastic bag and placed in an incubator (23˚C) for four days. Seedlings were transplanted into 20 L tubs filled with an aerated nutrient solution that contained 1 mM CaSO 4 , 1 mM K 2 HPO 4 , 1 mM KH 2 PO 4 , 2 mM MgSO 4 , and 0.2 g L −1 Fe-NaEDTA and micronutrients (20% of a modified Hoagland's solution with either 0.2 mM KNO 3 or 0.1 mM (NH 4 ) 2 HPO 4 as the N source, Epstein and Bloom, 2005). The nutrient solution was replaced weekly and an additional 0.2 mM of NO − 3 -or NH + 4 − N was added midweek until harvest. The solution volume was maintained by daily addition of deionized water. Solution pH varied between 6.8 and 7.0 for both of the N forms, and the NH + 4 and the NO − 3 solutions did not differ by more than 0.1 pH units.
The plants were grown in controlled environment chambers (Conviron,Winnipeg, Canada) set at 23/20˚C day/night at 60-70% relative humidity with a photoperiod of 15 h. The photosynthetic flux density was 375 µmol m −2 s −1 at plant height. Plants were subjected to one of three CO 2 concentrations: "subambient" (310 ± 30 ppm), "ambient" (410 ± 30 ppm), and "elevated" (720 ± 5 ppm). Subambient CO 2 concentrations were maintained by passing air that entered the growth chamber through wet soda lime, a mixture of KOH, NaOH, and Ca(OH) 2 that was replaced as needed. The elevated CO 2 conditions were maintained in an environmental chamber equipped with non-dispersive infrared analyzers for CO 2 (Horiba model APBA-250E) and valves that added pure CO 2 to the incoming air stream to hold the chamber concentration at 720 ppm.
The wheat was grown until all aboveground parts turned completely yellow. Plant matter was sorted into grain, chaff, shoots, and roots and dried for 48 h at 55˚C. Data on kernel number (KN), kernel mass, number of heads, kernels head −1 , and HI were collected prior to sample preparation for nutrient analysis. A portion of the grain was analyzed for phytate using a modification of the method as described by Haug and Lantzsch (1983). The remainder of the grain as well as the shoots and chaff was bulked into five repetitions per treatment and sent to the UC Davis Analytical Laboratory for nutrient analysis. The roots of plants for each CO 2 × N treatment became entangled within the same tub; therefore, we were unable to separate the roots of the individual plants for analysis. Root data are thus presented as means for each treatment with no standard errors or confidence intervals.
Data were analyzed using PROC MIXED (SAS 9.0 Cary, NC, USA). Nitrogen form and CO 2 factors were treated as fixed independent variables. We used the Tukey-Kramer Honestly Significant Difference test for mean separation. Probabilities less than 0.05 were considered significant. Because some of the transformed variables did not meet the assumption of homogeneity Frontiers in Plant Science | Plant Nutrition of variances, but one-way ANOVAs met the ANOVA assumptions, we analyzed the results via one-way ANOVAs to gain some information on the interactions between CO 2 and N form.

MODELING THE INFLUENCE OF N FORM ON Zn NUTRITION IN THE HUMAN DIET
We used a database derived from the United Nation's Food and Agriculture Organization (FAO)'s national food balance sheets (FBS) to estimate the average daily per capita dietary intake of zinc and phytate from 95 different food commodities in each of 176 countries. This database combines FAO data on per capita intake of food commodities with USDA data on the nutrient or phytate content of each of these commodities. More detailed discussion of the creation of this database for the International Zinc Collaborative Group may be found in Wuehler et al. (2005). Using this database, we produced two datasheets: one containing per capita daily dietary intake of zinc from each food commodity for each country and another containing per capita phytate intake from each food commodity for each country. To calculate total dietary zinc (TDZ) and total dietary phytate (TDP) per country, we summed across the rows of all food commodities for each respective country.
To determine the proportion of a population at risk for zinc deficiency from a hypothetical least developed country (LDC), we first calculated TDP and TDZ values for a set of 44 countries defined by the United Nations as being least developed. We took the mean TDP and TDZ values for these countries to represent a hypothetical "less developed country." To calculate the bioavailable zinc portion (TAZ; usually a small fraction of TDZ) we used the Miller equation (Equation 1: Miller et al., 2007).

