The response of the maize nitrate transport system to nitrogen demand and supply across the lifecycle

Authors

  • Trevor Garnett,

    Corresponding author
    1. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    • Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Vanessa Conn,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Darren Plett,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Simon Conn,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Juergen Zanghellini,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. Austrian Centre of Industrial Biotechnology, Vienna, Austria
    Search for more papers by this author
  • Nenah Mackenzie,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Akiko Enju,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Karen Francis,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Luke Holtham,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Ute Roessner,

    1. Australian Centre for Plant Functional Genomics, School of Botany, The University of Melbourne, Parkville, Vic., Australia
    2. Metabolomics Australia School of Botany, University of Melbourne, Vic, Australia
    Search for more papers by this author
  • Berin Boughton,

    1. Australian Centre for Plant Functional Genomics, School of Botany, The University of Melbourne, Parkville, Vic., Australia
    2. Metabolomics Australia School of Botany, University of Melbourne, Vic, Australia
    Search for more papers by this author
  • Antony Bacic,

    1. Australian Centre for Plant Functional Genomics, School of Botany, The University of Melbourne, Parkville, Vic., Australia
    2. Metabolomics Australia School of Botany, University of Melbourne, Vic, Australia
    Search for more papers by this author
  • Neil Shirley,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Antoni Rafalski,

    1. DuPont Crop Genetics, Wilmington, DE, USA
    Search for more papers by this author
  • Kanwarpal Dhugga,

    1. Agricultural Biotechnology, DuPont Pioneer, Johnston, IA, 50131, USA
    Search for more papers by this author
  • Mark Tester,

    1. Australian Centre for Plant Functional Genomics, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    2. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author
  • Brent N. Kaiser

    1. School of Agriculture, Food and Wine, Waite Research Institute, University of Adelaide, Adelaide, SA, Australia
    Search for more papers by this author

Author for correspondence:

Trevor Garnett

Tel: +61 408408085

Email: trevor.garnett@acpfg.com.au

Summary

  • An understanding of nitrate (math formula) uptake throughout the lifecycle of plants, and how this process responds to nitrogen (N) availability, is an important step towards the development of plants with improved nitrogen use efficiency (NUE).
  • math formula uptake capacity and transcript levels of putative high- and low-affinity math formula transporters (NRTs) were profiled across the lifecycle of dwarf maize (Zea mays) plants grown at reduced and adequate math formula.
  • Plants showed major changes in high-affinity math formula uptake capacity across the lifecycle, which varied with changing relative growth rates of roots and shoots. Transcript abundances of putative high-affinity NRTs (predominantly ZmNRT2.1 and ZmNRT2.2) were correlated with two distinct peaks in high-affinity root math formula uptake capacity and also N availability. The reduction in math formula supply during the lifecycle led to a dramatic increase in math formula uptake capacity, which preceded changes in transcript levels of NRTs, suggesting a model with short-term post-translational regulation and longer term transcriptional regulation of math formula uptake capacity.
  • These observations offer new insight into the control of math formula uptake by both plant developmental processes and N availability, and identify key control points that may be targeted by future plant improvement programmes to enhance N uptake relative to availability and/or demand.

Introduction

A vast amount (> 100 million tonnes) of nitrogen (N) fertilizer is applied to crops annually to maximize yield (FAO, 2006). However, in cereal production, only 40–50% of the applied N is actually taken up by the intended crop (Peoples et al., 1995; Sylvester-Bradley & Kindred, 2009). Given this low N uptake efficiency, we believe a better understanding of the N uptake process in cereals would help to identify the limiting factors contributing to poor N uptake efficiency and overall cereal nitrogen use efficiency (NUE). NUE, in this case, refers to grain yield per unit of available N in the soil (Moll et al., 1982; Dhugga & Waines, 1989; Good et al., 2004).

This study is focused on the uptake and use of nitrate (math formula), as it is the predominant form of N in most high-input agricultural soils (Wolt, 1994; Miller et al., 2007). Plant math formula uptake generally involves two types of transport system, one involving high-affinity (HATS) and the other low-affinity (LATS) transporters (Glass, 2003). In Arabidopsis, four math formula transporters (NRTs) have been linked to math formula uptake from the soil: NRT1.1 and NRT1.2 from the LATS class, and NRT2.1 and NRT2.2 from the HATS class (Tsay et al., 2007). NRT1.1 (Chl1) is unique among these in that it displays dual affinity towards math formula depending on its phosphorylation status (Liu et al., 1999). Although we now have some fundamental knowledge about the functionality of these transporters, our understanding of their roles and of the regulation of math formula uptake remains limited.

Certain aspects of the regulation of the Arabidopsis uptake system have been examined extensively. For example, the math formula uptake capacity of HATS shows strong induction when plants are exposed to math formula after a period of N starvation, and the uptake capacity is repressed following a period of sufficient math formula (Minotti et al., 1969; Jackson et al., 1973; Goyal & Huffaker, 1986; Aslam et al., 1993; Henriksen & Spanswick, 1993; Zhuo et al., 1999). This strong induction and repression are reflected in the transcript levels of AtNRT2.1 and AtNRT2.2, which follow the induction and repression of the uptake capacity (Zhuo et al., 1999; Okamoto et al., 2003). Redinbaugh & Campbell (1993) referred to this pattern of induction and repression as the primary math formula response. Whether this N response is relevant to longer time scales and to soil N characteristics of typical cropping soils has yet to be shown.

