HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gallais, A. Articles by Quilléré, I. Search for Related Content PubMed Articles by Gallais, A. Articles by Quilléré, I. Agricola Articles by Gallais, A. Articles by Quilléré, I. Related Collections Crop Physiology & Metabolism Nitrogen Crop Genetics

CROP BREEDING & GENETICS

Estimating the Proportion of Nitrogen Remobilization and of Postsilking Nitrogen Uptake Allocated to Maize Kernels by Nitrogen-15 Labeling

^a Station de Génétique Végétale, INRA-UPS-INAPG-CNRS, Ferme du Moulon, 91190 Gif/Yvette, France
^b Unité de Nutrition Azotée des Plantes, INRA route de St Cyr, 78026 Versailles Cedex, France
^c UMR INRA-USTL Stress abiotiques et différenciation des végétaux cultivés, Estrées-Mons, 80203 Péronne, France
^d Institut de Biotechnologie des Plantes, Université de Paris-Sud, 91405 Orsay Cedex, France

^* Corresponding author (gallais{at}moulon.inra.fr).

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Estimating the proportion of N remobilization and postsilking N uptake allocated to kernels may help to improve N-use efficiency in maize (Zea mays L.). In this study, we show theoretically and experimentally, that ¹⁵N labeling at the beginning of stem elongation can be used in the field to estimate the proportion of N remobilized from the vegetative parts to the kernels of maize by measuring ¹⁵N distribution only at maturity. In the same way, ¹⁵N labeling at silking allows a determination of the proportion of postsilking N uptake allocated to the kernels by measuring ¹⁵N distribution only at maturity. Two 1-yr experiments with three and four genotypes and one 2-yr experiment with testcross progenies from 66 recombinant inbred lines were developed. Nitrogen-15 labeling during vegetative growth provided an estimate of the proportion of N remobilized with a greater accuracy compared with the "balance" method, which computes the difference between the amount of N present in the stover at silking and the amount of N present in the stover at maturity. The validity of our ¹⁵N method is mainly based on the assumption that less than 15 to 20% of ¹⁵N is taken up after silking. The determination of the proportion of postsilking N uptake allocated to kernels requires assumptions that are more difficult to fulfill. Nitrogen-15 labeling in the field at the beginning of rapid stem growth appears to be a useful tool for studying the genetic variability of N remobilization using a large number of genotypes.

Abbreviations: NHI, nitrogen harvest index • RIL, recombinant inbred lines • RSA, relative specific allocation

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
IN CEREALS, during grain filling, amino acids from newly absorbed inorganic N and from leaf and stem protein remobilization are both translocated to the ear and further used for the synthesis of grain storage proteins. In maize (Zea mays L.), 35 to 65% of the grain N is remobilized from the stover depending on the environmental conditions and/or the genotype considered (Bertin and Gallais, 2000; Gallais and Coque, 2005). This also means that 35 to 65% of the grain N originates from postsilking N uptake from the soil. Remobilization and postsilking N uptake are thus essential components of N utilization efficiency. To develop cultivars exhibiting a better N-use efficiency, evaluating the contribution of N absorption and N remobilization to grain protein deposition should be useful (Beauchamp et al., 1976; Pollmer et al., 1979; Moll et al., 1982; Di Fonzo et al., 1993). For example, this could allow development of cultivars with a high N remobilization capacity for grain maize production at low N input. Cultivars maintaining N uptake and remobilization after silking may also prove useful for silage yields, in which protein content must be high both in the stover and in the grain (Gallais and Coque, 2005). However, it is still rather difficult to determine the two origins of grain N. The classical approach for evaluating the contribution of the two sources of grain N is based on a comparison of the total N budget of the whole plant at silking and of both the grain and the stover at maturity. Apparent N remobilization from the stover to the grain is then estimated by calculating the difference between the quantity of N in the stover at silking and the quantity of N in the stover at harvest (Moll et al., 1982; Di Fonzo et al., 1993; Rajcan and Tollenaar, 1999; Bertin and Gallais, 2000). Postsilking N uptake is estimated by calculating the difference between the quantity of N in the whole plant at harvest and the quantity of N in the whole plant at silking. This approach, called the balance method, leads to biased results for the proportion of remobilized N because the contribution of the roots is neglected and it is assumed that all the N taken up after silking is allocated to grain, which is not true (Gallais et al., 2006). Furthermore, due to plant sampling heterogeneity, the amount of N coming either from postsilking N remobilization or from postsilking N uptake is determined with poor precision because the two quantities of N that are used for the calculation, are themselves not accurately determined.