Equation 1 : Miller equation
Mean TDZ and TDP values were converted to mg mmol −1 and put into the Miller equation to compute the average per capita TAZ in our hypothetical LDC. The variables TDZ, TDP, and TAZ are described above, and A max , K P , and K R are constants as described in Miller et al. (2007).
We made an assumption that our hypothetical LDC receives half of its phytate and half of its zinc from wheat, which is roughly consistent with many of the LDCs in the FAO database. We analyzed the effect of elevated carbon dioxide levels on TDP, TDZ, and TAZ concentrations in a hypothetical LDC population for both NH + 4 and NO − 3 -supplied wheat. To calculate a new TAZ for wheat grown under elevated CO 2 conditions, we first calculated the percent change in TAZ from ambient to elevated levels for wheat receiving NH + 4 or NO − 3 . This computed percent change was then applied to half of the hypothetical TDZ and TDP; meanwhile, the other half of the hypothetical TDZ and TDP remained unmodified. Thus, the total new TDP and TDZ is the sum of the unmodified and modified portions. These new TDP and TDZ values for both NH + 4 and NO − 3 -supplied wheat were then put into the Miller equation to compute new hypothetical TAZ values for an LDC. Differences and corresponding percent changes between the new TAZ values and the original TAZ value for a LDC were computed to determine the overall affect of elevated CO 2 on TAZ in NH + 4 and NO − 3 -supplied wheat for an average developing world population. TAZ, TDP, and TDZ concentrations can only be compared within a single N form across the CO 2 concentrations due to methodological constraints of the model.

RESULTS
We divide the results here into three categories: first, biomass and yield data for the shoots, grain, and roots; second, tissue concentrations and whole plant micro-and macronutrient contents; and third, nutrient distribution among the different tissues. Values of the statistical significance of the results were place into a table (Table 1) in order to improve the readability of the text.

BIOMASS AND YIELD
Plants supplied NH + 4 vs. NO − 3 nutrition reacted differently to CO 2 enrichment (Figure 1; Table 1). Plants supplied NH + 4 differed across CO 2 treatments for most of the yield and biomass measurements. The greatest values typically were found at ambient CO 2 concentrations. Shoot, chaff, grain yield, number of heads, and KN were greatest at ambient CO 2 levels. Individual kernel mass was greatest under both ambient and elevated CO 2 treatments. HI and kernels head −1 showed no change across CO 2 treatments. In contrast, biomass and yield measures of NO − 3 -supplied plants did not differ among the three CO 2 concentrations.
At subambient CO 2 , differences between the NH + 4 and NO − 3 treatments occurred in shoot biomass and three of the yield components: kernel mass, head number, and kernels head −1 . Ammonium-supplied plants had a larger number of heads while NO − 3 -supplied plants had greater shoot biomass, kernel mass, and kernels head −1 . At ambient CO 2 , NH + 4 -supplied plants had a greater number of heads and greater chaff biomass. Plants supplied NO − 3 had a larger number of kernels head −1 . At elevated CO 2 , biomass and yield measures did not differ with N treatment.

ROOT
Roots had a smaller mean biomass when supplied NH + 4 than when supplied NO − 3 at all CO 2 concentrations (Figure 1). Both N treatments had the greatest biomass at ambient CO 2 , and the smallest at subambient CO 2 . The highest root:shoot ratios for both NH + 4 and NO − 3 -supplied plants were observed at ambient and elevated CO 2 . Ammonium-supplied plants always had lower root:shoot ratios and biomasses than NO − 3 -supplied plants at the same CO 2 concentration.

