|Home | About | Journals | Submit | Contact Us | Français|
Nitrogen-use efficiency (NUE) of cereals needs to be improved by nitrogen (N) management, traditional plant breeding methods and/or biotechnology, while maintaining or, optimally, increasing crop yields. The aims of this study were to compare spring-barley genotypes grown on different nitrogen levels in field and growth-chamber conditions to determine the effects on N uptake (NUpE) and N utilization efficiency (NUtE) and ultimately, NUE.
Morphological characteristics, seed yield and metabolite levels of 12 spring barley (Hordeum vulgare) genotypes were compared when grown at high and low nitrogen levels in field conditions during the 2007 and 2008 Canadian growing seasons, and in potted and hydroponic growth-chamber conditions. Genotypic NUpE, NUtE and NUE were calculated and compared between field and growth-chamber environments.
Growth chamber and field tests generally showed consistent NUE characteristics. In the field, Vivar, Excel and Ponoka, showed high NUE phenotypes across years and N levels. Vivar also had high NUE in growth-chamber trials, showing NUE across complex to simplistic growth environments. With the high NUE genotypes grown at low N in the field, NUtE predominates over NUpE. N metabolism-associated amino acid levels were different between roots (elevated glutamine) and shoots (elevated glutamate and alanine) of hydroponically grown genotypes. In field trials, metabolite levels were different between Kasota grown at high N (elevated glutamine) and Kasota at low N plus Vivar at either N condition.
Determining which trait(s) or gene(s) to target to improve barley NUE is important and can be facilitated using simplified growth approaches to help determine the NUE phenotype of various genotypes. The genotypes studied showed similar growth and NUE characteristics across field and growth-chamber tests demonstrating that simplified, low-variable growth environments can help pinpoint genetic targets for improving spring barley NUE.
Nitrogen (N) is an essential element for both crop development and yield. However, since the green revolution, 50 years ago, farmers tend to maximize N fertilization to maximize crop yield (Hirel et al., 2007). As the need for more food production increases, the global consumption of synthetic (commercial) and organic (manure) has increased at an even greater rate (Tilman et al., 2001). Although globally there has been an increase in N fertilizer use, this is not uniform in all countries or geographic areas (Vitousek et al., 2009). Some areas, such as North America, have steady increases in N fertilizer use, China is showing a substantial increase in use, while Denmark has a reduction in use and Africa has a chronic under-use of N fertilizers (Fixen and West, 2002; Mosier et al., 2004; Liu et al., 2008; Ju et al., 2009). There are several costs associated with both over and under-use of N fertilizers that are borne by all. Over-use of N fertilizers is economically costly to the low-value crop producers in terms of both the price of the N-fertilizer and the potential loss of yield, as reported by Ju et al. (2009). High N-fertilizer consumption is also environmentally damaging, with excess N lost by leaching into groundwater and runoff into surface water, plus ammonia volatization and production of NOx gases from denitrification polluting the atmosphere (Kaye and Hart, 1997; Galloway et al., 2008; Gruber and Galloway, 2008; Conley et al., 2009). Under-use of N fertilizers is costly to subsistence farmers who rely on their crops to yield enough food to feed their family nutritiously. Therefore, some areas of the world need to focus on reducing N fertilizer use, while other areas need to have greater access to N fertilizers.
Coupled to the rate of N fertilizer application is the inherent low nitrogen-use efficiency (NUE) of cereal crops. NUE of cereals is estimated to be 42 % and 29 % in developed and undeveloped nations, respectively (Pilbeam, 1996; Raun and Johnson, 1999; Hodge et al., 2000). There is a need to increase the NUE of cereal crops by either N management strategies, traditional plant breeding methods or biotechnology, while at least maintaining, or optimally, increasing crop productivity. Increasing cereal crop NUE would be economically beneficial to the low-value crop producer and environmentally beneficial to the world population.
NUE is a complex genetic trait comprising N uptake and N utilization. N uptake efficiency (NUpE) is the ability of the plant to take up N from the soil. N utilization efficiency (NUtE) is the ability of the plant to assimilate and remobilize the N taken up from the soil, producing amino acids to be used as N carriers or signalling and regulatory pathway components and ultimately to produce grain (Moll et al., 1982; Good et al., 1994; Moose and Below, 2009).
