Increased proline loading to phloem and its effects on nitrogen uptake and assimilation in water-stressed white clover (Trifolium repens)

Authors

  • Bok-Rye Lee,

    1. Department of Animal Science, Institute of Agricultural Science and Technology, Environment-Friendly Agriculture Research Center (EFARC), College of Agriculture & Life Science, Chonnam National University, Buk-Gwangju PO Box 205, Gwangju, 500-600, Korea
    2. Present address: Department of Metabolic Biology, John Innes Centre, Norwich Research Park, Colney, Norwich NR4 7UH, UK
    Search for more papers by this author
  • Yu Lan Jin,

    1. College of Chemistry and Pharmacy, Qingdao Agriculture University, Qingdao City, Shandong, China
    Search for more papers by this author
  • Jean-Christophe Avice,

    1. INRA, UMR INRA-UCBN, Écophysiologie Végétale, Agronomie et Nutritions NCS, Institut de Biologie Fondamentale et Appliquée, IFR 146 ICORE, Université de Caen, F-14032 Caen Cedex, France
    Search for more papers by this author
  • Jean-Bernard Cliquet,

    1. INRA, UMR INRA-UCBN, Écophysiologie Végétale, Agronomie et Nutritions NCS, Institut de Biologie Fondamentale et Appliquée, IFR 146 ICORE, Université de Caen, F-14032 Caen Cedex, France
    Search for more papers by this author
  • Alain Ourry,

    1. INRA, UMR INRA-UCBN, Écophysiologie Végétale, Agronomie et Nutritions NCS, Institut de Biologie Fondamentale et Appliquée, IFR 146 ICORE, Université de Caen, F-14032 Caen Cedex, France
    Search for more papers by this author
  • Tae-Hwan Kim

    1. Department of Animal Science, Institute of Agricultural Science and Technology, Environment-Friendly Agriculture Research Center (EFARC), College of Agriculture & Life Science, Chonnam National University, Buk-Gwangju PO Box 205, Gwangju, 500-600, Korea
    Search for more papers by this author

Author for correspondence:
Tae-Hwan Kim
Tel:+82-62-530-2126
Email: grassl@chonnam.ac.kr

Summary

  • • The aim of this study was to investigate the physiological significance of increased proline loading to phloem caused by water-deficit stress in relation to nitrogen (N) uptake and assimilation.
  • • N uptake and N assimilation were quantified by 15N tracing in well-watered (control) and water deficit-stressed white clover (Trifolium repens). De novo proline synthesis and proline loading to the phloem were also compared between treatments. The relationships among proline concentrations in phloem exudates, N uptake, and assimilation of newly absorbed N were assessed.
  • • The newly synthesized proline in the phloem exudates increased rapidly after 3 d of water deficit. The water-deficit treatment significantly reduced the maximum nitrate reductase activity (NRA), and also attenuated de novo synthesis of amino acids and proteins in the roots. The increase in proline concentrations in phloem exudates was closely related to reductions in NRA in the roots, N uptake, and the assimilation of newly absorbed N. The accumulation of proline induced in roots by exogenous proline and NH4Cl treatments was closely associated with the decrease in NRA.
  • • These results indicate that increased proline transport to roots via phloem caused by water deficit has a significant influence on the down-regulation of N uptake and the assimilation of newly absorbed N.

Introduction

Proline has been shown to accumulate in a variety of plant species under a broad range of stress conditions, including low water potential (Yancey et al., 1982; Verslues & Sharp, 1999; Kim et al., 2004). Proline functions as a ‘compatible solute’, that is, a compound that can accumulate to high concentrations within the cell cytoplasm without interfering with cellular structure or metabolism (Yancey et al., 1982; Samaras et al., 1995). Also, proline performs many other functions as a hydroxyl radical scavenger (Smirnoff & Cumbes, 1989), a reducer of acidity in the cell (Venkamp et al., 1989), or as a solute that protects macromolecules against denaturation (Schobert & Tschesche, 1978). Proline can also function as an organic nitrogen (N) and/or carbon reserve which can be utilized during recovery via proline degradation in the mitochondria (Trotel et al., 1996).

