A Role for Shoot Protein in Shoot–Root Dry Matter Allocation in Higher Plants

Abstract

• Background and Aims It is stated in many recent publications that nitrate () acts as a signal to regulate dry matter partitioning between the shoot and root of higher plants. Here we challenge this hypothesis and present evidence for the viewpoint that and other environmental effects on the shoot : root dry weight ratio (S:R) of higher plants are often related mechanistically to changes in shoot protein concentration.

• Methods The literature on environmental effects on S:R is reviewed, focusing on relationships between S:R, growth and leaf and protein concentrations. A series of experiments carried out to test the proposal that S:R is dependent on shoot protein concentration is highlighted and new data are presented for tobacco (Nicotiana tabacum).

• Key Results/Evidence Results from the literature and new data for tobacco show that S:R and leaf concentration are not significantly correlated over a range of environmental conditions. A mechanism involving the relative availability of C and N substrates for growth in shoots can explain how shoot protein concentration can influence shoot growth and hence root growth and S:R. Generally, results in the literature are compatible with the hypothesis that macronutrients, water, irradiance and CO₂ affect S:R through changes in shoot protein concentration. In detailed studies on several species, including tobacco, a linear regression model incorporating leaf soluble protein concentration and plant dry weight could explain the greater proportion of the variation in S:R within and between treatments over a wide range of conditions.

• Conclusions It is concluded that if can influence the S:R of higher plants, it does so only over a narrow range of conditions. Evidence is strong that environmental effects on S:R are often related mechanistically to their effects on shoot protein concentration.

INTRODUCTION

Changes in irradiance level, photoperiod and supply of CO₂, water and inorganic nutrients can affect the partitioning of dry matter between the shoot and root of higher plants (Andrews et al., 2001; Raven et al., 2005). Consistently, the shoot : root dry weight ratio (S:R) increases with increased supply over the range likely to occur in natural and agricultural soils and it is stated in many recent publications that acts as a signal to regulate dry matter partitioning between the shoot and root of higher plants (e.g. Forde, 2002; Kruse et al., 2002; Foyer et al., 2003; Santi et al., 2003; Scheible et al., 2004). This response of S:R to is viewed by some as one of several -specific effects which contribute to the regulation of plant metabolism and plant architecture (Stitt and Scheible, 1999; Forde, 2002; Scheible et al., 2004).

Here, firstly, we challenge the hypothesis that acts as a signal to regulate S:R. Secondly, we present evidence for the viewpoint that and other environmental effects on S:R of higher plants, are often related mechanistically to changes in shoot protein concentration. Results from the literature and new data for tobacco (Nicotiana tabacum) are utilized in our discussion.

MATERIALS AND METHODS

The initial and repeat experiments were carried out between 1–2 Sep. and 23–24 Oct. and 8–9 Sep. and 30–31 Oct. 2004 in a glasshouse under natural daylight at the University of Sunderland. The temperature was maintained above 15 °C, day and night. Seeds of Nicotiana tabacum L. ‘Petit Havana SR1’ were germinated in sieved John Innes seed compost (John Innes Manufacturers Association, Harrogate, UK) in the glasshouse. After 2–3 weeks, seedlings of approximately equal size were transferred to liquid culture and the different nutrient and irradiance treatments imposed. The treatments were complete nutrient solution (control) and low N, P, S, K and Mg as described in Andrews et al. (1999); low irradiance (6 % open ground PAR, complete nutrient medium) as described in Andrews et al. (2005); and different N forms, where 4 mol m⁻³ in the complete nutrient solution was replaced with 0·5 mol m⁻³ urea, 0·5 mol m⁻³ glutamine or 0·5 mol m⁻³ NH₄NO₃. In all treatments, concentrations of all macronutrients except the deficient nutrient were made equal to those in the control by the addition of the appropriate Na or Cl salt as required (Andrews et al., 1999). Plants were harvested at the onset of flowering and leaf soluble protein and NO₃-concentrations and shoot and root dry weight were determined as described in Andrews et al. (1999, 2005). Both experiments were of completely randomized design with three replicates for all treatments. Data from the two experiments were pooled for statistical analysis and presentation using one-way analysis of variance with nutrient/irradiance treatment as the variable. Linear and quadratic regression analysis was carried out on S:R v plant dry weight, leaf soluble protein concentration and leaf concentration. Variability quoted in the text is the standard error. Before giving these experimental data, evidence from the literature will be reviewed and discussed.