Barley (Hordeum vulgare) is the fourth major cereal crop grown worldwide, and the third major crop grown in the Canadian prairies. It is used for malting, feedstock and human consumption of the grain. The rate of N fertilizer application is critical for cereal crops, but especially for malting barley since N affects tiller number, grain yield and quality (Anderson et al., 2002). Barley kernels must be within a strict protein concentration range to be acceptable for malting (McKenzie, 2008).
Recently Anbessa et al. (2009) analysed the genetic variability in NUE of many spring-barley genotypes grown under field conditions in two different studies, and found that both genotypic variability and environment can account for differences in NUE. One of the genotypes tested, Vivar, showed a superior NUE when grown on either high or low N-containing soils, with a concomitant decrease in yield on the low-N soils of <10 % (Anbessa et al., 2009).
The long-term goal is to improve the NUE of barley using either plant breeding or biotechnology coupled with best management practices such as the 4R nutrient stewardship proposed by the International Plant Nutrition Institute (Bruulsema et al., 2009). For researchers interested in improving NUE in crops, it is critical that they test the improved genotypes in field trials to examine the resulting phenotypes in a realistic environment. However, it is also important to develop basic tools to evaluate genotypes, such as the use of potted and hydroponic growth-chamber methodologies for growth and subsequent analysis of tissue (Moose and Below, 2009). The development and use of these basic tools/protocols is important for transgenic studies and research on the effects of varying N concentrations, since they reduce environmental variables. As well, effective laboratory-based growth methods allows researchers to make use of the many ways to evaluate an NUE phenotype and isolate tissue-specific genetic targets, such as by transcriptomic, metabolomic and/or proteomic analysis (Beatty et al., 2009). The phenotype of hydroponically grown plants can be relatively easily characterized at both the shoot and root level, whereas analysis of field-grown plants tends to be limited to above-ground biomass tissue, since analysis of the root tissue is technically difficult. Root-level analysis is important to fully understand NUE in cereals (Garnett et al., 2009). As well, one of the important rationales for comparing growth-chamber data and field data is to compare the growth characteristics and phenotypes of identical genotypes from simplistic to complex environments.
In this study, four barley genotypes were grown in hydroponic and potted growth chambers and ten barley genotypes were grown in field trials. Two of the genotypes, Vivar and Kasota, were common to both growth-chamber and field-trial conditions. This study complements that of Anbessa et al. (2009) by comparing, on a smaller scale, growth characteristics, yield, metabolite concentrations, N uptake and N utilization and the overall NUE of spring-barley genotypes, grown in various N conditions. Growing spring barley plants hydroponically allowed N uptake to be investigated, growing them in pots allowed NUE to be investigated in a growth-chamber setting, while growing them in the field allowed both N uptake and N utilization to be investigated as well as overall NUE of the genotypes.
A field study was conducted at Lacombe (52°26′N, 113°44′W), Alberta during the 2007 and 2008 crop seasons. Soil type of the site was Black Chernozem. Seasonal (May–August) precipitation was 316 mm in 2007 and 280 mm in 2008.
Ten spring barley cultivars released in the last two decades from different barley breeding programmes in North America were grown under low- and high-fertility regimes. The cultivars included malt and feed types as well as two-and six-rowed types (Table 1). The experiment was laid out in a randomized complete block design with three replications. Each cultivar was planted in double plots. Seeding rate was about 200 viable seeds m−2 based on pre-planting germination tests.
The soil for the Lacombe moderately low-N field had been depleted of N by growing cereal crops [winter or spring triticale (× Triticosecale Wittmack)] for 2 years previous to the barley. The high-N field was maintained by ploughing in pulse or alfalfa (Medicago sativa) plus applying fertilizer. In both years, soil samples were taken before seeding and N content was determined at Bodycote Norwest Laboratory, Bodycote International PLC (Edmonton, AB). The concentration of N was determined in soils extracted in modified Kelowna solution by continuous flow colorimetry (Carter, 1993).
Plant samples were taken from a 0·21-m2 (0·42 × 0·5 m) area from one of the double plots at physiological maturity. These plants were then separated into leaves, stem and head and dried to a constant weight. At maturity, the second of the double plots was harvested for yield. Grain yield was assessed after drying the seed.
Tissue samples from barley grown in field conditions for metabolite analysis were measured to 0·5 g and cut from the penultimate leaf when the plants were just prior to anthesis. The tissue was flash frozen in liquid nitrogen and subsequently kept at −80 °C until the metabolites were extracted.