The relative roles of NRTs in the uptake of math formula from the soil remain unclear, but circumstantial evidence has been used to postulate their activities. First, the math formula concentration in agricultural soils is generally in the millimolar range (Wolt, 1994; Miller et al., 2007), well above the point at which the math formula HATS system would be saturated (c. 250 μM) (Siddiqi et al., 1990; Kronzucker et al., 1995; Garnett et al., 2003). Second, the location of the transporters within a root suggests variable roles in math formula uptake. AtNRT1.1 expression is localized in the tips of young roots (Huang et al., 1999; Guo et al., 2001), where roots first come into contact with the higher math formula concentrations of unexplored soil, whereas AtNRT2.1 is localized in the cortex of older parts of the root, where external math formula concentrations may be reduced following uptake at the root tip (Nazoa et al., 2003; Remans et al., 2006). Third, the pattern of NRT2 repression observed in roots exposed to sufficient N would seem to limit their relative importance to steady-state math formula uptake in N-rich soils. Given this evidence, it has been proposed that the LATS system is most probably responsible for the majority of math formula uptake from the soil (Glass, 2003).

Little is known about how math formula uptake is actually managed over the lifecycle of the plant, with many studies on math formula uptake focused on responses to perturbations, where external math formula availability is varied in order to explore math formula-dependent uptake responses. In one of the few published studies, Malagoli et al. (2004) measured the uptake capacity of the math formula HATS and LATS in oilseed rape over time, and their response to various factors, and used this information, together with the modelling of field data, to suggest that math formula HATS could supply most of the plants N requirements, even with high N availability. This work suggests that HATS are important in net math formula uptake, necessitating a re-examination of the respective roles of these two transport systems. A detailed analysis of math formula uptake capacity across the entire lifecycle is an important step towards the development of plants with enhanced N uptake capacity and efficiency, and may help to improve N fertilization practice where supply can be better matched to demand.

In this study, we have profiled the changes in math formula uptake capacity in maize plants across a broad developmental time period in response to either reduced or adequate math formula provision. During the lifecycle, the plants were changed between math formula treatments to help distinguish between developmental changes. Given the problems inherent in using a full-sized maize plant for such experiments, we used the dwarf maize ‘Gaspe Flint’, which has a lifecycle of just 60 d, allowing profiling across both vegetative and reproductive stages in a contained environment (Hourcade et al., 1986).

Materials and Methods

Plant growth

Seeds of dwarf maize (Zea mays L. var. Gaspe Flint) were germinated on moist filter paper for 4 d at 28°C. Seedlings were transferred to one of two 700-l ebb and flow hydroponic systems with the fill/drain cycles completed in 13 min. Initially, 150 plants were planted in each system. Plants were grown on mesh collars within tubes (300 mm × 50 mm), which kept the roots of adjacent plants separate, but allowed free access to solution. The hydroponic system was situated in a controlled environment room with a day : night cycle of 14 h : 10 h, 25°C : 20°C, at a flux density at canopy level of c. 500 μmol m−2 s−1. The nutrient solution was a modified Johnson's solution (Johnson et al., 1957) containing (in mM) 0.5 math formulaN, 0.8 K, 0.1 Ca, 0.5 Mg, 1 S and 0.5 P for the 0.5-mM math formula treatment, and 2.5 math formulaN, 1.8 K, 0.6 Ca, 0.5 Mg, 0.5 S and 0.5 P for the 2.5-mM math formula treatment. The choice of concentration was based on preliminary experiments, which suggested that the threshold math formula concentration eliciting a major N response was c. 0.5 mM and this appeared to be the case (Supporting Information Fig. S1). Both treatment solutions contained (in μM): 2 Mn, 2 Zn, 25 B, 0.5 Cu, 0.5 Mo and 100 Fe (as FeEDTA and ethylenediamine-N,N ′-bis(2-hydroxyphenylacetic acid (FeEDDHA)). Iron was supplemented twice weekly with the addition of Fe(NH4)2(SO4)2.6H2O (8 mg l−1). The solution pH was maintained between 5.9 and 6.1. math formula was monitored using an math formula electrode (TPS, Springwood, Qld, Australia) and maintained at the target concentration ± 10%. Other nutrients were monitored using an inductively coupled plasma optical emission spectrometer (ICP-OES: ARL 3580 B, ARL, Lausanne, Switzerland) and showed limited depletion between solution changes. Nutrient solutions were changed every 20 d.

Flux measurement

On sampling days, between 11:00 and 13:00 h, plants were transferred to a controlled environment room with conditions matching growth conditions (light, temperature and relative humidity) and into solutions matching growth solutions. The roots were then given a 5-min rinse with the same nutrient solution, but with either 50 or 250 μM math formula, followed by 10 min of exposure to the same solution, but with 15N-labelled math formula (15N 10%). In preliminary experiments, the flux measured at 50 and 250 μM math formula was found to be before (50 μM) and at the point of (250 μM) saturation of the HATS uptake system. At the end of the flux period, roots were rinsed for 2 min in matching, but unlabelled, solution. Two identical solutions were used for this rinse to allow an initial 5-s rinse to remove labelled solution adhering to the root surface. The flux timing was based on that used by Kronzucker et al. (1995) and chosen to minimize any possible efflux or transport to the shoot.

Roots were blotted, and the roots and separated shoots were weighed and dried at 65°C for 7 d, after which the roots were ground to a fine powder (Clarkson et al., 1996). Total N and 15N in the plant samples were determined with an isotope ratio mass spectrometer (Sercon, Crewe, Cheshire, UK). Unidirectional math formula influx was calculated on the basis of the 15N content of the root. The unidirectional math formula influx measured in this study is described as the uptake capacity of the plant at that point in the lifecycle.