An alternative tool to the balance method is ¹⁵N labeling. Methods based on short- or long-term labeling have been developed using plants grown either in hydroponics or more rarely in the field. Hydroponics allows ¹⁵N application during a well-defined period, either short (with pulse-chase experiments) or long (with quasi-state labeling), whereas in the field only long-term labeling is possible. We will consider here only long-term labeling in the field. In maize, long-term labeling has already been used for studying the distribution among organs of the N taken up at various developmental stages of the plant (Friedrich and Schrader, 1979; Below et al., 1985; Cliquet et al., 1990a, 1990b; Deléens et al., 1994; Ma and Dwyer, 1998). In other ¹⁵N labeling studies, the proportion of remobilized N has been estimated by calculating the difference between the quantity of ¹⁵N accumulated in the stover at silking and the quantity of ¹⁵N accumulated in the stover at maturity (Mae and Ohira, 1981; Ta and Weiland, 1992). However, while this method is presumably more accurate than the balance method for calculating a total N budget, there is also a bias in the estimation of the total N budget due to the allocation to the stover of a significant part of the N taken up after silking. Moreover, this method requires two separate analyses, one at silking and one at maturity.

Using the ¹⁵N labeling, the objective of our study is to propose and evaluate a method that is easy to use for determining, from plants grown in the field, the proportion of N remobilized and the proportion of postsilking N uptake allocated to the kernels.

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Theory of ¹⁵N labeling
Consider a plant part x (stover, grain, or whole plant). Let q'_x be the quantity of N from the labeled fertilizer, q_x be the quantity of ¹⁵N from the labeled fertilizer, A_0x and A_x the ¹⁵N abundance without and after labeling and Q'_x the quantity of total N in the plant part x after labeling. The ¹⁵N abundance is defined as the ratio of the amount of ¹⁵N to the total amount of N (¹⁴N + ¹⁵N). Thus, we can write (Cliquet et al., 1990a, 1990b; Deléens et al., 1994)

[1]

A_fert being the ¹⁵N fertilizer abundance, and E_fert and E_x the isotopic excesses for the fertilizer and the plant part x, respectively.

The relative specific allocation (RSA) can be defined as the proportion of newly incorporated ¹⁵N atoms relative to total atoms in the sample (Cliquet et al., 1990a, 1990b),

[2]

which corresponds to the proportion of N in the organ that comes from the labeled fertilizer.

Labeling during Vegetative Growth
With this type of labeling, to predict the proportion of N remobilized from stover to the grain we assume that (i) the ¹⁵N fertilizer is distributed just at the beginning of quick growth, with labeling of only a small soil compartment, in such a way that a large part of the total ¹⁵N uptake is assimilated before silking; (ii) there is no discrimination between both isotopes: the ¹⁵N is distributed as ¹⁴N within the plant and both isotopes are remobilized in the same proportion; (iii) RSA is the same for all vegetative organs, or if not, the proportion of N remobilized toward the kernels is the same for each organ; and (iv) there are no N losses. Therefore, with such assumptions, at harvest, the amount of ¹⁵N originating from the N fertilizer present in the whole plant (q_wp) will be equal to the amount of ¹⁵N from the N fertilizer present at silking. Consequently, if q_G is the amount of ¹⁵N from N fertilizer present in the grain originating from the ¹⁵N already present in the stover at silking, the ratio q_G/q_wp directly expresses the proportion of N remobilized, t_rem, defined as the ratio of grain N originating from remobilization to total N at silking:

[3]

with q_silk being the amount of ¹⁵N originating from the N fertilizer absorbed by the plant at silking. It is then only necessary to determine the amount of ¹⁵N in the grain and the whole plant at maturity. With the mass spectrometer, only the ¹⁵N abundance can be quantified, but not the ¹⁵N content. If A_0x is the natural abundance in the plant part considered, ¹⁵N abundance and ¹⁵N amount are related by the following equations:

[4]

where Q'_G and Q'_wp are, respectively, the N amount in the grain and the whole plant after labeling. Thus,

[5]

NHI' being the nitrogen harvest index for labeled plants, NHI' = Q'_G/Q'_wp. If, as in our proposed method, the quantity of N from the labeled fertilizer in each organ is low relative to the total amount of N within this organ, then NHI' can be replaced by the NHI estimated with nonlabeled plants.