Total plant nutrients
Total plant nutrients generally followed the same trend within N form, although NH + 4 -supplied plants exhibited a greater diversity of responses to increasing CO 2 concentrations ( Table 2). Total plant P, K, B, Ca, Mg, and Zn decreased with increasing CO 2 under NH + 4 , while S and Mn were highest under ambient CO 2 .
www.frontiersin.org  Figure 2). Calcium and Cu were highest under subambient CO 2 . Total N and S were greatest at subambient and elevated CO 2 . Nitrate-N was greatest at ambient CO 2 . Phosphorus was highest at elevated CO 2 concentrations. Boron, Zn, and Mn increased with CO 2 concentration.
Plants supplied NO − 3 showed significant variation across CO 2 treatments for K, Ca, Mg, B, Fe, Cu, Zn, and NO − 3 − N (Table 1; Figure 2). Calcium and Cu had the greatest concentrations at subambient CO 2 . The highest concentrations of B, Fe, and Zn occurred at subambient and elevated CO 2 . Potassium concentrations were highest at elevated CO 2 . Nitrate-N increased with CO 2 . Magnesium showed the opposite trend, decreasing with CO 2 concentration.
Differences between N forms were also evident.  Figure 3). The greatest concentrations of total N, P, K, Ca, and Cu were found at subambient CO 2 . Iron concentrations were high at both subambient and ambient CO 2 . Boron was equally high at subambient and elevated CO 2 . Manganese was greatest at elevated CO 2 . Nitrate-N decreased with increasing CO 2 .
Significant differences among the NO − 3 -supplied plants across CO 2 treatments were only observed in S and B. The greatest concentrations of B were found at subambient CO 2 . Sulfur was highest at ambient CO 2 .
Nitrogen form significantly affected grain nutrient concentrations (   at subambient CO 2 for NH + 4 -supplied plants (Figure 4). Subambient CO 2 also produced the lowest phytate concentrations in NO − 3 -supplied plants. NH + 4 -supplied plants had greater phytate concentrations than NO − 3 -supplied plants at subambient CO 2 , but not at the other CO 2 concentrations. Grain from plants grown under NH + 4 nutrition had roughly 7, 18, and 8% higher bioavailable Zn than NO − 3 -supplied plants at subambient, ambient, and elevated CO 2 , respectively (Figure 4).
Based on this phytate and bioavailable Zn data, we modeled how a human population from a LDC would be affected by www.frontiersin.org changes in atmospheric CO 2 concentrations ( Table 3). The calculations were based on differences among CO 2 concentrations; therefore, modeled TDZ, TDP, and TAZ values cannot be compared between NH + 4 and NO − 3 -supplied grain. Grain from plants supplied the different N forms behaved differently as CO 2 concentration increased. We found that under NH + 4 supply, TAZ would increase 3.6% with the rise in CO 2 from subambient to ambient, and decrease 1.6% with the rise from ambient to elevated CO 2 (Figure 4). Humans provided NO − 3 -supplied wheat would experience a decrease in TAZ of 3.5% going from subambient to ambient, and an increase 5.6% from ambient to elevated CO 2 (Figure 4).

Roots
Ammonium-supplied plants generally showed a trend toward decreasing nutrient concentrations with increasing CO 2 concentration while NO − 3 -supplied plants varied widely across CO 2 treatments ( Figure 5). The decrease in nutrient concentrations under NH + 4 supply corresponded to an increase in root mass. Nitratesupplied plants tended to have their highest nutrient concentrations in the ambient and elevated CO 2 treatments. Ammoniumsupplied plants had higher concentrations of Zn and Mn across all of the CO 2 treatments, as well as higher total N and Fe at subambient CO 2 . Nitrate-supplied plants typically had higher concentrations of the other nutrients at all CO 2 concentrations.

Distribution of nutrients
The distribution of nutrients and micronutrients among plant parts followed similar patterns in both the NH + 4 and NO − 3supplied plants, although the NH + 4 -supplied plant distributions were slightly more variable ( Table 4). Allocations to root and grain usually were greatest at ambient CO 2 , and those to chaff and shoots at either subambient or elevated CO 2 . Grain typically contained the largest proportion of total N, P, Zn, and Cu, although the organ with the largest percentage of Cu varied with CO 2 treatment among NO − 3 -supplied plants. Plants at subambient and elevated CO 2 allocated more Cu to the grain, while those at ambient CO 2 allocated more to the roots. In general shoots received the majority of K, S, B, Ca, and Mg for all N and CO 2 treatments. Ammonium-supplied plants allocated slightly more Mn to the roots at subambient CO 2 , but allocated increasing amounts to the shoots at the expense of the roots as CO 2 concentration increased. In contrast, NO − 3 -supplied plants allocated most of the Mn to the shoots. Ammonium-supplied plants typically allocated more resources to the chaff while NO − 3 -supplied plants allocated a greater percentage of elements to the roots.