Real-time quantitative PCR (Q-PCR)

On sampling days, root material was harvested between 5 and 7 h after the start of the light period. The whole root was excised and snap frozen in liquid N2 and stored at −80°C. RNA was extracted using the RNeasy Plant Mini Kit with on-column DNase treatment (Qiagen, Hilden, Germany), according to the manufacturer's instructions, before the RNA integrity was checked on a 1.2% (w/v) agarose gel. cDNA synthesis was performed on 1 μg of total RNA with oligo(dT)19 using SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions. Q-PCR was carried out as outlined in Burton et al. (2008). In this method, the amount of each amplicon in each cDNA is quantified with respect to a standard curve of the expected amplicon (typically, PCR efficiencies ranged between 0.85 and 1.05). Four control genes (ZmGaPDh, ZmActin, ZmTubulin and ZmElF1) were utilized for the calculation of the normalization factor. Q-PCR normalization was carried out as detailed in Vandesompele et al. (2002) and Burton et al. (2004). Q-PCR primers were designed for the closest maize homologues of the Arabidopsis NRTs (Plett et al., 2010). Q-PCR products were verified by sequencing, agarose gel electrophoresis and melt-curve analysis to confirm that a single PCR product was being amplified. All primer sequences and Q-PCR product information for control genes and NRT genes can be found in Table S1.

math formula determination

Tissue math formula content was determined via a previous method (Braun-SysteMatic, Methodenblatt N 60; Rayment & Higginson, 1992) scaled appropriately for assay in 96-well optical plates. Frozen and ground tissue (100 mg) was measured into 1.1-ml strip tubes in a 96-well format. Six-hundred microlitres of extraction buffer were added to each tube and the rack of tubes was shaken vigorously for 15 min in a cold room at 4°C. Extraction buffer comprised 50 mM Hepes (pH 7.5), 20% (v/v) glycerol, 1 mM EDTA, 1 mM ethylene glycol tetraacetic acid (EGTA), 0.1% (v/v) Triton X-100, 1 mM benzamidine and 1 mM 6-aminohexanoic acid. Racks were centrifuged at 3400 g at 4°C for 45 min and the supernatant was transferred to fresh tubes. Racks were centrifuged at 3400 g for an additional 45 min at 4°C and the supernatant was transferred to 96-well PCR plates. A clarified soluble extract (15 μl + 10 μl distilled H2O) was added to optical plates and 15 μl of freshly prepared 2 mM CuSO4 and 10 μl of 0.2 M hydrazine sulfate were added to each well. The plates were incubated for 5 min at 37°C and 15 μl of 1 M NaOH was added to each well. The plates were shaken and incubated for 10 min at 37°C. A solution (100 μl) containing equal parts 2.5% (w/v) sulfanilamide in 3.75 M HCl and 0.5% (w/v) N-ethylenediamine was added to each well and the plates were incubated at room temperature for 10 min. The absorbance was measured at 540 nm. KNO3 standards (15 μl) (0–75 nmol/15 μM) were run on each plate and were processed in the same manner as the samples above. The math formula content was expressed as nmoles of math formula per milligram of tissue fresh weight (FW).

Amino acid determination

Tissue amino acid concentration was determined using liquid chromatography electrospray ionization-mass spectrometry, as described by Boughton et al. (2011), once the samples had been derivatized following the method of Cohen & Michaud (1993).

Statistical analyses

Statistical analysis of biomass, flux and metabolite data was carried out using two-way analysis of variance (ANOVA). Data followed a normal distribution. Means of grain yield were tested for significance using a two-tailed t-test. The time course was repeated twice (flux analysis and transcript levels) with similar results.

Results

Biomass

As expected, under our steady-state hydroponic conditions, we observed no difference in either total root or shoot biomass when plants were grown in nutrient solution containing either reduced (0.5 mM) or adequate (2.5 mM) concentrations of math formula (Fig. 1a,b). With both math formula treatments, there was a considerable drop in the root to shoot ratio over the first 18 d after emergence (DAE), highlighting the rapid shoot growth of the plants in the early vegetative period (Fig. 1c). However, our treatments impacted upon the N content between 0.5- and 2.5-mM-grown plants (Fig. 1d). Shoot N concentration was significantly greater (< 0.001) in the whole shoots of plants grown at 2.5 mM math formula than in those of plants grown at 0.5 mM, but, in both treatments, the N concentration was above the critical concentration in the youngest fully expanded blade, which is around 2 mmol g−1 dry weight (DW) N (Reuter & Robinson, 1997). Based on in-season monitoring, these concentrations reflect agronomically realistic math formula concentrations (Miller et al., 2007) and, for the 0.5-mM treatment, represent reduced but not growth-impacting math formula levels, which is important in the context of this study. Irrespective of the math formula concentration supplied; there was a continual drop in tissue N across the lifecycle (Fig. 1d). There was no significant difference in final grain yields (grain DW (g), mean ± SEM: 0.5 mM, 1.85 ± 0.38 (= 12); 2.5 mM, 1.80 ± 0.24 (= 8)). The plants at each of the growth stages can be seen in Fig. S2.

Figure 1.