LEAF REGULATES S:R: LIMITATIONS

The assertion that regulates S:R is primarily based on results from studies on mutants and transformants of tobacco with decreased expression of nitrate reductase (NR). In particular, a highly significant positive correlation between leaf content and S:R was found for eight different genotypes growing at a wide range of supply (Scheible et al., 1997). However, plants can take up and utilize a range of forms of N, and S:R increases with N supply regardless of N form (Andrews et al., 1985, 1999, 2001, 2004a, b). There is evidence that nitrification of reduced N can occur within shoot tissue of some species (Watt and Cresswell, 1987; Hipkin et al., 2004) and, although levels of produced are likely to be low, it cannot be discounted that they are high enough to act as a ‘signal’ in some processes such as stomatal closure (Raven, 2003). However, it is stressed that S:R changes with supply over the range which affects growth (dry matter production). Indeed, there are several reports that S:R increases with increased supply above that which gives maximum growth: this effect is associated with increased tissue reduced N (total N – /NO₂ – N) concentration (Andrews et al., 1999). For example, for common bean (Phaseolus vulgaris), plant dry weight increased with increased supply from 0·5 to 4–6 mol m⁻³, then decreased with a further increase in supply to 10 mol m⁻³, but S:R and plant reduced-N content increased with increased supply throughout (Fig. 1). There was a strong positive correlation between S:R and tissue reduced N concentration (r=0·97, P < 0·0001). Leaf concentration differed to S:R in its response to supply and ranged from 6 to 14 µmol g⁻¹ d. wt at 0·5–4 mol m⁻³ then increased 20-fold with increased supply from 4·0 to 6·0 mol m⁻³, the range of supply where growth reached a maximum (Fig. 1). There are reports in the literature, for several other species, that leaf concentrations are low and change little with increased leaf supply until maximum growth is reached (Khamis and Lamaze, 1990; Zhen and Leigh, 1990; Dastgheib et al., 1995), although this is not always the case (Andrews et al., 1992).

Fig. 1.

The effect of different applied concentrations on total plant dry weight (DW), shoot to root dry weight ratio (S:R), leaf concentration and total plant reduced N concentration of common bean. Taken from Andrews et al. (1995).

In the study of Scheible et al. (1997), the major proportion of the change in S:R of tobacco was associated with exceptionally high leaf concentrations, which would rarely occur under natural or agricultural conditions. Specifically, values for S:R ranged from around 2 to 10, with those above 3·5 associated with leaf concentrations of 500–3000 µmol g⁻¹ d. wt. Also, there were deviations from the strong relationship between leaf content and S:R at low leaf concentrations. Plants that were grown on low had S:R values that lay below the regression line, while plants that were grown on alone or NH₄NO₃ had S:R values above the regression line. In addition, the relationship between leaf content and S:R did not hold under P deficiency. In relation to nutrition, it was suggested that discrepancies might be due to a restriction of root growth as a result of acidification; competition between root growth and assimilation in the roots or a separate signal from deficiency that is generated in N metabolism. It was argued that P deficiency acts via a separate signal from deficiency. Subsequent work on NR transformants of tobacco has shown that the relationship between S:R and leaf concentration does not hold at a twice ambient CO₂ concentration (Kruse et al., 2002).

Generally, levels in plant tissues are positively related to uptake but this need not necessarily be the case. It has been suggested that it is the influx of into the shoot, or xylem loading of which determines S:R (Stitt and Krapp, 1999; Kruse et al., 2002). However, neither of these hypotheses can explain the effects of different N forms on S:R. Also, in the study of Kruse et al. (2002), S:R and xylem sap concentrations were not significantly correlated across genotype and CO₂ concentrations. New data for tobacco, presented and discussed below, further emphasize the limitations of the proposal that leaf concentration regulates S:R (Table 1).

Table 1.

Shoot : root dry weight ratio (S:R), total plant dry weight and leaf soluble protein and concentrations for tobacco supplied basal nutrient solution (control, as N source); basal nutrient solution deficient in N, P, S, K or Mg; basal nutrient solution with replaced with urea, glutamine (Gln) or NH₄NO₃ as N source and basal nutrient solution under low irradiance (Ir)

	S:R	Dry weight	Protein (mg g−1 d. wt)	(µmol g−1 d. wt)
Low Mg	7·02	3·60	211	540 ± 36
Control	6·79	5·36	170	112 ± 3
Low K	6·26	1·38	140	409 ± 18
Low Ir	5·92	0·82	129	1801 ± 87
NH4NO3	4·51	2·23	79·8	59 ± 5
Gln	4·37	1·08	97·7	4 ± 1
Low S	4·24	3·47	74·8	159 ± 18
Low P	3·47	2·61	108	411 ± 14
Low N	2·58	1·58	34·3	21 ± 4
Urea	2·48	0·97	35·5	4 ± 1
LSD	0·275	0·412	13·48

Variability quoted for concentration is the standard error.