In a previous study, we found that the reduction in the quantity of N incorporated into the protein fraction caused by water deficit was concomitant with an accumulation of proline in both leaves and roots and with an increase in the inline image concentration in the roots of white clover (Trifolium repens; Kim et al., 2004). This suggests that proline accumulation is a symptom of water deficit associated with excess ammonia production in white clover. This may be attributed to the reduction in de novo protein synthesis under water stress. In this regard, it has been proposed that the accumulation of proline may be associated with protein synthesis and ammonia production in stressed plants. Brugière et al. (1999) showed that glutamine synthetase in phloem was important for the control of proline production in tobacco (Nicotiana tabacum). In white clover, Kim et al. (2004) reported that the accumulation of proline occurred in both the leaves and the roots, and that the rate of increase was higher in the roots, although proline is synthesized preferentially in the chloroplasts of leaves (Rayapati et al., 1989). This indicates that the proline generated in leaves may be transported from the shoot to the roots via the phloem. It has been relatively well established that amino acids, including Glu, Asp and Gln, are transported from the shoot to the root via the phloem (Peeters & Van Laere, 1994; Caputo & Barneix, 1997). It has also been established that amino acids transported to the roots result in an inhibition of inline image uptake and also act as the principal phloem-borne signals for feedback regulation of inline image uptake (Lee et al., 1992; Muller & Touraine, 1992; Vidmar et al., 2000; Beuve et al., 2004).

In this study, we hypothesized that proline accumulated in leaves as a result of water deficit would be transported via phloem. Consequently, proline accumulation could play a significant role in the regulation of the uptake and assimilation of nitrate. In order to test this hypothesis, direct quantifications of proline in phloem exudates, determination of nitrate reductase activity (NRA), and measurements of N uptake and the assimilation of newly absorbed inline image were conducted via 15N tracing in well-watered (control) and water deficit-stressed white clover. The physiological significance of the increased phloem loading of proline in response to water-deficit stress, in terms of effects on the rate of N uptake and its assimilation into amino acids and proteins, was examined. Moreover, an additional experiment was carried out to determine how the endogenous proline concentration responds to exogenous proline or excess N and to assess its relationship with N assimilation.

Materials and Methods

Plant culture

Seeds of white clover (Trifolium repens L. cv. Huia) were surface-sterilized (Kim et al., 1991). Seeds were germinated on wet filter paper in Petri dishes at 30°C in darkness. After 3 d, eight uniform seedlings per pot were planted and then thinned to three after 2 wk. Two-litre pots were filled with a mixture of soil, perlite and cocopeat (50 : 40 : 10, w/w/w), and then watered once to full capacity with 400 ml of complete nutrient solution (Kim et al., 1991), in order to prevent mineral nutrient deficiencies. The seedlings were cultivated in a glasshouse with a day:night mean temperature of 27 : 20°C, and a relative humidity of 65 : 80%. Natural light was supplemented by metal halide lamps which generated c. 400 µmol photons m−2 s−1 at the canopy height for 16 h d−1.

Water-deficit imposition and inline image labelling

The plants were watered regularly to field capacity until the full vegetative stage (c. 12 wk after sowing). At this stage, water-deficit stress was imposed by reducing the volume of water supplied per day. In preliminary tests using the same soil, we monitored changes in leaf water potential (Ψw) occurring in response to daily irrigation with different volumes of water. The resultant data indicated that predawn Ψw was abruptly reduced after the first 2 d when soil was irrigated with < 40 ml per pot d−1, while Ψw was maintained at between –0.84 and –1.02 MPa for 7 d with 400 ml. For these reasons, 400 and 40 ml of daily irrigation per pot were supplied to the well-watered (control) and water-deficit treatments, respectively; half the volume of daily irrigation for each treatment was supplied at 10:00 h and the other half at 16:00 h.

For the 15N feeding of the well-watered treatment (control), 200 ml of 15N solution (1 mm K15NO3 with 5.0 15N atom % excess) was evenly administered via three porous plastic tubes placed on each pot at 10:00 h and 16:00 h, respectively (Kim et al., 2004). For the water-deficit treatment, 20 ml of 15N solution, containing the same percentage of excess 15N atoms and the same amount of N as was applied to the control pot, was administered. The inline image feeding was conducted over 24 h (for the pots sampled on day 1) or 48 h (for sampling days 3, 5, and 7) before each harvest date.