DISCUSSION
No other study to our knowledge has examined the influence of N form (NH + 4 vs. NO − 3 ) on plant nutrient relations at three different atmospheric CO 2 concentrations. Overall, N form affected growth, total plant nutrient contents, and nutrient distribution in senescing wheat shoots, grain, and roots. The influence of NH + 4 and NO − 3 on growth and nutrient status were so distinct that they should be treated as separate nutrients and not bundled into a general category of N nutrition. Wheat size and nutrition at senescence responded to CO 2 concentration in a non-linear manner. As was previously shown (Bloom et al., 2012), we found that plants supplied with NH + 4 were more responsive to CO 2 concentration than those supplied with NO − 3 . Although not explicitly addressed here because of the heterogeneity of variances, interactions between CO 2 and N treatments likely existed for a number of the biomass and nutrient measures. Most nutrient concentrations were generally higher in NH + 4supplied plants, with the exceptions of NO − 3 − N , Mg, B, and Mn, which were generally higher in NO − 3 -supplied plants. Phytate, which hinders human absorption of Zn and Fe (Raboy, 2009), showed little variation at ambient and elevated CO 2 between NH + 4 and NO − 3 -supplied plants, which, in conjunction with the observed greater bioavailable of Zn in NH + 4 -supplied plants, may have consequences for human nutrition. Distribution of nutrients to the shoots, roots, chaff, and grain in response to CO 2 concentration and N form was also non-linear and varied by nutrient.

BIOMASS AND YIELD
The data support our hypothesis that NO − 3 -supplied plants would show a more limited biomass and yield enhancement with CO 2 enrichment than NH + 4 -supplied plants. Nevertheless, mean biomass and yield decreased from ambient to elevated CO 2 in both NO − 3 -and NH + 4 -supplied plants in contrast to biomass increases in prior work on wheat seedlings (Bloom et al., 2002). NO − 3supplied plants allocated more biomass to roots and had larger root:shoot ratios than NH + 4 -supplied plants regardless of CO 2 concentrations as has been reported previously (Wang and Below,    Table 1 for statistical significance).  Table 1 for statistical significance). 1995; Bloom et al., 2002), but increased root mass at elevated CO 2 concentration for NO − 3 -supplied plants reported previously (Bloom et al., 2002) were not observed here. The shoot biomass data suggest that growth differences measured early in the lifespan of wheat supplied with NH + 4 or NO − 3 or NH + 4 (i.e., greater shoot biomass in plants supplied NH + 4 relative to those supplied NO − 3 at elevated CO 2 concentrations; Bloom et al., 2002) do not necessarily carry through to senescence. This may be due in part to a shift in NO − 3 assimilation to the root (Kruse et al., 2003), allowing NO − 3 -supplied plants to compensate for the decrease in shoot NO − 3 assimilation that occurs at elevated atmospheric CO 2 concentrations (Bloom et al., 2002(Bloom et al., , 2010(Bloom et al., , 2012.

Frontiers in Plant Science | Plant Nutrition
The decrease in yield and biomass measures at elevated CO 2 concentrations does not agree with field observations where wheat yields as well as overall biomass increased with elevated CO 2 (Högy and Fangmeier, 2008;Taub et al., 2008). Similarly, our results that the greatest values for other yield measures (e.g., heads, kernel mass, KN) occurred at ambient CO 2 concentrations varies from the literature. High CO 2 has been found to increase flowering tillers (Havelka et al., 1984;Fangmeier et al., 1996), KN (McKee et al., 1997), and kernel mass (i.e., thousand grain weight; McKee  , 1997). Conflicting results, however, have also been reported (e.g., Havelka et al., 1984). Many of the field and open top chamber studies were grown under natural light and thus received substantially greater photosynthetic flux density than our chamber-grown plants. These higher light conditions would be more favorable to biomass accumulation. Also, these studies typically applied high amounts of mixed N fertilizer (e.g., NH 4 NO 3 ), and yields and biomass have been found to be greater under mixed N nutrition than under either NH + 4 or NO − 3 alone (Cox and Reisenauer, 1973;Gentry et al., 1989;Heberer and Below, 1989;Wang and Below, 1995). Finally, the wheat cultivar we used (T. aestivum cv. Veery 10) is a short-statured variety that has rarely been used in other studies and may have accounted for some of the differences between our study and other published data.
Our results that NH + 4 -supplied plants had greater yield and yield components than NO − 3 -supplied plants at ambient CO 2 have been observed previously (Wang and Below, 1996;Chen et al., 1998). Wang and Below (1995) observed greater numbers of kernels head −1 and KN in plants supplied NO − 3 that was not observed here. Their study, however, supplied NH + 4 at relatively high levels (∼8.9 vs. 0.2 mM NH + 4 − N in our study). Several studies (Bennett and Adams, 1970;Cox and Reisenauer, 1973) have found that incipient NH + 4 toxicity can start appearing at N levels as low as 0.08-0.2 mM NH + 4 , although the onset of NH + 4 toxicity depends on light level (Magalhaes and Wilcox, 1984;Britto and Kronzucker, 2002) and solution pH (Findenegg, 1987). The poorer performance of the NH + 4 treatment in Wang and Below (1995), therefore, might derive from NH + 4 toxicity. We have previously determined that the 0.2 mM NH + 4 -supplied to our plants to be sufficiently high for normal growth, but low enough to avoid toxicity problems under our experimental conditions (Bloom et al., 2002).