The main problem is to determine RSA_whole-plant and NHI, which are both defined at the whole-plant level, including roots. If we discard the contribution of the root system to the whole-plant N remobilization process, t_rem can be estimated by measuring only ¹⁵N and N distribution in the grain and in the stover at harvest. Previous studies (Cliquet et al., 1990b; Ta and Weiland, 1992) have shown that, after labeling during vegetative growth, ¹⁵N present in the roots at maturity contributes less than 5% to the amount of total ¹⁵N present in the whole plant. With this low root contribution, the bias in the prediction of t_rem is expected to be low.

If some residual ¹⁵N is still taken up from the soil after silking, the formula presented in Eq. [5] gives a biased estimate of the proportion for N remobilized. Indeed, in such a situation, let q_abs be the amount of residual ¹⁵N originating from the fertilizer N that is taken up after silking, q_Gabs the amount of ¹⁵N originating from postsilking ¹⁵N uptake allocated to the grain and q_rem the amount of ¹⁵N resulting from remobilization. The expression of total amount of ¹⁵N originating from fertilizer that is accumulated in the grain can be written as follows:

[6]

where q_rem is the ¹⁵N accumulated in the grain originating from the stover by remobilization, q_Gabs is the amount of ¹⁵N accumulated in the grain originating from the postsilking ¹⁵N uptake, q_abs is the total postsilking ¹⁵N uptake, and t_G is the proportion of postsilking assimilated N which is allocated to kernels (see below the case of ¹⁵N labeling at silking).

Therefore, if we ignore the presence of postsilking ¹⁵N uptake, an estimate of the proportion of remobilized N will be

[7]

where s = q_abs/q_wp is the proportion of the whole-plant ¹⁵N at maturity that is taken up after silking. When s = 0, there is no ¹⁵N residue after silking; however, in practice, there is always a certain amount of residual ¹⁵N. The use of Eq. [7] with different values of s and t_G leads to the conclusion that the overestimation of t_rem will generally be less than 0.05, even in the presence of 20% residual postsilking ¹⁵N uptake, when at least 15% of postflowering N uptake is allocated to the stover (Fig. 1 ). As an example, taking realistic values for the parameters as estimated in our experiments, with t_rem = 0.65, s = 0.24, and t_G = 0.84, t'_rem = 0.69 instead of 0.65. With s = 0.10, the overestimation is only 0.02.

View larger version (19K): [in this window] [in a new window]   Figure 1. The expected bias in the prediction of the proportion of remobilized N according to Eq. [7] in terms of the proportion of postsilking N uptake allocated to the grain and post silking 15N uptake proportion. Graphs correspond to a true remobilization proportion of 0.65 (continuous lines) and 0.69 (broken lines). For example, with a true remobilization proportion of 0.65, a proportion of postsilking N uptake allocated to the grain of 0.80 and a postsilking 15N uptake proportion of 0.20, the bias is just 0.03. With the same proportion of postsilking N uptake allocated to the grain and a postsilking 15N uptake of 0.30, the bias is between 0.04 and 0.05.

Using the relationship between ¹⁵N amount and ¹⁵N abundance we obtain a formula analogous to Eq. [5]:

[8]

Again, if the quantity of N from the labeled fertilizer in each organ is low relative to the total amount of N within this organ, then NHI' can be replaced by NHI estimated using nonlabeled plants. By neglecting the root contribution to RSA_whole-plant at maturity, t_G can be easily determined by the analysis of ¹⁵N abundance and total N in the grain and in the shoots at maturity.

With the simultaneous estimation of t_G and t_rem, it is possible to derive an unbiased estimation of t_rem, even in presence of residual postsilking N uptake. In this case, using the previous notation, we obtain

[9]

Field Experiments
Two preliminary ¹⁵N labeling experiments, with only three and four genotypes, have been conducted at Thiverval-Grignon (France) in 2001 to 2002. Protocols for ¹⁵N labeling in the field are those described by Gallais et al. (2006). In 2003 and 2004, an experiment was conducted using a larger number of genotypes.