A potted growth-chamber experiment was conducted on Vivar, Kasota, Morex and CDC-Copeland under varying N conditions. The barley seeds were planted directly into seeding trays filled with Sunshine mix #4 aggregate plus (Sun Gro Horticultural Canada, Seba Beach, Alberta, Canada), at a 2·5 cm depth.
Six replicate plants for each of the barley genotypes tested and for each nitrogen treatment were selected based on uniformity of leaf surface area and height and transplanted into pots containing Sunshine mix #4 in a growth chamber (21 °C day/18 °C night, 16 h/8 h light/dark photoperiod, 60 % humidity, photon flux density, 500 µmol m−2 s−1 photosynthetically active radiation). A basic fertilizer mix was developed and used as a constant source of phosphorus and potassium, with no added nitrogen (K2HPO4, KH2PO4, KCl, H3BO3, CuCl2·2H2O, Fe-EDTA, MnCl2·4H2O, Na2MoO4·2H2O, ZnCl2, Na2EDTA·2H2O). The nitrogen supplied was a chemically defined mixture of ammonium nitrate, potassium nitrate and urea that was made to four different percentages; 100, 75, 50 and 25 % total N, supplying 0·1308 g N, 0·0981 g N, 0·0654 g N and 0·0327 g N per plant per treatment, respectively. The plants were fertilized every 2 weeks, until anthesis, when each genotype had 50 % of the plants with a spike at the top of the boot. Anthesis was also the first harvest time, this was 54 d-after-sowing (DAS) for Morex, 67 DAS for Vivar and Kasota and 75 DAS for CDC-Copeland. At this point the flag leaf width and dry weight were measured. For dry weight the plant was cut at the base of the shoots and dried at 55 °C for 1 week before the weight measurement was taken. The plants were harvested a final time at maturity and the dry weight and total seed weight were measured.
A hydroponic growth-chamber experiment was conducted on Vivar, Kasota, Morex and CDC-Copeland under varying N conditions. Approximately 250–300 seeds from each of the four genotypes were sterilized with 50 % bleach followed by sterile distilled water, aerated for 24 h then dehusked. Two weeks after seed germination, plants were selected according to uniformity of biomass, height and health and transferred to polyethylene buckets containing modified Gries nutrient solution #1 (Gries et al., 1995), containing either 0·5, 2, 4 or 8 mm nitrate with four replicates per N condition and grown in a growth chamber (18 °C day/15 °C night, 16 h/8 h light/dark photoperiod, 60 % humidity, photon flux density, 500 µmol m−2 s−1 photosynthetically active radiation). The plants were supported in the buckets using a Plexiglas cover containing plant support slots that the tillers were inserted through, similar to the method used by Chaney et al. (1992). A polyurethane foam was used to cushion the tillers at the slot. The nitrate concentration and pH in the hydroponic solutions were monitored closely, and the pH was adjusted to 5·3. Two weeks after transfer, 28 d-after-germination (DAG), when plants were at the end of active tillering, the hydroponic solution was changed for fresh solution and the first harvest was conducted. A further 2 weeks from the time of the first harvest, 42 DAG, at a stage just prior to anthesis, the hydroponic solution was changed for the last time and the final harvest was conducted. Shoot and root dry-weight samples were taken at both harvest points (only data from the second harvest are shown). For dry weight the plant was separated into shoots and roots and dried at 55 °C for 1 week before the weight measurement was taken. Fresh weight samples of 0·3 g were taken from roots and shoots at the second harvest point and flash frozen in liquid nitrogen and subsequently kept at −80 °C until the metabolites were extracted.
Samples for metabolite analysis, weighing 0·3–0·5 g, were taken from field-grown (penultimate leaf; just prior to anthesis; four biological replicates) and hydroponically (penultimate leaf and actively growing root; prior to anthesis; three biological replicates) grown Vivar and Kasota. The plant samples were immediately flash frozen and stored at −80 °C, then ground in liquid nitrogen with a mortar and pestle when ready for extraction. Ground samples were transferred to an Eppendorf tube and weighed on an analytical balance. Methanol/chloroform/water (65 : 25 : 15) mixture was added to the ground sample at a ratio of 1 : 3 w/v, followed by 1 µL of internal standard amino acid solution (0·4 m Norvaline and Sarcosine in ddH2O). The sample was mixed with a vortexer three times for 10 s each, then centrifuged at 10 000 × g for 5 min at 4 °C. The supernate (500 µL) was collected into another Eppendorf tube and 100 µL chloroform followed by 150 µL water were added to the sample tube. The sample was mixed again and centrifuged at 10 000 × g for 5 min at 4 °C. The upper aqueous phase (400 µL) was collected and filtered through a Millex GX 0·22um disc filter (Millipore, Billerica, USA). The filtrate was then passed through a Biomax 5KNMWL membrane 0·5-mL filter (Millipore). The final filtrate was kept in −80 °C until metabolite analysis by HPLC.