Measurements and sampling

Leaf water potential (Ψw) was immediately determined as the petiole xylem-pressure potential using a pressure chamber (PMS Instrument Co., Corvallis, OR, USA). The net photosynthetic rate was measured using a portable photosynthesis measurement system (LI-6400; Li-Cor, Lincoln, NE, USA) at 22°C on leaves under ambient atmospheric CO2 concentrations (c. 400 ppm) at 1000 µmol m−2 s−1 photosynthetic photon flux (PPF) provided by an LED light. It was measured on the youngest fully expanded leaf, 4 h after the beginning of the photoperiod under glasshouse conditions. The first sampling on day 0 was conducted immediately before the water stress treatment (the first 15NO3 feeding) at 10:00 h. Additional sampling times were 1, 3, 5, and 7 d after treatment. The harvested plants were separated into leaves and roots. All plant samples were frozen immediately with liquid nitrogen, and then freeze-dried, weighed, ground, and stored in vacuum desiccators for further analysis.

Collection of phloem exudates and analysis of 15N-proline in phloem exudates

The phloem exudates were collected in EDTA using the facilitated diffusion method, in accordance with the methods described by Bourgis et al. (1999). The petioles of 15 leaves were cut and immersed rapidly in 20 ml of 20 mm EDTA solution (pH 7.0). After 5 min, the EDTA solution (pH 7.0) was discarded in order to prevent contamination with xylem exudates or cellular fluids. The leaves were rinsed, transferred to another 2 ml of 5 mm EDTA solution in 4-ml tubes, and maintained for 6 h in a growth chamber with 95% relative humidity in darkness in order to prevent transpiration. The exudation solution (5 mm EDTA, pH 7.0) was renewed every 2 h during the 6-h collection period. The samples were concentrated via freeze-drying and maintained at −70°C before proline analysis. The phloem exudates were poured onto cation exchange columns of Dowex 50W × 8 200–400 mesh in the H+ form (Bio-Rad, Richmond, CA, USA). The columns were then washed in 10 ml of distilled water and the proline was eluted with 15 ml (6 m) of NH4OH. This fraction was freeze-dried and re-suspended in 1 ml of 80% methanol for proline analysis. The 15N atom percentage proline excess was determined using a GC-MS system (Fisons MD800 quadrupole, Fisons, Paris, France). An aliquot of 0.5 ml of the proline fraction was taken directly into a silanized vial and dried under nitrogen at 30°C with 50 µl of (2.5 mm) aminobutyric acid added as an internal standard. The dry samples were re-suspended in 50 µl of derivatizing mixture consisting of MTBSTFA (N-tert-Butyldimethylsilyl-N-methyltrifluoroace tamide) (Sigma):α-acetonitrile:triethylamine (15 : 15 : 1, v/v/v), and heated for 30 min at 90°C. A 0.6-µl aliquot of the t-butyldimethylsilyl derivatives obtained was injected into the gas chromatograph fitted with an on-column injector, and a 30 m methylpolysiloxane, 0.25-µm film thickness, fused silica capillary column. Helium was utilized as a carrier gas with a flow rate of 1 ml min−1, and the oven was temperature-programmed to remain at 60°C for 1 min, and then to increase from 60 to 120°C at 30°C min−1, and from 120 to 260°C at 8°C min−1. Mass spectra were acquired with an electron energy of 70 eV, and the mass range was scanned from mass-to-charge ratios of 50–650 with a total scan time of 0.9 s.