2001 Experiment
In this experiment, a single-cross hybrid (Déa) and its parental lines F2 and Io were studied. Each genotype was grown in one plot of about 550 m², at 80000 plants ha^–1, in 80 cm–spaced rows with an application of 170 kg N ha^–1. Nitrogen-15–labeled nitrate was distributed at the beginning of stem elongation on three distant microplots, each with 12 consecutive plants in a row. Using a syringe, 1.25 mg of ¹⁵N was injected around each individual plant by applying 200 mL of a solution of KNO₃ at 1.94% ¹⁵N atom excess. To study the ¹⁵N distribution within the plant, labeled plants were harvested at silking and grain maturity from microplots, one for each stage. For a given harvest, six normally developed plants were selected from the 12 labeled plants of a microplot to have three "replicates" of two pooled plants. For each replicate, leaves and stalks + sheaths at silking and leaves, stalk + sheaths, husks + cobs, and kernels at maturity, were separated for further analyses of their dry matter content, total N content, and ¹⁵N abundance.

2002 Experiment
Four commercial hybrids (Déa, Anjou 285, Nicco, and Tarro) were grown at 100000 plants ha^–1 in 80 cm–spaced rows, with an application of 180 kg N ha^–1, in four replicates in a randomized block design (Pommel et al., 2005). Nitrogen-15–labeled nitrate was provided at two stages of plant development: at the beginning of stem elongation and at silking. For the labeling during vegetative growth in each replication, ¹⁵N-labeled nitrate was distributed on two distant microplots of four plants in a row; one microplot for sampling at silking and the other for sampling at maturity. Nitrogen-15 nitrate was applied according to the same protocol as in 2001. For the labeling at silking, 2.5 mg of ¹⁵N was provided to each individual plant via a syringe injection to the soil of 200 mL of a solution of KNO₃ at 4.91% ¹⁵N atom excess. For each sampling, two normally developed plants were selected from one labeled microplot in each replication. Leaves and stalks + sheaths at silking and leaves, stalks + sheaths, husks + cobs, and kernels at maturity were separated for the analysis of dry matter content, total N content, and ¹⁵N abundance.

In this paper, for both preliminary experiments, we have grouped together leaves and stalks + sheaths + cobs to have only two plant parts, stover and grain. For each experiment remobilization was also estimated by the balance method, from the sampling at silking and maturity.

2003–2004 Experiment
In this experiment, the maize plants were a set of recombinant inbred lines (RIL) from the cross of the two inbred lines F2 and Io (studied in 2001), and evaluated for combining ability with a common inbred line tester, as described by Bertin and Gallais (2000). Only the 66 RIL common to both years are considered in this paper. Plants were grown in 2003 and 2004 in the field at Gif sur Yvette (France) at about 90000 plants ha^–1 and with a N fertilization of 155 kg ha^–1. A randomized block design with three replicates and two-row plots (5 m long and 0.8 m between rows) was used. The two types of ¹⁵N labeling (i.e., during vegetative growth and at silking) were performed in both years. For labeling during vegetative growth in each two-row plot, for each replication, two microplots of 10 consecutive plants were utilized. A 1.5-L solution of KNO₃ at 4.91% ¹⁵N atom excess was sprayed on each microplot to distribute, on average, 1 mg ¹⁵N per plant. In 2004, the same protocol was applied, except that 10 d after spraying the ¹⁵N labeled KNO₃ solution, 3 L of water was applied to each microplot to facilitate ¹⁵N distribution within the soil and to have a chase effect. For the labeling at silking, only two replicates were analyzed, each with the same type of microplots as previously used, two microplots in 2003 and only one microplot in 2004. An application of 2.5 mg ¹⁵N was made per plant in a solution of KNO₃ at 4.91% ¹⁵N atom excess. In 2003, to facilitate ¹⁵N distribution within the soil, the ¹⁵N solution was sprayed at two dates, silking and silking + 10 d, with 1.5 L per microplot, at each date. In 2004, 3 L of the ¹⁵N solution was directly sprayed at silking.

At silking, for all treatments six to eight plants were sampled from each microplot. Furthermore, for labeling during vegetative growth, six to eight plants were selected at random from each plot. This sampling was required for the determination of postsilking remobilization and N uptake by the balance method and also for the determination of the proportion of postsilking ¹⁵N uptake. At maturity, six to eight plants were harvested per microplot and they were divided into two parts (stover = leaves + stalks + sheaths) and kernels. Cobs known to be very poor in N at maturity were not considered. Following each sampling at maturity, control plants were also selected using the same protocol as used for treated plants, to allow determination of ¹⁵N abundance in each organ in the absence of ¹⁵N fertilization.