PLANT NUTRIENTS
Our second hypothesis, that nutrient concentrations are differentially affected by the inorganic N form supplied to the plants and CO 2 enrichment, was supported by our data. CO 2 concentration www.frontiersin.org   and N form interactions may alter tissue demands for nutrients. For many nutrients, ratios between different elements are typically maintained within a narrow range (Garten, 1976;Bloom et al., 1985;Loladze, 2002). CO 2 concentration and N form may disturb the balance between different nutrients, leading to a cascade of changes in demand, accumulation, and allocation among the different plant tissues (e.g., Loladze, 2002;Högy and Fangmeier, 2008;Natali et al., 2009). Nitrate-supplied plants accumulated the greatest amounts of nutrients at ambient CO 2 ( Table 2). Some portion of the greater response of NH + 4 -supplied plants to CO 2 derived from a dilution effect from the greater biomass at ambient CO 2 concentrations (Figures 2 and 3). Total amounts of nutrients tended to decline with CO 2 enrichment for NH + 4 -supplied plants, which had the greatest amounts of macro/micronutrients at subambient CO 2 ( Table 2). These results have not been observed in other published studies (e.g., Fangmeier et al., 1997;Wu et al., 2004). Growth chamber studies, however, tend to have more exaggerated differences among treatments than field and greenhouse experiments (Högy and Fangmeier, 2008), and N source cannot be well-controlled in field and greenhouse experiments.
The observed increase in NO − 3 −N concentration with CO 2 concentration in NO − 3 -supplied plants has been reported previously (Bloom et al., 2002), and adds further support to the hypothesis that elevated CO 2 concentrations and the resulting decrease in photorespiration inhibit shoot NO − 3 photoassimilation. Nevertheless, tissue NO − 3 − N concentrations observed here were substantially lower than those in the earlier study (Bloom et al., 2002). Again, this may derive from difference in life stages in the two studies. Most of the N available to the plant for grain filling comes from N translocation rather than uptake from the substrate (Simpson et al., 1983). Probably, the plants continued to assimilate plant NO − 3 using a non-photorespiratory dependent process such as root assimilation after root N uptake slowed or stopped. Loss of NO − 3 through root efflux to the nutrient solution also may have contributed to the lower concentration of NO − 3 − N . The partitioning and accumulation of all mineral elements was affected in some manner by the CO 2 treatment and N form supplied to the plants. Observations that cation concentrations decrease under NH + 4 supply (e.g., Cox and Reisenauer, 1973;Gashaw and Mugwira, 1981;Wang and Below, 1998) relative to NO − 3 supply were not apparent in this study. Again, this could be partly due to the relatively low concentration of NH + 4 -supplied in our study, the age of the plants at harvest, and differences among wheat cultivars.
Allocation of nutrients within the plant followed similar trends for both N forms, with the exceptions of Mn and Cu (Table 2). Interestingly, in NO − 3 -supplied plants, shoot Mn concentrations increased slightly with CO 2 , and these plants allocated far more Mn to the shoots than NH + 4 -supplied plants at all CO 2 concentrations. Manganese (Mn 2+ ) has been found to activate Rubisco in place of Mg 2+ and the Rubisco-Mn complex has been observed to decrease Rubisco carboxylase activity while minimally affecting or even enhancing oxygenase activity (Jordan and Ogren, 1983). The slight increase in shoot Mn with CO 2 corresponded to a large 23% decrease in Mg concentration. Manganese, which can act as a cofactor for glutamine synthetase (Smirnoff and Stewart, 1987), was also the only nutrient that NH + 4 -supplied plants allocated a www.frontiersin.org greater percentage to the roots at the expense of the shoots. NO − 3supplied plants typically allocated a higher percentage of most nutrients to the roots, as has been reported previously Below, 1995, 1998).