Nitrate reductase activity

Maximum NRA was extracted from an aliquot of leaf tissue that had been reduced to a fine powder in a mortar with liquid N (Foyer et al., 1998). The extraction buffer, which consisted of 50 mm 4-orpholinepropanesulfonic acid (Mops)-KOH, pH 7.8, 5 mm NaF, 1 mm Na2MoO4, 10 mm flavin adenosine dinucleotide (FAD), 1 mm leupeptin, 1 mm microcystin, 0.2 g g−1 FW polyvinylpyrrolidone (PVP), 2 mmβ-mercaptoethanol and 5 mm EDTA, was added to the leaf powder (1 ml 50 mg−1 FW). The homogenate was then centrifuged at 4°C for 5 min at 12 000 g. The NRA in the supernatant was assayed immediately. The reaction mixture consisted of 50 mm Mops-KOH buffer, pH 7.5, containing 1 mm NaF, 10 mm KNO3, 0.17 mm NADH and 5 mm EDTA. The reaction was stopped after 16 min by the addition of an equal volume of sulfanilamide (1%, w/v in 3 N HCl) followed by n-napthylethylenediamine dihydrochloride (0.02%, w/v), and the A540 was measured.

Chemical fractionation and isotope analysis

Approximately 200 mg of finely ground freeze-dried samples was extracted twice with 100 mm sodium-phosphate buffer (pH 7.5) at 4°C. The proteins in the combined supernatant were precipitated with 80% (v/v) acetone and centrifuged for 10 min at 10 000 g at 4°C. The resultant pellets, which corresponded to the soluble protein fractions, were re-suspended in 0.5 ml of extraction buffer (Kim et al., 2004). inline image and amino acids in the soluble fraction were separated further with a Dowex 50 H+ column, as described by Kim et al. (2004). The aliquot obtained after protein precipitation was evaporated under vacuum at 4°C, precipitated with 95% ethanol at –20°C, and centrifuged. The resultant supernatant was passed through a H+ column (Dowex 50W × 8). The pH of the solution collected from the H+ column was adjusted to a neutral pH, and this solution was concentrated to a final volume of 0.5 ml (inline image fraction). The amino acids were eluted from the Dowex 50W × 8 column with 25 ml of 0.5 N HCl and concentrated to 1 ml. For the fractionated liquid samples, an appropriate sample volume (usually 0.1 ml) containing more than the minimum quantity of N (25 µg N) was dropped into a tin capsule that was cooled with liquid nitrogen. The contents of the tin capsules were then freeze-dried. Freeze-dried power samples (1–5 mg) were weighed (±10 µg) in tin capsules for the determination of total N. The N content and 15N atom % of all fractions were determined using N single mode analysis on a continuous flow isotope mass spectrometer (IsoPrime; GV Instrument, Manchester, UK) linked to a C/N analyser (EA 3000; EuroVector, Milan, Italy). The obtained 15N abundances were converted to relative specific activities (RSAs, i.e. % of recently incorporated atoms relative to the total numbers of atoms in the sample) via the following equation:

image(Eqn 1)

in which the natural 15N atom % was the 15N atom % of non-15N-fed plants.

The amounts of newly absorbed N (NAN) in the nitrogenous compounds were calculated as follows:

NAN = (RSA × N content measured in a compound)/100 (Eqn 2)

Exogenous proline and excess N treatment

The roots of white clover plants (10-wk-old) grown on the soil mixture were washed free of soil with tap water. The plants were grown hydroponically in a continuously aerated nutrient solution as described previously (Kim et al., 1991) for 2 wk. To begin the treatment, the plants were fed with a new nutrient solution containing various concentrations of proline (0, 10, 20 and 40 mm) and NH4Cl (0, 5, 10 and 20 mm). After 4 d of treatment, plants were separated into shoots and roots. The plant tissues were quickly fixed in liquid nitrogen and stored in a deep freeze before analysis. The proline concentration and NRA were analysed.

Statistical analysis and presentation of data

A completely randomized design was utilized with three replicates for two water levels and five sampling dates. An individual pot containing five plants represented a replicate. Duncan's multiple range test was employed to compare the means of separate replicates. Regression analysis was also undertaken to determine the closeness of the relationship between the measured variables. Unless stated otherwise, the conclusions are predicated on differences between the means, with the significance level set at P < 0.05.