Determination of Nitrogen Content and ¹⁵N Abundance
In all experiments, after drying and weighing of each plant part, the material was ground to obtain a homogeneous fine powder. A subsample of 2.5 mg was used to determine total N content and ¹⁵N abundance via an automated N analyzer (NA1500 Carlo-Erba, Milan, Italy) coupled to an isotope ratio mass spectrometer (Optima, Micromass, Manchester, UK) calibrated for measuring ¹⁵N natural abundance. Using the formula ¹⁵N = (¹⁵N/¹⁴N – 1)1000, from the ¹⁵N given by the mass spectrometer, the ¹⁵N abundance (A) was calculated to determine the ¹⁵N content of the sample. For each plant organ termed X, the atom percentage excess (E_X = A_X% – A_0X%) was then calculated, A_X% representing the ¹⁵N abundance percentage in the organ X considered and A_0X% representing the natural ¹⁵N abundance percentage of the same organ X. A_0X% was close to 0.36634%, a value that corresponds to the natural abundance of atmospheric dinitrogen (N₂).

From the dry matter amount and N content of each plant organ, the N amount was determined and the NHI defined as the ratio of N amount in the kernels to total N amount in the aerial part of the plant.

Statistical Analyses
ANOVA with fixed effects was applied to each studied trait in the 2001 and 2002 experiments. In the 2003–2004 experiment, genotype effect was considered as random because the set of studied genotypes was a random sample of an RIL population. The accuracy of estimation of a given trait was considered through the coefficient of variation. Furthermore, for the 2003–2004 experiment, broad-sense heritabilities (h²) were computed for the proportion of N remobilized estimated via the ¹⁵N method and by the balance method. To derive h² from the estimated variance components, we have considered the means of 2 yr with three replications per year. Regression analysis with the constraint of a zero intercept was also used to study the relationship between genotype means for RSA_grain and RSA_whole-plant or RSA_stover, because as shown by Gallais et al. (2006), a zero intercept was expected. For correlation studies, we considered phenotypic correlations between genotype means.

	RESULTS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Estimating Proportion of Remobilized Nitrogen
In the 2001 experiment, on average, the RSA_grain/RSA_whole-plant ratio was 0.96. This ratio was determined with a high accuracy, as shown by the very low CV of 0.8% (Table 1). The genotypic effect was also highly significant, with a RSA_grain/RSA_whole-plant ratio for F2 being significantly greater than that for Io and the Déa hybrid (Table 2). For the NHI the genotype effect was significant. The Déa hybrid had a greater NHI than its parents. This was mainly due to a higher harvest index (grain yield/shoot dry matter) for the hybrid (0.45) than for its parents (0.38 and 0.33). The predicted proportion of remobilized N was estimated with a better accuracy (CV = 5.5%) using ¹⁵N labeling than by the balance method (CV 16.1%). For this predicted remobilization, the genotypic effect was significant with a higher value (0.65) for the Déa hybrid than its parents (0.60 for F2 and 0.53 for Io). Such a ranking is clearly due to the difference in NHI. Due to a large experimental error, ¹⁵N uptake after silking showed no significant differences among the three genotypes with, on average, the same amount of ¹⁵N being present at maturity as at silking (Table 1).

View this table: [in this window] [in a new window]   Table 1. Means, coefficients of variation and F values for 2001 and 2002 experiments for 15N labeling during vegetative growth.

View this table: [in this window] [in a new window]   Table 2. Predicted and apparent remobilization values in the 2001 experiment.