Growth parameters across the dwarf maize (Zea mays) Gaspe Flint lifecycle of plants grown at either 0.5 mM (open squares) or 2.5 mM (closed squares) math formula. (a) Shoot dry weight (DW), (b) root DW, (c) DW root : shoot ratio and (d) shoot nitrogen (N) concentration (mmol g−1 DW). Fitted curves are as described in the text. There was no significant difference between treatments for shoot biomass, root biomass or root : shoot ratio, and so there is just one fit to the pooled data. Values are means ± SEM (= 8, except for (d) where = 4). *Points significantly different between the two growth conditions (P < 0.05).

math formula flux capacity

Unidirectional math formula HATS flux (e.g. high-affinity math formula uptake capacity) into the root at external concentrations of 50 and 250 μM was determined across various stages of the lifecycle of both 0.5- and 2.5-mM-grown plants (Siddiqi et al., 1990; Kronzucker et al., 1995; Garnett et al., 2003). We observed large, but parallel, fluctuations in HATS math formula uptake capacity over time at both math formula concentrations (Fig. 2a,b), where math formula uptake capacity peaked twice, one coinciding with early vegetative growth (15 DAE) and the other just before flowering (26 DAE). The reduction in uptake capacity between these two peaks (22 DAE) was considerable when measured at 50 μM (c. 20% of the peak value). Apart from the two peaks and the intervening drop, math formula uptake capacity decreased continually from 15 DAE.

Figure 2.

Unidirectional math formula influx into the roots of dwarf maize (Zea mays) Gaspe Flint throughout the lifecycle of plants grown at either 0.5 mM (open squares) or 2.5 mM (closed squares) math formula. Nitrate influx was measured using 15N-labelled math formula over a 10-min influx period with either (a) 50 μM math formula or (b) 250 μM math formula. Values are means ± SEM, = 4. *Points significantly different between the two growth conditions (< 0.05).

HATS uptake capacity of the 0.5-mM-grown plants across most of the lifecycle was generally higher than that of the 2.5-mM-grown plants (c. 50% at 50 μM and c. 40% at 250 μM). This was particularly evident during the early vegetative period of growth (up to 18 DAE), where the math formula uptake capacity measured in 50 μM was enhanced significantly in the plants grown at low external math formula concentrations (Fig. 2a). However, when averaged across the lifecycle, the math formula fluxes measured at 250 μM were c. 20% higher than those measured at 50 μM.

N uptake

To better understand the relationship between growth and N uptake, shoot and root growth, together with tissue N, was used to calculate N uptake over the lifecycle. As there was no difference between treatments for root or shoot biomass, the data were pooled for model fitting irrespective of the treatments. The initial shoot growth rate was much higher than the root growth rate, and a modified exponential function was required to describe the apparent change in the shoot growth rate early after germination, whereas the root data were accurately fitted with an exponential function (Figs 1, S3). Both functions accurately fitted the data, with coefficients of determination (R2) of 0.988 and 0.992 for root and shoot, respectively (Table S2). To model the N content, an allometric relation between N content and shoot biomass (Lemaire & Salette, 1984) was fitted (Fig. S3, inset). To avoid division by zero, N = α/(γ + DWSβ) was used as a fitting function, rather than the usual power law. Here, N denotes shoot N, DWS is the shoot dry weight and α, β and γ are fitting parameters, listed in Table S2. As a result, an improvement was seen in the goodness of fit from R2 = 0.996 to R2 = 0.999. Root N concentration was constant throughout the lifecycle.

Shoot and root dry weight (DW(t)) and N content (N(DW)) were used to calculate the net N uptake of the plants (Ntot(t) = NS·DWS(t) + NR·DWR(t)). The N uptake per g DWR as a function of time (t) is illustrated in Fig. 3 (lines without symbols) and is compared with the experimentally determined math formula uptake capacity for both treatments (open and filled squares). All four datasets showed a comparable peak around day 15. The experimentally measured second peak around day 26 was less pronounced in the calculated values, where a plateau rather than a peak structure was visible. Both features can be understood in terms of the initial mismatch between the root and shoot growth rate.

Figure 3.

Dwarf maize (Zea mays) Gaspe Flint whole-plant net nitrogen (N) uptake per gram root dry weight (DW) as a function of time. Net uptake was calculated from the fitted curves for shoot DW, root DW and shoot N, as shown in Fig. 1 and detailed in the text. Net N uptake is compared with the experimentally determined nitrate flux capacity at 50 μM for different nitrogen treatments (0.5 mM, open squares; 2.5 mM, closed squares; values are means ± SEM, = 4).

Up until 12 DAE, shoots grew almost six times faster than roots (Figs 1a,b, S3). During this time, the N concentration in the shoots remained approximately constant at 3.9 mmol g−1 DW. It would appear that the elevation in N uptake capacity observed by the roots (Figs 2, 3) is a response to meet plant demand for N. Between 10 and 20 DAE, the overall shoot growth rate dropped by > 75%, reaching a final value of 0.0032 h−1. The reduction in shoot growth reduced overall plant demand for N, which was correlated with the observed decrease in the measured math formula uptake capacity beginning from 13 DAE. Similarly, for the second peak (Fig. 2), the exponential phase of shoot growth during this period was roughly 1.3 times faster than that of root growth. Again, it would appear that there is a mismatch in growth-dependent N demand relative to N availability, requiring an up-regulation of N import mechanisms (Fig. 3, see also Fig. S3). However, up-regulation was reduced relative to that of the first peak (Fig. 3). During this period, N concentrations in the shoot decreased from 3.9 mmol g−1 DW at 15 DAE to 2.5 mmol g−1 DW at 40 DAE.