Results

Effect of water deficit on leaf water potential, photosynthesis rate, biomass and proline concentration

The leaf water potential (Ψw) showed a minimum value of −2.11 MPa following 7 d of water-deficit treatment, whereas no significant changes occurred in the controls (Table 1). The rate of photosynthesis decreased rapidly to 57.3% of the control value on the first day, and then continued to decrease to 1.44% of the control value after 7 d (Table 1). The dry leaf mass was unaffected by water-deficit treatment for the first 3 d. A significant reduction (−31.5% of control) of the dry leaf mass was observed after 5 d, but only at day 7 for dry root mass. Water deficit significantly increased the proline concentrations in both the leaves and the root tissues, whereas no significant changes were noted in the well-watered (control) plants. The increase of proline accumulation under water-deficit stress was higher in the roots than in the leaves. The highest proline concentration, 9.4-fold higher than the control value in leaves and 23.4-fold higher in roots, was recorded on day 7.

Table 1.  Changes in leaf water potential, photosynthesis rate, biomass and proline concentration of white clover plants under well-watered (control) or water deficit-stress conditions for 7 d
Leaf water parameters/treatmentDays after treatment
01357
  1. Values are the means of three replicates with three plants each. Values in a vertical column or a horizontal row followed by different letters are significantly different at P < 0.05 according to Duncan's multiple range test.

Water potential (Ψw, MPa)
Control−0.84 a−0.94 ab−0.92 a−1.01 b−1.02 b
Water-stressed−0.85 a−1.13 c−1.61 d−1.94 e−2.11 f
Photosynthesis rate (µmol CO2 m−2 s−1)
Control7.02 b10.80 a9.80 a10.25 a10.13 a
Water-stressed7.24 b4.61 c1.82 d0.19 e0.14 e
Biomass (DM; g per plant)
Leaves
Control5.80 bc6.47 bc7.45 bc8.97 a9.40 a
Water-stressed5.70 c6.56 bc6.72 bc6.14 bc6.64 bc
Roots
Control2.22 b2.21 b2.45 b2.67 ab3.54 a
Water-stressed2.18 b2.12 b1.99 b2.52 b2.29 b
Proline (mg g−1 DW)
Leaves
Control0.65 d0.83 d1.06 d0.90 d0.58 d
Water-stressed0.64 d1.31 d2.67 c3.50 b5.46 a
Roots
Control0.33 c0.45 c0.35 c0.29 c0.29 c
Water-stressed0.33 c0.77 c4.23 b4.42 b6.65 a

Amount of newly absorbed N in plants

The effects of water deficit on N uptake, as estimated from the quantity of 15N distributed to the leaves and roots, are shown in Fig. 1. The quantity of 15N in the leaves and roots increased linearly in both the water-deficit and control plants. However, the absolute quantity of 15N and the rate of increase were significantly lower in the water-stressed plants. On day 7 after treatment, the quantity of 15N in water-stressed plants was reduced significantly to 23.5% and 68%, respectively, in the leaves and roots as compared with the control plants. The total amount of newly absorbed 15N in whole plants was 3.96 mg g−1 DW in the water-stressed plants and 8.38 mg g−1 DW in the controls. This indicates that a 52.7% decrease in N uptake was induced by 7 d of water-deficit treatment.

Figure 1.

Total amount of 15N distributed to the leaves (open bar) and roots (closed bar) of white clover (Trifolium repens) under well-watered (WW) and water deficit-stress (DS) conditions over the 7 d of the experimental period. Bars labelled with the same letters are not significantly different (P > 0.05) according to Duncan's multiple range test.

Nitrate reductase activity, amino acids, and protein incorporation in the roots

In an effort to determine the effects of water deficit on N assimilation in the roots, maximum NRA in the roots was measured and the quantities of 15N-amino acids and 15N-proteins in roots were also quantified (Fig. 2). No significant difference in NRA between treatments was observed within 3 d. On day 7, NRA in the water-stressed plants was reduced significantly and represented 70% of the control values (Fig. 2a). The total amount of N in inline image (newly absorbed (15N-nitrate) + previously stored inline image (14N-nitrate)) in water deficit-stressed plants was slightly higher than that in the control plants only on day 5 (Fig. 2b). Water-deficit stress induced a substantial reduction in 15N-amino acids from day 1. The quantity of 15N-amino acids in the water-deficit roots was 32% lower than that in the control plants on day 7 (Fig. 2c). The reduction in 15N-proteins in the water-stressed roots was already significant (P < 0.05) on the first day of treatment. This then continued until day 7 (14.6% lower than control) (Fig. 2d).