For the 2003 and 2004 experiments, the relationships between RSA_grain and RSA_whole-plant were strong (the slope of the regression line with a zero intercept was b = 0.99 with a correlation coefficient r = 0.99 (P 0.001) in 2003, and b = 0.96 and r = 0.98 (P 0.001) in 2004, and b = 0.97 and r = 0.98 (P 0.001) for the pooled data of the 2 yr (Fig. 2 ). Taking the pooled data for the 2 yr, the slope of the regression line of RSA_grain onto RSA_stover with a zero intercept was b = 0.92 with r = 0.86 (P 0.001). On average, the RSA_grain/RSA_whole-plant ratio, which is close to the slope of the regression line RSA_grain onto RSA_whole-plant with a zero intercept, was 0.99 in 2003 and 0.96 in 2004 and 0.97 using the pooled data of the 2 yr. This ratio showed a better accuracy than the RSA_grain or RSA_whole-plant, without a significant genotype effect (Table 3). The NHI was estimated with good accuracy (CV = 4.4%) with a highly significant genotype effect. As a consequence, the predicted proportion of N remobilized after silking using ¹⁵N was determined with relatively good accuracy (CV = 6.2%) and a heritability of 0.52 (confidence interval: 0.28–0.68), whereas the proportion estimated by the balance method was determined with a lower accuracy (CV = 15.8%) and a heritability of 0.27 (confidence interval: 0–0.50). Furthermore, the correlation between the 2 yr for the estimates of the proportion of N remobilized using ¹⁵N was significant (r = 0.42, P 0.001), whereas it was not significant (r = 0.20) between estimates calculated using the balance method.

View larger version (12K): [in this window] [in a new window]   Figure 2. Relationship between RSAgrain and RSAwhole-plant according to the stage of 15N labeling (during vegetative growth, full squares, and at silking, open circles) with the pooled data from the 2003–2004 experiment (66 genotypes); b is the slope of the regression with zero intercept.

View this table: [in this window] [in a new window]   Table 3. Results from the 2003–2004 experiment for 15N labeling during vegetative growth.

Estimating the Proportion of Nitrogen Postsilking Uptake Allocated to Kernels
From the 2002 experiment, on average, the ratio RSA_grain/RSA_whole-plant = 1.23. Its CV was low (3.1%), but the genotype effect was not significant (Table 4). The proportion of postsilking N uptake allocated to the kernel (t_G) also showed no significant differences among genotypes with an average of 83%. This means that 17% of the N taken up after silking was allocated to the stover. This parameter was also estimated with a low CV (3.4%).

View this table: [in this window] [in a new window]   Table 4. Results from the 2002 experiment for 15N labeling at silking.

View this table: [in this window] [in a new window]   Table 5. Results from the 2003–2004 experiment for 15N labeling at silking.

	DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The correlations between the NHI and both the predicted proportion of N remobilized using ¹⁵N labeling and the proportion of N taken up after silking allocated to the kernels were highly significant and very similar. Such a high correlations were expected from Eq. [5] and [8], with a high accuracy and low variation in the RSA_grain/RSA_whole-plant ratio. Note that, with labeling during vegetative growth, in the absence of postsilking N uptake, NHI would be equal to the proportion of remobilized N. Then, in Eq. [5], the multiplication by the RSA_grain/RSA_whole-plant ratio can be interpreted as a correction to the NHI to estimate the proportion of N remobilized in the presence of postsilking N uptake. However, our results show that the effect of such a correction is low because the ratio RSA_grain/RSA_whole-plant is always close to 1 for every genotype (Fig. 2). This means that the proportion of N remobilized can be predicted by the NHI. With labeling at silking and in the absence of remobilization (all the grain N coming from uptake), NHI is equal to the proportion of N taken up after silking allocated to the grain. In the presence of remobilization, it is necessary to weight NHI by a coefficient which can be higher or lower than 1 to predict the true proportion of N taken up after silking that is allocated to the grain. Our results show that NHI alone is a better predictor of the proportion of N remobilized than of the proportion of N taken up after silking that is allocated to the kernels.

Assumptions underlying the prediction of both proportions are mainly relative to the ¹⁵N fluxes (Gallais et al., 2006). For determining the proportion of N remobilized with a labeling during vegetative phase, the main assumption concerns the proportion of residual postsilking ¹⁵N uptake, which must be less than about 15%. On average of the 4 yr of experiment, this proportion was 18.5%. Its variation between 0 and 32% was difficult to relate to soil and climatic conditions. As 2003 was dryer than 2004, the drought could have slowed down the diffusion of the ¹⁵N solution within the soil. Furthermore, in 2004, the spray of 3 L of water on each microplot 10 d after ¹⁵N application could have facilitated the ¹⁵N uptake before silking. However, in the 2003–2004 experiment, it was possible to derive an unbiased estimate of the proportion of N remobilized by using the estimate of the proportion of N taken up after silking allocated to kernels. This unbiased estimate can be considered closer to the real proportion of N remobilized than the estimate by the balance method. Indeed, with the balance method, it is assumed that all N assimilated after silking is allocated to the kernels and N from root remobilization is neglected, whereas ¹⁵N method suppresses such assumptions. Furthermore, the correlation between uncorrected and corrected estimates of the proportion of N remobilized (r = 0.78, P 0.001) higher than that between the estimates by the balance method and the corrected ¹⁵N estimates (r = 0.58, P 0.001), shows that the uncorrected estimates were closer than the real values than the balance estimates. Thus, in spite of a significant postsilking ¹⁵N uptake, the ranking of the genotypes according to the biased but accurate estimate of the proportion of N remobilized by using labeling during vegetative growth must approximately correspond to the ranking according to the true proportion of N remobilized.