All HPLC runs were performed on an Agilent 1200SL HPLC system (Agilent Technologies, Mississauga, Canada) equated with a Zorbax Eclipse Plus C18 3·0 × 100 mm 1·8-μm column (Agilent Technologies, Mississauga, Canada), with both a diode-array detector SL and a fluorescence detector. The instrument configuration was set up according to the published Agilent Technologies Application Note (Woodword et al., 2007) with the following exceptions: (a) a new gradient timetable was implemented since a 100-mm column was used; (b) the detection wavelengths for both diode array and fluorescence were switched at 10·9 min; and (c) the temperature of the column was set to 40 °C going in and 30 °C coming out.
All chromatography was processed in Agilent ChemStation software suite (Agilent Technologies). The area under the signal peaks was calibrated by the standard calibration curve and normalized by the internal standard Norvaline concentration, then all the concentration data were back calculated to the original concentration in the samples.
Dried above-ground biomass plant samples from the field trials were sent to BodyCote Northwest Laboratory for tissue N analysis. Dried shoot samples from the hydroponic trials were analysed for total Kjeldahl N using the improved colorimetric determination of Willis et al. (1996) with the following modifications: (a) 0·1 g of dried plant tissue was ground in 2 mL perchloric acid, with a pinch of sand, for 4 min, using a mortar and pestle; (b) the sample was centrifuged for 2 min at 11 000 × g and 100 µL 3 m potassium hydroxide added to the supernatant. This was incubated on ice for 30 min then centrifuged for 2 min at 11 000 × g. The supernatant was analysed for total Kjeldahl N as per the protocol.
The NUE calculations used in this study were developed by Moll et al. (1982) where NUE is defined as the grain yield (Gw) per unit of supplied N (Ns; Gw Ns−1), NUpE is the total above-ground plant biomass N (Nt) per unit of supplied N (Nt Ns−1) and the NUtE is the grain yield per total above-ground plant biomass N (Gw Nt−1).
For analyses of shoot dry weight, flag leaf width and grain yield of potted growth chamber-grown barley, the results are presented as average values from six individual plants with standard errors. For analysis of shoot, root and total dry weights, total N from shoots and metabolites of hydroponically grown barley, the results are presented as average values from three individual plants with standard errors. For analysis of grain yield and total N from above-ground biomass of field-grown barley the results are presented as average values from three individual plants with standard errors. Statistical analyses were performed using the ANOVA and Tukey's Honestly Significant Difference test functions of the SYSTAT Software, Inc. (Chicago, IL, USA). Principal component analysis (PCA) of the metabolite concentrations was performed using an online web server called MetaboAnalyst (Xia et al., 2009).
The field-grown spring barley grain yields (Table 2) showed that in 2007 the average yield was higher in the high-N field, while in 2008 the average yield was higher in the low-N trial. The overall grain yield trends were very similar to the plant-height trends (Fig. S1 in Supplementary data, available online) in that the low-N treatment in 2007 had the overall lowest grain yields for all genotypes.
The four genotypes showed some effects from being grown in N-limiting conditions (Fig. 1). Genotype (P ≈ 0·000) and N concentration (P ≤ 0·05) were significant factors contributing to shoot dry weight at anthesis, with genotype as most significant. All the genoptype pairings were significantly different in shoot weight (P ≤ 0·01) except for Vivar versus CDC-Copeland. Barley grown at 25 and 75 % N were significantly different (P ≤ 0·045) in shoot dry weight, with the higher N supply allowing for higher shoot dry weight.
The flag leaf width measurements at anthesis showed that genotype (P ≈ 0·000) and N concentration (P ≤ 0·01) were both significant factors contributing to this characteristic. All genotype pairings were significantly different (P ≤ 0·048) except for Morex versus Kasota. Barley grown at 25 versus 75 % N (P ≤ 0·027) and 25 versus 100 % N (P ≤ 0·001) were significantly different in flag leaf width. The plants supplied with higher N supply grew wider flag leaves, except for Morex which showed a slightly lower flag leaf width at 100 % N than at 75 % N, although not significantly so.