The math formula HATS uptake capacity in the 0.5-mM-grown plants was remarkably similar to the net N uptake rate as calculated from the plant N content (Fig. 3), suggesting that there was little overall LATS input. However, in the 2.5-mM treatment, the uptake capacity of HATS was c. 50% of the actual uptake rate and, given the math formula concentration of this treatment, suggests that there is significant LATS contribution to the net math formula uptake under these conditions. This was supported by our data from experiments in which the LATS capacity was measured at 1 and 4 mM, and was found to be 30% and 100% of the HATS uptake capacity (0–20 DAE), respectively (Fig. S4). This indicates that the LATS uptake capacity measured at 2.5 mM would be close to our estimation of 50%.

To further distinguish between developmental and N responses, a subset of plants was subjected to a change in math formula concentration. At day 15, plants were moved from 0.5 to 2.5 mM math formula (N-inc) and, likewise, plants were moved from 2.5 to 0.5 mM (N-red), a process also repeated at day 22. When the math formula flux capacity was first measured, 3 d after changing math formula concentrations, at both day 15 and day 22, N-red treatments led to a substantial increase in math formula flux capacity (Fig. 4). In N-red treatments at day 15, the initial doubling in uptake capacity relative to plants maintained at 2.5 mM math formula was, nonetheless, followed by a reduction in uptake capacity at day 22 observed in plants with constant math formula concentration. Following the day 22 dip, the uptake capacity returned to a level higher than that of plants kept at 2.5 mM math formula. N-red treatments at day 22 showed a dramatic increase in uptake capacity at day 25. N-inc treatments (plants moved from 0.5 to 2.5 mM math formula) had approximately half the uptake capacity of plants kept at 0.5 mM, and this was maintained until day 40 (Fig. 4).

Figure 4.

Unidirectional math formula influx into the roots of dwarf maize (Zea mays) Gaspe Flint plants grown at either 0.5 mM or 2.5 mM math formula and moved to (a) higher or (b) lower math formulaconcentration at either day 15 or day 22 post-emergence. Nitrate influx was measured using 15N-labelled math formula over a 10-min influx period with 50 μM math formula. Values are means ± SEM, = 4. Dashed lines without symbols are the fluxes presented in Fig. 2(a).

Developmental and nutritional changes to NRT transcript levels

The recent completion of the maize genome sequence provided the opportunity to complete a rigorous survey of cereal homologues to the Arabidopsis NRT genes (Plett et al., 2010), and the naming conventions put forward in that paper are used here. There are currently four NRT genes thought to be involved in root math formula uptake in Arabidopsis (Tsay et al., 2007). However, given the dichotomy between the Arabidopsis NRTs and the cereal NRTs identified by Plett et al. (2010), it was decided to quantify the developmental expression pattern for the relevant maize NRT1, NRT2 and NRT3(NAR2) orthologues of all the known Arabidopsis NRTs on plants grown at either 0.5 or 2.5 mM math formula.

At the whole-root level, transcript levels of the putative HATS genes ZmNRT2.1 and ZmNRT2.2 were significantly more represented in the total RNA pool than those of the other NRT2 or NRT1 genes examined (Figs 5, S5). This may represent simple differences in RNA and/or protein stability between the classes of transport proteins, or may reflect defined roles with respect to math formula transport (Fig. 4). This latter point is suggested by the expression pattern of ZmNRT2.1 and ZmNRT2.2 across the lifecycle, where transcript responses showed remarkable similarity to the patterns observed in the uptake measurements (Fig. 5, see also Figs 2 and 3). Interestingly, both ZmNRT2.1 and ZmNRT2.2 transcript levels were found to be higher in the roots of plants grown at 0.5 mM math formula than in the roots of those grown at 2.5 mM, indicating an N-dependent response; this contrasts with most other NRT genes, where differences in N availability had less of an impact. Notwithstanding the variation in transcript levels of ZmNRT2.1 and ZmNRT2.2 across the lifecycle and the N treatments, the baseline transcript levels from which they varied were also very high, being 200–300-fold higher than the other transporters NRT2 or NRT1 (ZmNRT1.1B) (Fig. 5). Across the lifecycle, this baseline showed a reduction for both transporters, but was far more pronounced for ZmNRT2.1. With regard to the other NRT2s, ZmNRT2.3 showed much lower transcript levels and, although there were similar fluctuations across the lifecycle, there were no clear differences between N treatments. ZmNRT2.5 expression was only detectable in the plants grown in the reduced math formula treatment, with significant variation across the lifecycle.

Figure 5.

Root transcript levels of various putative high- and low-affinity (NRT1, NRT2 and NRT3) math formula transporters throughout the lifecycle of dwarf maize (Zea mays) Gaspe Flint. Plants were grown in nutrient solution containing either 0.5 mM (open squares) or 2.5 mM (closed squares) math formula. The broken lines correspond to maximum math formula uptake capacity as shown by the 15N unidirectional flux analysis (see Fig. 2). Each data point is normalized against control genes, as described in the text. Values are means ± SEM (= 4). *Points significantly different between the two growth conditions (P < 0.05).

Transcript levels of ZmNRT1.1A, ZmNRT1.1B and ZmNRT1.2 were 1000-fold less than those of ZmNRT2.1 and ZmNRT2.2, and did not show the same pattern of variation over the lifecycle as the ZmNRT2s (Fig. 5). Both ZmNRT1.1A and ZmNRT1.1B showed a peak commencing at 13 DAE, coinciding with the ZmNRT2 peak. ZmNRT1.2 showed very low transcript levels until 34 DAE, from where they increased 10-fold. Apart from ZmNRT1.5A, there were no consistent differences in transcript levels of the NRT1s that corresponded to treatment differences in either growth or uptake capacity. ZmNRT1.5A transcript levels were higher in 0.5-mM math formula plants and had a profile matching that of ZmNRT2.1 and ZmNRT2.2. Transcript levels of ZmNRT1.1D, ZmNRT1.3, ZmNRT1.4A, ZmNRT1.4B and ZmNRT1.5B were all very low (Fig. S5), whereas ZmNRT1.1C was undetectable.