Figure 2.

Changes in maximum nitrate reductase activity (NRA; a), total nitrate pool (inline image + inline image; b), 15N-amino acids (c) and 15N-proteins (d) in roots of white clover (Trifolium repens) under well-watered (open circles) or water deficit-stress (closed circles) conditions. Each value is expressed as the mean ± SE for n = 3. The percentages of inline image on days 1, 3, 5 and 5 were 4.90, 7.89, 6.16, and 6.06% in well-watered and 3.17, 4.83, 5.66 and 3.28% in water deficit-stressed plants, respectively.

Amount of proline in phloem exudates

De novo proline synthesis from newly absorbed N (15N-proline, Fig. 3a) and the amount of previously synthesized proline (14N-proline, Fig. 3b) in the phloem exudates are shown in Fig. 3. The 15N-proline in the phloem exudates was not significantly different between treatments within 1 d. This was then significantly increased by water deficit, particularly after 3 d. On day 7, 15N-proline in the phloem exudates was 12.8-fold higher than in the controls. The nonlabelled 14N-proline also increased linearly from day 1 to day 5 in water-stressed plants (4.4-fold higher than the control), while less substantial changes were observed within the range 5.2–11.3 µg 6 h−1 in the control plants.

Figure 3.

Changes in the amount of 15N-proline in the phloem exudates (a) and nonlabelled 14N concentration in phloem exudates (b) of well-watered (open circles) and water deficit-stressed (closed circles) white clover (Trifolium repens). Each value is the mean ± SE for n = 3.

The relationship between proline in phloem exudates and N assimilation

Correlation of proline in phloem exudates with N uptake, NRA in the roots, total amount of 15N-amino acids, and 15N-proteins in whole plants as affected by water-deficit treatment were respectively assessed (Fig. 4). The measurements were normalized by obtaining the difference (Δ) between the values measured in the water deficit-stressed plants and in the well-watered control plants. The increases in proline in phloem exudates induced by water-deficit treatments were strongly associated with the reduction in total 15N uptake (r = 0.822, P < 0.001; Fig. 4a) and the reduction in NRA in the roots (r = 0.816, P < 0.001; Fig. 4b). The relationships with the total amount of N newly assimilated into amino acids (r = 0.840, P < 0.001; Fig. 4c) and into proteins (r = 0.881, P < 0.001; Fig. 4d) in whole plants were also found to be significant.

Figure 4.

Correlation between proline in phloem exudates and newly absorbed nitrogen (N) (a), maximum nitrate reductase activity (NRA; b) in roots, 15N-total amino acids (c), and 15N-total proteins (d) in whole plants of white clover (Trifolium repens), as affected by water-deficit treatment. Values are normalized to the difference (inline image) between the values of the water deficit-stressed and well-watered (control) plants. Numbers labelling each point indicate the days after treatment (n = 15).

Effects of exogenous proline and excess N supply on endogenous proline concentration and inline image assimilation in roots

To assess further the physiological significance of proline accumulation in regulating N assimilation, proline and NRA were analysed in the roots of white clover grown in the presence of various concentrations of proline and NH4Cl. As indicated in Fig. 5, the endogenous proline concentration significantly increased with increasing exogenous proline (Fig. 5a) and NH4Cl (Fig. 5b) supply. There was a gradual decrease in NRA in response to the increase in the exogenous proline supply (Fig. 5c). The exogenous of NH4Cl also significantly reduced NRA (except between 5 and 10 mm NH4Cl) (Fig. 5d). The linear correlation between proline concentration and NRA in roots exposed to different concentrations of exogenous proline and NH4Cl was evaluated (Fig. 6). The increase in endogenous proline, affected by exogenous proline (r = 0.940, P < 0.001) and NH4Cl (r = 0.709, P < 0.01) supply, was strongly associated with the reduction in NRA in roots.

Figure 5.