Genotype and N concentration were both equally significant factors (P ≈ 0·000 each) contributing to the shoot dry weight of the potted barley at maturity. All the genotypes versus CDC-Copeland were significantly different in shoot dry weight (P ≤ 0·001), while no other pairings showed any significant differences. Barley grown at 25 and 50 % N versus 100 % N (both at P ≈ 0·000) were significantly different in shoot dry weight. Like the shoot dry weight at anthesis, the higher N supply allowed for a larger shoot dry weight at maturity too. By maturity though, Morex, Vivar and Kasota were all very similar in shoot dry weight to each other.
Genotype (P ≈ 0·00) and N concentration (P ≤ 0·09) were both significant factors contributing to grain yield of the potted growth chamber-grown barley. As well, the different genotypes responded to differing N concentrations significantly for grain yield (P ≈ 0·00). All genotype pairings were significantly different (P ≤ 0·16) except for Morex versus Vivar. Barley grown at 50 % N versus 100 % N was significantly different in grain yield (P ≤ 0·05). Morex and Vivar both showed higher grain yields with higher N supply; however, Kasota and CDC-Copeland both had lower grain yields at 100 % N supply as compared with the other three, lower N supplies.
Figure 2 shows the shoot, root and total dry weight of four genotypes at varying N treatments at 42 DAG, just prior to anthesis.
Genotype was not a significant factor, while N concentration was a significant factor (P ≈ 0·000) contributing to shoot dry weight of the hydroponic barley. The shoot dry weight of barley grown at 0·5 mm N was significantly different from the rest of the dry weights (P ≈ 0·000), being much lower in weight for all genotypes than the barley grown at the higher N supply.
Genotype (P ≤ 0·048) and N concentration (P ≈ 0·000) were both significant factors contributing to root dry weight of hydroponic barley. Also, different genotypes respond to differing N concentrations significantly (P ≤ 0·004) for root dry weight. Kasota and CDC-Copeland were significantly different from each other (P ≤ 0·046), while all other pairings were not. Barley grown at 0·5 mm N was significantly different for root dry weight from the other three higher N supplies (P ≤ 0·002) while the root dry weights of barley grown at 2 mm N was also significantly different from the other root dry weights of barley grown at any other N supply (P ≤ 0·004). All genotypes showed suppressed root dry weights at the lowest N supply, and for the highest N supply, except for Kasota which had the largest root dry weight at 8 mm N. Although the root dry weights were very low for the 0·5 mm N supply, they were also much longer than the roots of the plants grown at the higher N levels (data not shown).
For the total dry weights of the hydroponically grown barley, genotype was not significant while N concentration was a significant factor (P ≈ 0·000). Barley grown at 0·5 mm N was significantly lower in total dry weight than the barley grown at 2, 4 or 8 mm N (P ≈ 0·000). There was no significant difference in total dry weight of the plants grown at the higher N supply.
Amino acid concentrations were measured in the shoots of Vivar and Kasota grown in field trials in 2008 at both low and high available N (Table 3). Six amino acids – asparagine (asn), histidine (his), arginine (arg), methionine (met), tryptophan (trp) and proline (pro) –increased significantly (P ≤ 0·09) in Kasota when grown on high N compared with low N. There were no significantly different amino acid concentration changes in Vivar grown on high to low N. Four amino acids – glutamate (glu), glycine (gly), threonine (thr) and pro –were measured at significantly (P ≤ 0·09) higher levels in Kasota grown at high N versus Vivar at high N. Six amino acids – glu, asn, glutamine (gln), thr, alanine (ala) and leu (leu) – were found to increase significantly in Kasota again, versus Vivar, when both genotypes were grown at low N.
PCA of these data retained 92·3 % of the variation in PC1 and 4·8 % in PC2, accounting for 97·1 % of the variation (Fig. 3A). The amino acid data grouped between genotypes when grown at high N only. One group comprised Kasota shoots grown at high N while the other group was Vivar at both N treatments and Kasota at low N. These first two principal components were significant (P < 0·0) using the permutation test 500 times. The loading plot for Fig. 3A (data not shown) showed that the gln concentration was much higher in the high N-grown Kasota than the rest of the analysed tissue. Pro, asn and thr concentrations were also elevated in Kasota shoots at high N, but to a lesser degree than gln.