VIEWPOINT: N AND OTHER RESOURCES AFFECT S:R THROUGH EFFECTS ON SHOOT PROTEIN

The mechanism

Various mechanisms other than signalling have been proposed to explain the N effect on S:R (Bastow-Wilson, 1988; Andrews et al., 1999, 2001). Bastow-Wilson (1988) reviewed models for the control of S:R and concluded that changes in S:R in response to deficits of macronutrients, water, irradiance and CO₂ usually conform to the Thornley (1972) model. In this model, the factors that determine S:R are the supply of C and N substrates by the shoot and root, respectively, transport of these substrates between shoot and root and their incorporation into plant structure. It was argued that structural growth of shoot and root is co-limited by the local C and N substrate concentrations and that this growth acts as a sink for substrates to which further substrates diffuse from the points of supply. It was further argued that the rate of transport of C and N substrate from shoot to root and root to shoot, respectively, is proportional to the concentration gradient divided by a resistance. Hence, a decrease in C substrate acquisition would result in an increase in S:R while a decrease in N substrate acquisition would cause S:R to decrease. A weakness of this model is that, although there is strong evidence that transport of C from shoot to root is driven by a concentration gradient of C substrate, N transport from root to shoot occurs primarily via mass flow through the xylem, driven by transpiration (Pate, 1980; Dewar, 1993).

Dewar (1993) developed the Thornley (1972) model such that a fraction of the N taken up by the root is translocated in the xylem transpiration stream from the root to the shoot where it is transferred laterally to the shoot phloem. The remaining fraction of the N taken up is transferred directly to the root phloem. Also, a fraction of the N translocated to the shoot is subsequently translocated back to the root in the phloem at a rate determined by the shoot to root gradient of labile C in accordance with the Münch pressure flow mechanism. Shoot and root growth rates are considered to be functions of local water potentials and labile C and N concentrations. It is assumed that the plant water balance is in instantaneous equilibrium for given values of shoot and root structure, so that the rate of shoot transpiration is equal to the rate of uptake of water by the root. Shoot and root water potential are calculated directly in terms of shoot and root dry matter and the rate of transpiration. Similarly, the proportion of N taken up that is allocated to the shoot is in direct proportion to the fraction of plant biomass contained in the shoot. Shoot N substrate is carried to the root as phloem translocate at a rate determined by the gradient of C substrate concentration. It was argued that as long as the fraction of N taken up that is transported in the xylem to the shoot is less than, or equal to, the shoot fraction then the Münch pressure flow mechanism of phloem translocation would always ensure that there is a higher labile N concentration in the root than in the shoot, opposite to the concentration gradient of labile C. The Dewar model (Dewar, 1993), as with that of Thornley (1972), relies on the existence of a gradient of N substrate between the root and shoot with the N substrate concentration greater in root than shoot. This will not always be the case. For example, although under nutrition, the root is the main site of uptake, the site of assimilation (production of amino acids) is the source of N that will be used for growth. Considerable data indicate that for many higher plants, the shoot is the main site of assimilation at low and high external concentrations (Andrews, 1986). Also, typically, there is little transported in the phloem (Pate, 1980; Andrews et al., 2004a). The Thornley and Dewar models (Thornley, 1972; Dewar, 1993) cannot explain a decrease in S:R, with decreased supply, for species which have the shoot as their main site of assimilation at low and high supply. Nevertheless, generally, predictions made from the Thornley/Dewar models, relating S:R to the relative availability of C and N substrate for growth and empirical/functional equilibrium/optimization models relating S:R to tissue N concentration, are in good agreement with experimental data (e.g. Ågren and Ingestad, 1987; Levin et al., 1989; Hilbert and Reynolds, 1991; Gleeson, 1993; Ågren and Franklin, 2003). However, as discussed below, across N form or different macronutrient treatments, S:R is more closely correlated with leaf soluble protein concentration than with leaf, shoot or plant N concentration.