Nevertheless, for a labeling during the vegetative phase, some modifications in the techniques of ¹⁵N distribution in the field can be used to limit postsilking ¹⁵N uptake. Experiments are in progress to evaluate whether it is possible to reduce the risk of significant residual ¹⁵N uptake after silking by providing ¹⁵N earlier (i.e., before stem elongation). As only a small soil compartment is labeled, we expect the same curves of ¹⁵N uptake as those observed by Sheehy et al. (2004) in rice (Oryza sativa L.) (Fig. 3 ). This could be performed without the risk of increasing heterogeneous ¹⁵N distribution within the plant because during growth before silking, there is probably a redistribution of absorbed N among the leaves (Gallais et al., 2006). Such a phenomenon contributes to fulfilling the assumption of homogeneous ¹⁵N distribution within the plant. It was to favor a homogeneous distribution that we chose to spray ¹⁵N just before stem elongation, which corresponds to the beginning of active growth.

View larger version (16K): [in this window] [in a new window]   Figure 3. Expected effect of the stage of 15N distribution on the 15N uptake and residual 15N uptake after silking. With an early stage of 15N labeling (e.g., at two adult leaf stage) the residual 15N uptake after silking could be about only 5%, whereas with a later 15N labeling (e.g., at the six-leaf stage), it could be higher than 15%.

For a labeling just after silking, the assumptions required for interpreting the results in our ¹⁵N labeling experiment appear more difficult to fulfill than for labeling at the beginning of stem elongation (Gallais et al., 2006). This could explain the significant genotype x year interaction found for the estimates of the proportion of postsilking N uptake allocated to the grain, with the absence of genotype effect in spite of a high accuracy. The consequence was a lower heritability for the proportion of postsilking N uptake allocated to the grain than for the proportion of N remobilized.

In conclusion, the ¹⁵N method appears to be very efficient for determining the proportion of N remobilized. It provided more accuracy and thus a higher heritability than the balance method. Furthermore, with the same number of plots, it did not appear to be more expensive than the balance method that needs three N analyses (one at silking and two at maturity) instead of two at maturity for the ¹⁵N method. Finally, the advantage of the method proposed is that it can be easily applied in the field, on a large number of genotypes, simultaneously with an agronomic evaluation. It can thus be a useful tool in breeding maize for N-use efficiency, to develop cultivars with high or low N remobilization.

	ACKNOWLEDGMENTS

 
The authors are very grateful to Dr. Peter Lea for his careful English revision.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication August 16, 2006.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