Amino acid concentrations were measured in the roots and shoots of Vivar and Kasota grown in hydroponic growth-chamber conditions at 0·5, 4 and 8 mm nitrate (Table 4). Significant differences (P ≤ 0·05) in amino acid concentrations were determined for a variety of comparisons. Four amino acids – his, phenylalanine (phe), isoleucine (ile) and lysine (lys) – were all significantly reduced in concentration in the shoots of Kasota grown at 4 mm N versus 0·5 mm N, whereas in the roots, aspartate was the only amino acid significantly changed in Kasota grown at 0·5 versus 4 mm N, being found in increased concentration at the higher N supply. There was no significant change in amino acid concentrations in the shoots of Kasota grown at either 4 or 8 mm N, while, in the roots, thr and tyrosine (tyr) both significantly increased in Kasota grown at 8 mm N versus 4 mm N. Glu significantly increased in the shoots of Vivar grown at 4 mm N versus 0·5 mm N, while, in the roots, glu and ala both significantly increased in Vivar grown at 4 mm N versus the lower N supply. When Vivar and Kasota, both grown at 0·5 mm N, were compared there was an increase in phe in the shoots in Kasota, but no significant change in the root amino acid concentrations. When the amino acid concentrations are compared between 0·5 mm N-grown Kasota roots and shoots, six were significantly different – asp, glu, serine (ser), thr, ala and phe –all with higher concentrations in the shoot. There were no significant differences between 4 mm N-grown Kasota roots and shoots. When Kasota was grown on 8 mm N there were again six amino acids with significantly different concentrations – glu, gly, ala, tyr, trp, and leu – but this time gly, ala, trp and leu were higher in concentration in the roots than the shoots. Comparison of the roots and shoots of 0·5 mm N-grown Vivar showed five significantly different concentrations of amino acids – glu, thr, ala, trp, and leu – with the last two higher in the roots than the shoots. At 4 mm N, Vivar had significantly different concentrations of glu, gln and lys in the roots and the shoots, with gln and lys higher in the roots than the shoots. At 8 mm N, asp, tyr, ile and leu were found in significantly different concentrations between the Vivar roots and shoots, asp being higher in the shoots and the rest higher in the roots.
PCA of these data retained 77·5 % of the variation in PC 1 and 12·4 % in PC 2, accounting for 89·9 % of the variation (Fig. 3B). The amino acid data grouped mainly between tissue type and moderately by N treatment, but not according to genotype. One of the two main groups comprised shoots from both genotypes while the other group was roots from both genotypes. For the secondary groups, one was of both genotypes and tissue types grown with 0·5 mm nitrate, while the other was both genotypes and tissue types grown at 4 and 8 mm nitrate. These first two principal components were significant (P < 0·0) using the permutation test 500 times. The loading plot for Fig. 3B (data not shown) showed that the gln concentration was much higher in the roots, while glu and ala concentrations were much higher in the shoots. Asn and ser were also elevated mainly in shoots, while thre was elevated in either shoots or roots, independent of either N treatment or genotype.
The NUpE, NUtE and NUE were calculated for the ten barley genotypes grown in field conditions at low and high N for 2007 and 2008 (Table 2). NUE was calculated for the four barley genotypes grown in potted growth-chamber conditions (Table 5). NUpE was calculated for the four barley genotypes grown hydroponically in growth-chamber conditions (Table 6).
In the low-N field conditions for both 2007 and 2008, the above-average NUE genotypes were also all above average in NUtE, but not NUpE, with the exception of Excel in 2007 and Sundre in 2008 (Table 2). This was not the case for the genotypes grown on high N – in this treatment some of the above-average NUE genotypes were also either above average for NUtE or NUpE, but not all. It was rare to find an above-average NUE genotype that was also above average for both NUpE and NUtE. Vivar grown with high N was the only genotype above average for all three N efficiency components for 2007 and 2008. In 2008, Sundre and Xena grown on high N, also showed above-average NUE, NUpE and NUtE values. For the rest of the above-average NUE genotypes, they were either also above average for NUpE or above average for NUtE (Table 2).
In general there were three genotypes, Vivar, Excel and Ponoka that were consistently above average for NUE at both N treatments over both growing seasons. There were no consistently below average genotypes for both N treatments in both years; however, Bentley was below average for NUE for both years at low-N conditions while Seebe was below average for both years at high N conditions.
The NUE genotypes in this study were not grouped according to spike head or malt/feed types.