Influence of exogenous proline and NH4Cl on proline concentration (a, b) and nitrate reductase activity (NRA; c, d) in roots of white clover (Trifolium repens). Each value is the mean ± SE for n = 3. Bars labelled with the same letters are not significantly different (P > 0.05) according to Duncan's multiple range test.

Figure 6.

The relationship between proline concentration and nitrate reductase activity (NRA) in roots of white clover (Trifolium repens) under exogenous proline (closed symbols) and exogenous NH4Cl (open symbols) conditions. Numbers labelling each point indicate the concentration of exogenous proline (0, 10, 20 and 40 mm) and NH4Cl (0, 5, 10 and 20 mm) (n = 12).

Discussion

Ψw showed a minimum value of −2.11 MPa after 7 d of water-deficit treatment (Table 1). This value was quite similar to that previously recorded (−2.23 MPa) in white clover exposed to 9 d of drying (Singh et al., 2000). The reduction in Ψw may be a consequence of low water uptake and hydraulic flow rates within plants, or of high rates of water loss (Bittman & Simpson, 1989). The data obtained in the current study confirm that the reduction in Ψw with the development of water-deficit stress is responsible for the observed reduction in photosynthetic CO2 assimilation rates (Foyer et al., 1998; Costa França et al., 2000), representing a 36.3% decrease in the photosynthesis rate within the first day of water-deficit treatment (Table 1). A significant (P < 0.05) decrease in DW was recorded on day 5 of the water-deficit treatment.

In this experiment, the quantity of newly absorbed N distributed to both the leaves and the roots decreased significantly within 1 d of water-deficit treatment and continued to diminish until day 7 (Fig. 1). This clearly shows that the white clover plants exposed to a water deficit suffered a restriction of inline image acquisition from the soil. Water deficit can initially disrupt root membrane integrity (Carvajal et al., 1999), which results in a reduction in the uptake of inline image (Klobus et al., 1988; Chaillou et al., 1994). In addition, the quantity of 15N incorporated into amino acids (15N-amino acids) and proteins (15N-proteins) from newly absorbed inline image in leaves on day 7 decreased significantly, by 58.6% and 74.5%, respectively, compared with the control plants (data not shown). The data obtained in the present study suggest that, as the demand for N for amino acids and protein synthesis decreases in leaves,inline image reduction in roots concomitantly decreases (Fig. 2a), and that this decline might result in an accumulation of inline image in the vacuoles of roots (Fig. 2b). High inline image contents in roots are associated with a restriction of inline image uptake from the soil through a reduction in membrane permeability (Raison, 1985). Siddiqi et al. (1989) reported a negative correlation between inline image influx and root inline image content. Water deficit caused a gradual reduction in 15N-amino acids (Fig. 2c) and 15N-proteins (Fig. 2d) in the roots, indicating that in situ inline image reduction in roots was reduced as a result of water deficit.