In a solid substrate (e.g. soil, sand, vermiculite/perlite), the growth rate for a range of plant species increased with increased supply, from a very low value at 0·1 mol m⁻³ applied to a maximum in the range 1–5 mol m⁻³ , then changed little or decreased with increased supply to 20 mol m⁻³, whilst tissue N concentration and S:R increased with increased throughout (Andrews et al., 2001; Fig. 1). For several species grown on , a significant, a positive linear relationship was found between S:R and whole plant or shoot N concentration per unit dry weight. There are reports that for plants of similar dry weight, S:R is greater with than with as N source (Scheible et al., 1997; Andrews et al., 2001, and references therein). However, where tested, the tissue N concentration for plants of similar dry weight was also greater with than with . For barley (Hordeum vulgare) and non-N₂-fixing common bean (Phaseolus vulgaris), the relationship between S:R and plant N concentration was similar with or as N source (Andrews et al., 1999). However, for non-N₂-fixing pea (Pisum sativum) there was a strong positive correlation between S:R and plant and shoot N concentration with or as N source, but the relationships between S:R and plant and shoot N concentration were substantially different with the two N forms (Andrews et al., 1999). For pea, the relationship between S:R and leaf soluble protein concentration was similar with the two N forms.

There are several reports for grain legumes that S:R is greater for N₂-fixing plants than for -fed plants of similar dry weight; this difference was related to an increased sink for photosynthate imposed by the nodules (Marschner, 1995). Andrews et al. (2004b) examined the relationships between S:R, growth and leaf soluble protein concentrations for pea inoculated with Rhizobium leguminosarum and supplied with low (0·5 mol m⁻³) and uninoculated plants supplied with a range of concentrations from 0·5 to 10 mol m⁻³. Inoculation and increased supply to 4 mol m⁻³, resulted in increases in S:R, growth and leaf soluble protein concentration. S:R and leaf protein concentration were as great for inoculated plants as for plants on 4 mol m⁻³ supply, although plant dry weight was 55 % greater with the treatment. A linear regression model incorporating leaf soluble protein concentration and plant dry weight could explain 78 % of the variation in S:R of plants within and between the inoculated and uninoculated plant treatments. Omission of data for the inoculated plants from this analysis reduced this value by 2 % to 76 %. Thus, if there is a nodulation-specific effect on S:R, it appears to be insignificant outside the effects of nodulation on leaf protein concentration and growth.

Our view is that the increase in S:R with increased N supply, regardless of its effect on growth (but excluding or toxicity effects), is due to an increase in N relative to C substrate for shoot growth in conjunction with the proximity of the shoot to the C supply. Specifically, increased N supply results in increases in N uptake, N assimilation and tissue organic N concentration. The increase in organic N concentration is likely to be due to increases in a range of N-containing molecules, but mainly amino acids, soluble protein and insoluble membrane-bound proteins with the relative proportions of each dependent on environmental conditions (Millard, 1988; Evans and Seemann, 1989; Andrews et al., 1999). Nitrogen uptake, N assimilation and protein synthesis are energy-requiring processes, hence the increase in organic N concentration reflects an increased proportion of energy/C derived from photosynthesis being utilized in processing N. However, N is a component of chlorophyll and photosynthetic enzymes and hence can influence photosynthesis greatly (Lawlor, 2002). If increased processing of N results in increased photosynthate available for growth, shoot dry weight will increase relative to root dry weight due to proximity of the shoot to the C source and increased organic N available for growth. Also, if growth increases, part of the N effect on S:R may be a growth/ontogenetic effect, although for several species under steady-state nutrition (constant internal N and constant relative growth rate), S:R was found to remain constant at a value dependent on tissue N concentration (Ågren and Ingestad, 1987; Ingestad and Ågren, 1991; Ågren and Franklin, 2003). Thus growth-related changes in S:R may be due to how nutrients are applied over time. Possible effects of growth on S:R need further testing. Nitrogen productivity (C gain per unit N per unit time) decreases with increased organic N concentration. If organic N concentration increases but the photosynthate available for growth changes little or decreases, S:R will still increase as again the shoot will realize a greater proportion of its growth potential due to its proximity to the source of C and the availability of reduced N for growth. It is proposed that shoot protein concentration is of particular importance as this reflects the availability of N substrate and N catalyst for shoot growth (Andrews et al., 1999, 2001). This hypothesis is independent of the form of N nutrition and the site of N assimilation and is similar to the Thornley model (Thornley, 1972), in that structural growth is co-limited by local C and N substrate concentrations and C transport from shoot to root is driven by a concentration gradient of C substrate, but it does not rely on a gradient of N between root and shoot. It is our view that, as for N, other environmental effects on S:R are often primarily mediated through effects on leaf protein concentration and hence shoot and then plant growth. The evidence for this hypothesis is now discussed.