The NUE of four genotypes grown in potted growth-chamber conditions were calculated (Table 5). CDC-Copeland, a two-row malt variety, had poor grain yields when grown at high N, resulting in poor NUE, although at the two lower N treatments the grain yield was improved. Conversely, Morex, a six-row feed variety, increased grain yield with increased available N. The top two NUE genotypes were Morex and Vivar, at all four N treatments, while Kasota and CDC-Copeland trailed behind in NUE, both faring the poorest at 100 % N.
The NUpE was calculated for the four genoptypes grown hydroponically at the four different N concentrations supplied (Table 6). All genotypes showed a decrease in NUpE as the N supply increased. The genotypes were also fairly consistent with each other as to the level of NUpE at each N supply, except for Vivar which showed a much higher NUpE at 0·5 mm N than the other three genotypes.
Many researchers have reported that the NUE of the modern barley genotypes has been improved over the older genotypes (Anbessa et al., 2009; Sylvester-Bradley and Kindred, 2009). In a recent review of UK-based cereal crops by Sylvester-Bradley and Kindred (2009) it appears spring-barley breeding has been more successful at increasing crop yields without increasing the optimal N application than wheat breeding. Furthermore, these researchers noted that, although both N capture (uptake) and N conversion (utilization) play a role in improving barley NUE, N capture plays the predominant role. From the limited data presented here it appears that the roles that N uptake and N utilization play in the NUE of the barley genotypes grown in field conditions depends on the level of N supplied. At low supplied N, NUtE plays the predominate role in the high NUE genotypes and at high supplied N, either N utilization or NUpE or both predominate in the high-NUE genotypes. The genotypes with the highest NUE values, grown at high N are the ones showing both N efficiency components as above average. This study correlates with another limited data study by Moll et al. (1982) calculating NUE and N efficiency components for maize. Those researchers also determined that at low supplied N, the genetic variation in NUE of maize was related to NUtE, while at high N genetic variation in NUE was due to a mix of N uptake and N utilization efficiencies. To improve NUE, N uptake or N utilization, or both components, need to become efficient (Moose and Below, 2009). It has been suggested that for the best NUE results, both N efficiency components should be targeted for improvement under low-N conditions (Moll et al., 1982). This present field study shows that modern barley breeding has produced NUE varieties that have only increased NUtE when grown in low-N conditions. A targeted transgenic approach may be required to improve these varieties further to become efficient in both N utilization and N uptake in low-N conditions.
Of the ten spring-barley genotypes tested in field conditions, Vivar, Excel and Ponoka, were consistently high in NUE across growth years and N levels. Vivar grown in potted growth-chamber trials also had high NUE; therefore, this genotype is consistently high in NUE across complex to simplistic growth environments.
All the hydroponically grown genotypes showed a decrease in NUpE as the N supply increased. This suggests that when barley is grown hydroponically there is less pressure on the plants to be efficient at N uptake when there is enough or abundant N supply at 4 mm and 8 mm N. Root architecture of the hydroponically grown barley was also seen to be affected by the available N, in that the dry weight was reduced in the low-N environment while the root length was increased. This is opposite to the N uptake efficiencies calculated from the field-grown barley; however, both field-grown barley and hydroponically grown barley were consistent in showing Vivar as the highest NUpE genotype of the tested barley.
The average grain yields of the spring barley grown in field conditions (Table 2) showed that in 2007 the average yield was larger in the high-N field; however, in 2008 the average yield was larger in the low-N trial. This effect is most likely to be due to the precipitation pattern and field topography, allowing for increased soil moisture in some areas and not in others (Table S1 in Supplementary data, available online).
Other field-based studies have shown differences in the NUE of barley genotypes (Sinebo et al., 2004; Abeledo et al., 2008; Anbessa et al., 2009). Anbessa et al. (2009) determined that NUE-superior spring-barley genotypes were not correlated to spike-type group, and the high NUE genotypes were Vivar and Xena, both modern varieties that showed highest grain yield per unit of available N under low-N conditions and averaged over many different soil types. Also no correlation was seen between high NUE and spike-type group in field conditions. In the potted growth-chamber trial, only one of the four genotypes tested – CDC-Copeland – was a two-row type. CDC-Copeland was sensitive to N treatment in growth-chamber conditions, faring poorer in terms of shoot and root biomass, grain yield and NUE than the six-row types at the four different N levels. There has been a report of yield differences between two- and six-row spike-type winter barley genotypes grown in field conditions, where six-row types out-yielded two-row types by 4–11 %, depending on the available N level (Le Gouis et al., 1999). A growth-chamber trial with a larger number of two -and six-row types would be required to determine the validity of a spike-type difference in NUE under those conditions.