Phytate, which forms complexes with divalent cations, has been found to hinder human Zn and Fe absorption during digestion and thus has been labeled an "anti-nutrient." It may serve a number of valuable functions, however, including roles as an anti-oxidant and anti-cancer agent (Raboy, 2009). Phytate is also the major repository of grain P, and variation in P supply to the developing seed is the major determinant of net seed phytate accumulation (Raboy, 1997(Raboy, , 2009Cakmak et al., 2010). To our knowledge, no published studies have explicitly looked at how phytate is affected by CO 2 concentration. Elevated CO 2 has been found to have a much larger negative impact on Zn and Fe concentrations than on P in wheat (Loladze, 2002;Cakmak et al., 2010). Several studies (e.g., Fangmeier et al., 1999;Högy and Fangmeier, 2008) have observed that P increases slightly with CO 2 concentration, and because the majority of P is tied up in phytate, this may cause increases in grain phytate concentrations as atmospheric CO 2 rises. As a result, bioavailable Zn and Fe-Zn and Fe not bound to phytate -is expected to decrease even further (Loladze, 2002).
Nonetheless, we did not observe such trends in macro-and micronutrient concentrations in this study. The mechanism behind these contrasting results is not clear, although the environmental conditions and nutrient solution in which the plants were grown likely had some role. The modeled data demonstrated only a small negative impact of CO 2 concentration on bioavailable Zn concentrations (Table 4), which was unexpected. Indeed, the grain from NO − 3 -supplied plants actually showed a slight increase in bioavailable Zn between ambient and elevated CO 2 . These results combined with the differences in grain bioavailable Zn between NH + 4 and NO − 3 -supplied plants demonstrates that N form may differentially affect the nutritional status of this important nutrient, especially in less developed countries that might be more dependent on phytate-rich grains for their Zn nutrition ( Table 3). The milling process removes some, if not most, of the phytate and grain mineral content with the bran fraction of the grain (Guttieri et al., 2006). Regardless, with over 50% of the human population suffering from Zn deficiencies, even small increases in bioavailable Zn would be beneficial (Loladze, 2002). This modeling exercise, however, is not a prediction of how increasing CO 2 will affect wheat nutrition so much as illustrates that N source may mediate, to some extent, the effects of CO 2 on phytate and bioavailable Zn, and that N source will become an even more important agricultural consideration in the future.
In summary, both CO 2 concentration and N form strongly affect biomass and yield in hydroponically grown wheat, as well as nutrient concentrations in above-and belowground tissues. Interactions among plant nutrient concentrations, CO 2 concentrations, and N form are complex and non-linear. The impact of N form and CO 2 concentration on the mechanisms affecting nutrient accumulation and distribution requires further research and extension to more realistic and agriculturally relevant growing conditions found in greenhouse and field studies. Of course, in greenhouse and field studies, control of N source is limited and control of atmospheric CO 2 concentration is expensive. The effects of CO 2 and N form on agriculture and human nutrition observed here are interesting and suggest a new area of research on mitigating the effects of climate change on agriculture. The supply of fertilizers (e.g., urea, NH 4 NO 3 , anhydrous NH 3 , organic amendments) or addition of nitrification inhibitors that increase the amount of available NH + 4 may have beneficial effects for human nutrition, particularly in regards to micronutrient deficiencies such as Zn and Fe that currently affect billions of people worldwide. In the face of the potentially negative consequences of climate change on agriculture, all avenues of mitigation must be examined, and even small improvements may prove worthwhile.