The transcript levels of ZmNRT3.1A were 20–100-fold lower than those of ZmNRT2.1 and ZmNRT2.2, but showed the same increase in transcript abundance at 18 and 28 DAE (Fig. 5e). ZmNRT3.1A differs in that it also has a third large peak just before 40 DAE. This third peak showed little difference between the two math formula treatments. The profile of ZmNRT3.2 was more similar to those of ZmNRT2.1/2.2, but the levels were much lower and there were no treatment differences. Transcript levels of ZmNRT3.1B were very low (Fig. S5).

As was seen with plants maintained at constant concentrations, when plants were swapped between math formula treatments at days 15 and 22, the genes that showed the greatest response to N were ZmNRT2.1, ZmNRT2.2, ZmNRT2.5 and ZmNRT1.5a (Figs 6, S6). The patterns of response for ZmNRT2.1 and ZmNRT2.2 were very similar, with plants with increased math formula (N-inc) having lower transcript levels than plants with decreased math formula concentration (N-red).

Figure 6.

Transcript levels of various putative high- and low-affinity (NRT1, NRT2 and NRT3) math formula transporters in roots of dwarf maize (Zea mays) Gaspe Flint plants grown at either 0.5 or 2.5 mM math formula and moved to increased (upper panel) or decreased (lower panel) math formulaconcentration at either day 15 or day 22 post-emergence. Each data point is normalized against control genes, as described in the text. Values are means ± SEM, = 4. Dashed lines without symbols are the transcript values of plants maintained with constant nitrate, as presented in Fig. 4.

The transcript profiles of these N-responsive genes were interesting in that, immediately after transfer to reduced math formula, transcript levels continued with the same trend as before the change in math formula, i.e. they kept decreasing, whereas, at the same time, there was a doubling in uptake capacity (Figs 4b, 6). By contrast, ZmNRT2.1, ZmNRT2.2, ZmNRT2.5 and ZmNRT1.5A all showed a peak in transcript level at day 25, a peak only previously seen in ZmNRT2.5. The transcript levels for ZmNRT2.5 were the most N responsive, with plants moved to higher math formula (N-inc) having no measurable transcripts, whereas those with decreased math formula (N-red) having similar peaks to those maintained at 0.5 mM math formula. ZmNRT1.5A, the only N-responsive ZmNRT1, again showed a major peak in transcript levels at day 25, but none at day 29.

Tissue math formula

Leaf math formula concentrations differed between math formula treatments (< 0.01). In general, leaves of 2.5-mM-treated plants had higher concentrations of math formula. At most time points, the trend in math formula concentration was mirrored between the two treatments, with the exception that leaf math formula in the 0.5-mM treatment was higher than that in the 2.5-mM treatment at 29 and 34 DAE. For both treatments, leaf math formula concentrations before anthesis remained high, but then dropped dramatically after 28 DAE (Fig. 7). There was a more consistent trend in root math formula, with 2.5-mM-treated roots often having higher levels than those exposed to 0.5 mM math formula. Over time, the root trend was similar between treatments in that, at 20 DAE, there was a doubling of root math formula in both treatments and, by 29 DAE, both treatments showed a major drop in root math formula. In the 0.5-mM-grown plants, there was a major spike in root math formula at day 39, a peak also seen in leaf math formula (Fig. 7).

Figure 7.

Nitrate concentration in youngest collared leaf (a) and root (b) tissue of dwarf maize (Zea mays) Gaspe Flint plants grown at either 0.5 mM (open squares) or 2.5 mM (closed squares) math formula. The broken lines correspond to the maximum math formula uptake capacity as shown by the 15N unidirectional flux analysis (see Fig. 2). Values are means ± SEM, = 4. *Points significantly different between the two growth conditions (P < 0.05).

Amino acids

The free amino acid levels showed similar trends in the two math formula treatments (Fig. 8). Apart from the first measurement, where free amino acids in the shoots were very low, root amino acid levels were consistently lower than shoot levels, and this difference increased after day 30, when the shoot level increased but the root level remained the same. For the roots, there was an initial decrease, followed by a peak at 20 DAE, which was common to both treatments. In the shoots, the patterns were less consistent between treatments, with fluctuations showing little correlation.

Figure 8.

Total free amino acid (AA) concentration in root (a) and youngest collared leaf (b) tissue of dwarf maize (Zea mays) Gaspe Flint plants grown at either 0.5 mM (open squares) or 2.5 mM (closed squares) math formula. The broken lines correspond to the maximum math formula uptake capacity, as shown by the 15N unidirectional flux analysis (see Fig. 2). Values are means ± SEM, = 4. *Points significantly different between the two growth conditions (P < 0.05).