• Beauchamp, E.G., L.W. Kannenberg, and R.B. Hunter. 1976. Nitrogen accumulation and translocation in corn. Agron J. 68:418–422.[Abstract/Free Full Text]
• Below, F.E., S.J. Crafts-Brandner, J.E. Harper, and R.H. Hageman. 1985. Uptake, distribution, and remobilization of 15N-labelled urea applied to maize canopies. Agron. J. 77:412–415.[Abstract/Free Full Text]
• Bertin, P., and A. Gallais. 2000. Genetic variation for nitrogen use efficiency in a set of recombinant inbred lines: I. Agrophysiological results. Maydica 45:53–66.[Web of Science]
• Cliquet, J.B., E. Deléens, A. Bousser, M. Martin, J.L.C. Lescure, J.L. Prioul, A. Mariotti, and J.F. Morot-Gaudry. 1990a. Estimation of C and N allocation during stem elongation by ¹³C and ¹⁵N tracing in Zea mays L. Plant Physiol. 92:79–87.[Abstract/Free Full Text]
• Cliquet, J.B., E. Deléens, and A. Mariotti. 1990b. C and N mobilization from stalk and leaves during kernel filling by ¹³C and ¹⁵N tracing in Zea mays L. Plant Physiol. 94:1547–1553.[Abstract/Free Full Text]
• Coque, M., P. Bertin, B. Hirel, and A. Gallais. 2005. Genetic variation and QTLs for ¹⁵N natural abundance in a set of maize recombinant inbred lines. Field Crop. Res. 97:310–321.
• Deléens, E., J.B. Cliquet, and J.L. Prioul. 1994. Use of ¹³C and ¹⁵N plant label near natural abundance for monitoring carbon and nitrogen partitioning. Aust. J. Plant Physiol. 21:133–146.[Medline]
• Di Fonzo, N., M. Motto, T. Maggiore, R. Sabatino, and F. Salamini. 1993. N-uptake, translocation, and relationships among N-related traits in maize as affected by genotype. Agronomie 2:789–796.
• Friedrich, J.W., and L.E. Schrader. 1979. N deprivation in maize during grain filling: II. Remobilization of ¹⁵N and ³⁵S and the relationship between N and S accumulation. Agron. J. 71:466–472.[Web of Science]
• Gallais, A., and M. Coque. 2005. Genetic variation and selection for nitrogen use efficiency in maize, a synthesis. Maydica 50:531–547.
• Gallais, A., M. Coque, I. Quilléré, J.L. Prioul, and B. Hirel. 2006. Modelling post-silking N-fluxes in maize using ¹⁵N-labeling field experiments. New Phytol. 172(4):696–707.[CrossRef][Web of Science][Medline]
• Kichey, T., B. Hirel, E. Heumez, F. Dubois, and J. Le Gouis. 2007. Postanthesis nitrogen uptake and remobilization to the grain of winter wheat correlations with agronomic traits and nitrogen physiological markers. Field Crop Res. (in press).
• Ma, B.L., and L.M. Dwyer. 1998. Nitrogen uptake and use of two contrasting maize hybrids differing in leaf senescence. Plant Soil 199:283–291.[CrossRef][Web of Science]
• Mae, T., and K. Ohira. 1981. The remobilization of nitrogen related to leaf growth and senescence in rice plants (Oryza sativa L.). Plant Cell Physiol. 24:1079–1086.
• Moll, R.H., E.J. Kamprath, and W.A. Jackson. 1982. Analysis and interpretation of factors which contribute to efficiency of nitrogen utilization. Agron. J. 74:562–564.[Abstract/Free Full Text]
• Pollmer, W.G., D. Eberhard, D. Klein, and B.S. Dhillon. 1979. Genetic control of nitrogen uptake and translocation in maize. Crop Sci. 19:82–86.[Abstract/Free Full Text]
• Pommel, B., A. Gallais, M. Coque, I. Quilléré, B. Hirel, J.L. Prioul, B. Andrieu, and M. Floriot. 2005. Carbon and nitrogen allocation and grain filling in three maize hybrids differing in leaf senescence. Eur. J. Agron. 24:203–211.[CrossRef]
• Rajcan, I., and M. Tollenaar. 1999. Source:sink ratio and leaf senescence in maize: II. Nitrogen metabolism during grain filling. Field Crop. Res. 60:255–265.[CrossRef]
• Sharpe, R.R., and L.A. Harper. 1997. Apparent atmospheric nitrogen loss from hydroponically grown corn. Agron. J. 89:605–609.[Web of Science]
• Sheehy, J.E., M. Mnzava, K.G. Cassman, P.L. Mitchell, P. Pablico, R.P. Robles, and A.B. Ferrer. 2004. Uptake of nitrogen by rice studied with a ¹⁵N point-placement technique. Plant Soil 259:259–265.[CrossRef][Web of Science]
• Ta, C.T., and R.T. Weiland. 1992. Nitrogen partitioning in maize during ear development. Crop Sci. 32:443–451.[Abstract/Free Full Text]

This article has been cited by other articles:

				M. Coque and A. Gallais Genetic Variation for Nitrogen Remobilization and Postsilking Nitrogen Uptake in Maize Recombinant Inbred Lines: Heritabilities and Correlations among Traits Crop Sci., September 1, 2007; 47(5): 1787 - 1796. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gallais, A. Articles by Quilléré, I. Search for Related Content PubMed Articles by Gallais, A. Articles by Quilléré, I. Agricola Articles by Gallais, A. Articles by Quilléré, I. Related Collections Crop Physiology & Metabolism Nitrogen Crop Genetics