The metabolite analysis illustrated the importance of examining genotypes undergoing NUE improvement, by either breeding or biotechnology, using a growth system such as hydroponics with fewer variables and with the ability to allow researchers to analyse roots as well as shoots easily. There was a clear concentration difference in some key N metabolism-associated amino acids between roots and shoots of the genotypes grown hydroponically, although there was little difference in metabolite concentrations according to genotype. The roots of both Vivar and Kasota were elevated in gln when grown at 4 mm and 8 mm N, while the shoots of both genotypes were elevated in glu and ala at all three N concentrations examined. Correlating this with NUpE calculations for the hydroponically grown genotypes suggests that when NUpE is high, at the lowest N supply the N is not only efficiently taken up by the roots but is also efficiently utilized in the roots and mobilized to the shoots. Once N is taken up by the plant it is very quickly assimilated into amino acids, which can take place in the root or shoot. Key N metabolism amino acids such as glm, glu, ala, asp and asn have been found to play roles as N carriers throughout tissues and be involved in signalling and regulatory pathways important to the plant N status and growth (Moose and Below, 2009). The metabolite analysis suggests that glu and ala are transported as N carriers to the shoots for further assimilation and possibly as signalling or regulatory components.
The field-grown samples were from shoots only and showed a difference between Kasota grown at high N from Kasota grown at low N and Vivar grown at either N condition based on glutamine concentrations. Glutamine, asn, his, arg and pro were highly elevated in Kasota at high N compared with Kasota grown at low N or to Vivar at either N condition. There may have been no difference in metabolite concentrations between Kasota and Vivar grown at low-N conditions in 2008 because both of these genotypes showed high NUE during that year, whereas Kasota grown at high N conditions was average for NUE in 2008. Interestingly, the calculated NUpE of Kasota grown on high N was higher than the average for that year and location while the NUtE of Kasota on high N was lower than the average, suggesting that the barley may have been efficient at uptake but not utilization, as reflected in the significantly higher amino acid concentrations still found in the shoot upon sampling. The amino acids still present in the shoot may represent lower utilization efficiency in mobilizing the amino acids into protein or some other form of storage. The metabolite concentrations may be indicative of an NUE phenotype; further analysis with larger data sets would be useful to perform to examine this possibility.
NUE of barley can be improved to allow for growth of barley on lower N soils with no reduction in grain yield per unit of available N. Improving the existing genotypes of barley can be done by identifying the genes and or traits that allow for adaptation to low N availability by either improved N uptake or N utilization or a combination of both. The high-NUE barley grown at low available N had high N utilization efficiencies but not NUpE. It may be possible to improve these genotypes further using a transgenic approach with a genetic construct known to improve NUpE of crop plants. Alanine aminotransferase (alaAT) has been over-expressed in canola and rice resulting in NUE transgenic plants that show elevated alaAT expression and high N uptake efficiencies (Shrawat et al., 2008; Beatty et al., 2009). Engineering a modern barley variety with a transgene such as alaAT, or another N uptake-related gene, may produce a genotype that could both utilize and take up N efficiently when grown with low available N. Determining other trait(s) or gene(s) to target to improve NUE in barley is important for this research to succeed. This goal can be facilitated by using simplified growth approaches that allow for the use of molecular tools such as transcriptomics, proteomics and metabolomics to help determine the genetic background of a particular phenotype.
In this study, the morphological and NUE components of a variety of spring-barley genotypes, grown at differing N levels in complex (field) to simplistic (growth-chamber) environments, were examined and compared. Despite the challenges of comparing field and laboratory data and the limited data set, the analyses suggested that these genotypes had similar growth and NUE characteristics across these environments. Using simplified, less-variable growth environments is another way to help pinpoint traits and genetic targets of interest for improving NUE in spring barley.
Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. Fig. S1: Morphological characteristics at plant maturity of ten spring-barley genotypes grown in field conditions at low and high supplied N for the 2007 and 2008. Table S1: Precipitation and mean air temperature data for the May to August growing seasons in 2007 and 2008 in Lacombe, Alberta, Canada.
We thank Cole Kushner and Rebecca Milkovich for their technical assistance in growing and measuring the spring-barley genotypes in the potted and hydroponic growth-chamber experiments. As well, we thank Yee Ying for the statistical analysis. This work was supported by the Natural Sciences and Engineering Research Council of Canada Discovery Grant (grant number G121210410); and the Alberta Crop Industry Development Fund (grant number C800000153).