Discussion

Across the lifecycle of Gaspe Flint, math formula uptake capacity changed c. 10-fold irrespective of external N availability. This change was characterized by distinct peaks and troughs in math formula uptake capacity, with a general trend towards decreased math formula uptake capacity as plants grew to maturity, but correlation with plant N demand (Figs 2, 3). There was also clear evidence that math formula uptake responded positively to reduced N supply, with increased math formula uptake capacity in the lower N treatment (Fig. 2). The transcript profiles of the NRTs suggested that changes in uptake capacity, in response to math formula supply and demand, were linked to changes in expression of the putative high-affinity NRTs ZmNRT2.1 and ZmNRT2.2. Their expression profiles, in response to N supply and time, provided strong correlative evidence of their in planta roles in math formula uptake. When N supply was varied (N-inc or N-red), the commonality in change to ZmNRT2.1 and ZmNRT2.2 transcript levels and associated change in math formula flux capacity further supported this role. We believe that the highly dynamic nature of N acquisition displayed here and the strong relationship to N provision provide new insights into the regulation of math formula uptake which may lead to the manipulation of N uptake efficiency and, ultimately, NUE in plants.

math formula uptake capacity responding to demand

The math formula uptake capacity was extremely variable across the lifecycle. It has long been suggested that the growth rate determines the N uptake rate (Clement et al., 1978; Lemaire & Salette, 1984; Clarkson et al., 1986). The data presented here support this hypothesis, with the relative differences in growth rate between shoots and roots leading to variability in N demand and changes in math formula uptake capacity (Figs 1, 3). In both treatments, we showed that math formula uptake capacity increased with peaks in shoot growth and, consequently, N demand, but also decreased rapidly when shoot growth decreased, creating a characteristic trough in math formula uptake capacity (Figs 1, 3). We propose that, during this period, the plants grown in 0.5 mM math formula were responding to N limitation and it was plasticity in math formula uptake capacity (HATS) that allowed sufficient N uptake to match the growth rate of the plants grown in 2.5 mM math formula. This plasticity is highlighted by the rapid changes in math formula uptake capacity observed in plants that were changed between math formula treatments.

The manner in which math formula uptake capacity changes in plants with a sustained reduction in the availability of N remains unclear. Most of the literature presents responses in uptake capacity when N is resupplied to plants after a period of reduced N availability, normally resulting in a transient increase in measured math formula uptake capacity (Lee, 1982; Lee & Drew, 1986; Lee & Rudge, 1986; Morgan & Jackson, 1988; Siddiqi et al., 1989). Indeed, there are few results in the literature with which to compare these lifecycle variations in uptake capacity. The work of Malagoli et al. (2004) with oilseed rape is closest in terms of measuring the uptake capacity over the lifecycle. Similar to this study, a spike in math formula uptake capacity was observed corresponding to the time of flowering; however, earlier changes in math formula flux capacity (as observed in the study) were not measured.

Transcript levels of ZmNRTs

The measurement of unidirectional math formula influx at 50 and 250 μM was chosen to describe the uptake capacity of the math formula HATS. Based on reliable estimates from the literature, the math formula HATS for most plants are saturated at c. 250 μM (Siddiqi et al., 1990; Kronzucker et al., 1995; Garnett et al., 2003). Given the relatively high math formula concentrations, at least in the 2.5-mM treatment, which were well above the point at which HATS would be saturated, it was anticipated that LATS would be responsible for much of the uptake. We also expected that there would be little variability in HATS activity based on the steady-state conditions in which we grew the plants, where constitutive HATS (cHATS) activity would be predicted to dominate, and induced HATS (iHATS) would be repressed after continued exposure to math formula. However, this was not the case in either treatment, as evidenced by the influx analysis described above, and in the expression patterns of the NRT gene families, where the math formula HATS responded intimately to math formula supply and demand.

Previous evidence has suggested that HATS transcript levels are generally negatively regulated when N levels are high (e.g. 0.5–2.5 mM math formula) (Filleur et al., 2001; Okamoto et al., 2003, 2006; Santi et al., 2003; Liu et al., 2009). However, in this study, we found the opposite, where the baseline transcript levels of ZmNRT2.1 and ZmNRT2.2 were generally much higher than for any of the other transporters, regardless of the external N supply. Following the paradigm suggested by Glass (2003), the role of the HATS system is to acquire math formula only when soil solution concentrations are low, well below the consistent levels of 0.5 or 2.5 mM used here. However, the high abundance of ZmNRT2.1 and ZmNRT2.2 transcripts, independent of the external N supply, suggests alternative roles for these gene products.

The high level of transcripts of the two putative HATS (ZmNRT2.1 and ZmNRT2.2) contrasts with the low transcript levels observed for the putative LATS, the ZmNRT1s, across the lifecycle. Despite differences in the abundance of LATS and HATS transcripts, there were some parallels in the expression patterns, particularly during the initial peak in math formula uptake capacity (Figs 5, S5). These data support previous reports (Ho et al., 2009) of a possible link between NRT1 and NRT2 transport systems, although, in maize, the relationship may only extend to the early vegetative stage in which math formula uptake capacity is at its maximum. Although the transcript levels of ZmNRT2.5 were very low, the observation that transcripts were only detected in the reduced math formula treatment suggests that this putative transporter may play an important role in low N responses.

The delivery of math formula into the xylem in Arabidopsis has been suggested to involve the NRT AtNRT1.5 (Lin et al. 2008). Unlike other ZmNRT1 genes, ZmNRT1.5A showed a similar transcript profile to ZmNRT2.1/2.2 and was responsive to the 0.5-mM treatment, this being consistent with a possible role in loading math formula into the xylem in maize.

The transcript levels of ZmNRT3.1A were closest in terms of absolute levels to ZmNRT2.1/2.2. There is good evidence that AtNRT3.1 is essential to the function of the AtNRT2s (Okamoto et al., 2006; Orsel et al., 2006; Wirth et al., 2007). Based on transcript levels and the similarity in pattern across the lifecycle, this would also seem to be true for the maize homologues.