Proline in the roots accumulated significantly within 1 d of water-deficit treatment (Table 1), which may reflect its transport from leaves to roots through the phloem, because the primary site at which proline biosynthesis occurs is the chloroplasts in the leaves (Rayapati et al., 1989). It was found that proline synthesis in maize (Zea mays) primary root tips was insufficient to meet the requirement for proline (Oaks, 1966; Verslues & Sharp, 1999) and, therefore, that increased transport of proline to the root tip was the source of low water potential-induced proline accumulation (Verslues & Sharp, 1999). As expected, water deficit caused a marked increase in the phloem loading of newly synthesized proline (15N-proline; Fig. 3a) and previously stored proline (14N-proline; Fig. 3b). Similar results were obtained by Girousse et al. (1996), who detected up to a 60-fold increase in the proline concentration of phloem sap collected from the stems of water-stressed alfalfa (Medicago Sativa) plants. The changes in the N status of the shoot (e.g. a decrease in de novo synthesis of amino acids and proteins under water-deficit conditions in this study; data not shown) might be rapidly reflected in the rate of proline delivery to the roots via the phloem. Given that amino acids are the major constituents of the phloem, they would appear to be the prime candidates for long-distance signals that are capable of transmitting information regarding the N status of the shoot. Also, Udea et al. (2007) found that the proline requirement in barley (Hordeum vulgare) roots is possibly supported by phloem translocation from source leaves during water stress, and that HvProT, a proline transporter, facilitates proline uptake in the root cap or cortical cells of the apical region. Phloem-borne amino acids have been identified as good candidates for shoot-derived signals for the feedback regulation of inline image uptake (Cooper & Clarkson, 1989). In this study we detected a strong correlation between the increase in proline in phloem exudates and the reduction in the newly absorbed inline image (r = 0.822, P < 0.001; Fig. 4a). We suggest that increased proline loading to the phloem as a result of water-deficit stress is an important factor in the regulation of inline image uptake. With regard to the other amino acids, similar results suggested that glutamine may be the principal down-regulator of inline image uptake (Vidmar et al., 2000). Also, gamma-aminobutyric acid (GABA) is a nonprotein amino acid which was reported to accumulate at high concentrations in plant tissues and regulate negative feedback acting on inline image uptake (Muller & Touraine, 1992; Bouchéet al., 2003; Beuve et al., 2004; Dluzniewska et al, 2007). 14N-NMR, 13N and 15N labelling, and standard uptake experiments were utilized by Lee et al. (1992) in an effort to assess the effects of potential assimilation inhibitors (Glu, Gln, Ala and Asn) on inline image and inline image uptake by maize roots. These results show that an excess of free amino acids could signal an ample N supply and thus inhibit inline image uptake. Moreover, the increase in proline in the phloem exudates induced by water deficit was strongly associated with the reduction in NRA (r = 0.816, P < 0.001; Fig. 4b), and the reduction in total amino acids (r = 0.840, P < 0.001; Fig. 4c) and proteins (r = 0.881, P < 0.001; Fig. 4d) newly incorporated at the whole-plant level. To clarify the role of proline in the regulation of inline image assimilation, the relationship between proline concentration and NRA in roots, affected by exogenous proline or excess mineral N, was assessed. An exogenous supply of proline or excess N resulted in an increase in the proline concentration (Fig. 5a,b). A negative relationship between proline content and NRA was found (Fig. 6). These results clearly indicate that increased proline loading to the phloem and endogenous proline accumulation in roots result in inhibition of the assimilation of newly absorbed inline image. In our previous works, a linear relationship was found between proline accumulation, which could be associated with excess ammonia, and restriction of de novo protein and amino acid synthesis under water-deficit stress conditions (Kim et al., 2004; Lee et al., 2005). These results were consistent with those obtained by Lovatt (1990) and Lazcano-Ferrat & Lovatt (1999), who provided evidence that abiotic stresses reduce leaf and shoot growth and increase ammonia production, leading to its accumulation or its removal through de novo synthesis of protein amino acids such as proline and arginine.

We demonstrated a strong relationship between the decrease in newly absorbed inline image in the roots and the total amount of reduced N (15N-amino acids + 15N-proteins) derived from the newly absorbed inline image, but a much lower correlation with NRA in the roots (data not shown). Thus, the rate of inline image reduction in situ would be controlled primarily by the rate of inline image uptake, rather than by alterations in NRA. However, it remains unclear whether or not the reduction in N uptake would be responsible for the restriction of shoot growth via the reduction in amino acid supply for protein synthesis or vice versa. In this study, the increase in de novo proline synthesis and proline loading to the phloem coincided with a reduction in de novo amino acid and protein synthesis in the leaves (within 1 d of water-deficit treatment; data not shown), but the restriction of leaf growth occurred later (from 5 d). These results suggest that water deficit principally affects N utilization in the shoot with a concomitant reduction in N uptake as the result of negative feedback regulation by the increased proline concentration and its phloem loading. Despite the evidence that inline image uptake could be limited to shoot growth in white clover exposed to water-deficit conditions, further studies will be required in order to assess the effects of water-deficit conditions on the balance between C assimilation and N assimilation, as these two processes are positively correlated (Foyer et al., 1998).

Acknowledgements

This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2007-313-F00056). We would like to thank Patrick Beauclair (photosynthesis measurements), Marie-Paule Bataillé, and Raphaël Segura (IRMS analysis) for providing equipment and for technical support. We also thank Dr Stanislav Kopriva for helpful discussion.

Ancillary