The regulation of math formula uptake capacity

There is a correlation between the math formula uptake capacity of HATS and the transcript levels of both ZmNRT2.1 and ZmNRT2.2. This has been found in plants other than maize, and has been proposed as evidence of the involvement of NRT2s in math formula uptake (Forde & Clarkson, 1999; Lejay et al., 1999; Zhuo et al., 1999; Okamoto et al., 2003). Combined with the impairment of math formula uptake associated with reduced transcript levels in Arabidopsis AtNRT2.1 and AtNRT2.2 knockout mutants (Filleur et al., 2001), this led to the proposal that uptake via AtNRT2.1 and AtNRT2.2 is regulated at the transcriptional level. However, transcript levels may not equate to levels of functional protein. Wirth et al. (2007) suggested that the NRT2s in Arabidopsis are long-lived proteins, and showed that the level of AtNRT2.1 protein was independent of transcript level or changes in uptake capacity, suggesting that there is considerable post-translational control of NRT2-mediated math formula uptake.

The results presented here are compatible with a model that combines both transcriptional and post-translational control of math formula uptake capacity (Fig. 9). In this model, the total concentration of ZmNRT2.1 and ZmNRT2.2 protein is predicted to be proportional to the sum of the ZmNRT2.1 and ZmNRT2.2 transcript levels at any given day plus, based on an estimated protein lifespan of NRT2 proteins of c. 5 d (Wirth et al., 2007), the sum of the transcript levels for the previous 4 d. This 5-d lifespan is based on Wirth et al. (2007), but estimates with a range of lifespans are shown in Fig. S7. This estimated protein concentration represents the maximal uptake capacity of NRT2.1 and 2.2 at a given day, the actual uptake capacity being dependent on the amount of post-translational inhibition, which could be through allosteric inhibition, phosphorylation or, given the results of Yong et al. (2010), perhaps a result of NRT2/NRT3(NAR2) complexes being removed from the plasma membrane.

Figure 9.

Predicted ZmNRT2.1/2.2 protein levels based on a protein lifespan of 5 d, and estimated as the sum of ZmNRT2.1 and ZmNRT2.2 transcripts at day x and those of the four previous days, in dwarf maize (Zea mays) Gaspe Flint plants grown at either (a) 0.5 mM or (b) 2.5 mM math formula. Transcript levels are the summed ZmNRT2.1 and ZmNRT2.2 transcripts that were present individually in Fig. 4(a), and the flux capacity is as presented in Fig. 2(a). (b) includes the flux capacity for plants grown at 0.5 mM math formula, but then moved to 2.5 mM math formula at day 15, as presented in Fig. 4(b).

As presented in Fig. 9, this model predicts that, up to day 15, the math formula uptake capacity is equal to the potential uptake capacity, after which the actual uptake capacity measured is then reduced and becomes less than the potential uptake capacity. At day 22, the measured uptake capacity increases through the utilization of the potential uptake capacity without a transcriptional response. This changes at day 27 where, based on our model, the NRT2 protein levels are insufficient to provide the required uptake capacity, this leading to the transcriptional peak observed at day 29. In terms of the plants moved from 2.5 to 0.5 mM math formula at day 15, the initial increase in uptake capacity seen at day 18 in Figs 4(b), 9(b) would be the result of a release of post-translational inhibition, and hence increased uptake capacity without a comparable increase in transcript levels (Fig. 6). The peak in NRT2.1 transcript levels at day 25 would be caused by the number of NRT2.1 proteins in these plants previously exposed to a much higher math formula concentration not providing sufficient uptake capacity, even with no post-translational inhibition. This model predicts that transcription will provide the long-term regulation of math formula uptake capacity, with short-term uptake capacity regulated via the post-translational regulation of the existing transport capacity, this short-term regulation being important for N homeostasis.

The current model of the regulation of math formula uptake by the plant N status (tissue concentration of math formula itself or a downstream assimilate, such as amino acids) has been described in numerous reviews (Cooper & Clarkson, 1989; Imsande & Touraine, 1994; Forde, 2002; Miller et al., 2008; Gojon et al., 2009). The two-component model of math formula uptake capacity regulation described above requires two triggers in its regulation, one a transcriptional trigger and another that determines the extent of post-translational inhibition. Given the major drop in transcript levels beginning at day 18 until day 22, it may be that the trigger for the transcriptional response is the root amino acid/math formula level, which increases and reaches a peak at day 22 (Figs 7, 8). The decrease in uptake capacity beginning at day 15, which we propose is caused by an increase in post-translational inhibition, could be triggered by shoot amino acid/math formula levels which peak at this point.

NUE increases through increased uptake capacity with reduced N availability

The results provide clear evidence that math formula uptake capacity in maize changes dynamically across the developmental growth cycle in response to changes in demand. As suggested previously (Filleur et al., 2001; Okamoto et al., 2003, 2006; Santi et al., 2003; Liu et al., 2009), math formula uptake capacity is highly responsive to N availability and NRT2.1 and NRT2.2 transcription is most likely linked to this response. The focus of future work will be to analyse NRT protein levels, global gene expression and metabolite concentrations at key points of the lifecycle with the aim of gaining a better understanding of how math formula transport is regulated. Such knowledge may benefit programmes directed at increasing NUE and, more specifically, N uptake efficiency in maize.

Acknowledgements

The authors gratefully acknowledge the technical assistance of Stephanie Feakin and Jaskaranbir Kaur, and Steve Tyerman for critical reading of the manuscript. This project was funded by the Australian Centre for Plant Functional Genomics, DuPont Pioneer, Australian Research Council Linkage Grant (LP0776635) to B.N.K., M.T. (University of Adelaide) A.R. and K.D. (DuPont Pioneer).

Ancillary