The evidence: literature on root-acquired resources

A series of studies has been carried out to test the proposal that root-acquired resources affect S:R through effects on shoot protein concentration; leaf soluble protein concentration was used as a measure of shoot protein status. If this proposal is correct, then across different environmental variables, there should be a positive correlation between S:R and shoot protein concentration. Andrews et al. (1999) examined relationships between S:R, total plant dry weight, shoot and plant N concentration and leaf soluble protein concentration for pea, common bean and wheat (Triticum aestivum) under different nutrient deficiencies. The effect of nutrient deficiency on S:R was dependent on plant species, specific nutrient and experiment. For example, for all species, in all experiments, S:R decreased with decreased N or P supply while, for Mg deficiency, S:R consistently increased substantially with pea or bean but did not change or decreased with wheat, depending on the experiment. However, despite these differences, a linear regression model incorporating leaf soluble protein concentration and plant dry weight could explain >80 % of the variation in S:R within and between treatments for pea supplied with different concentrations of or , pea and common bean supplied with different concentrations of N, P, K and Mg, and wheat supplied with different concentrations of N, P, K, Mg, Ca and S. Similarly for annual ryegrass (Lolium multiflorum) in a separate study, in which S:R decreased under N, P or S deficiency but increased under Mg, K or Ca deficiency, or when was replaced by in the complete nutrient medium, a linear regression model incorporating leaf soluble protein concentration could explain 84 % of the variation in S:R within and across treatments (Andrews et al., 2001). In the study of Andrews et al. (1999), the relationship between S:R and leaf soluble protein concentration was, generally, much stronger than that between S:R and leaf N, shoot N or plant N concentration. This indicates that leaf soluble protein concentration is more important than overall plant N status in determining S:R.

Generally, S:R increases with increased water supply over the range which causes increased growth (McDonald and Davies, 1996; Andrews et al., 2001). Often this response is likely to have been at least in part a growth/development effect but there are reports for many species that protein synthesis decreases under limiting water supply (Lawlor and Cornic, 2002), thus water could act on S:R via its effect on protein synthesis as well as growth. When tested, results obtained were consistent with this proposal (Andrews et al., 2001). For example, for Himalayan balsam (Impatiens glandulifera) supplied with 0·05–0·4 ml water g⁻¹ substrate, plant dry weight increased with water supply to 0·25 ml g⁻¹ substrate and then decreased with increased water supply thereafter (Andrews et al., 2001). Here, S:R and leaf protein concentration changed little with increased water supply to 0·15 ml g⁻¹ substrate, then decreased steadily with increased water supply thereafter. A linear regression model using leaf soluble protein concentration could explain 84 % of the variation in S:R across water treatments. Thus, results are consistent with the proposal that N form, macronutrient, and water effects on S:R are often primarily mediated through their effects on protein synthesis and growth. Leaf concentration was not measured in these studies but for pea and annual ryegrass supplied with different concentrations of or (Andrews et al., 1999, 2001) and inoculated and uninoculated pea (Andrews et al., 2004b), it seems unlikely that S:R and leaf concentration would have been strongly correlated.

The evidence: literature on shoot-acquired resources

Often, but not invariably, the S:R and leaf protein concentration per unit dry weight of higher plants decrease with the increased growth associated with increased irradiance level or photoperiod (Andrews et al., 2001, and references therein). Thus irradiance could affect S:R via its effect on shoot protein concentration in accordance with our hypothesis. Detailed studies on irradiance effects on growth, S:R and tissue N and protein concentrations provide evidence that this is the case. For example, Ingestad and McDonald (1989) found that for birch (Betula pendula), dry weight increased but S:R and tissue N concentration decreased with increased irradiance over a wide range of supply, and concluded that irradiance affected S:R to an extent corresponding to its effect on the N status of the plant. Also, for Tradescantia fluminensis supplied with 5 mol m⁻³ , plant dry weight increased with increased irradiance from 1 % to around 50 % relative irradiance (Ir; open ground irradiance = 100 % Ir), then changed little with increased irradiance thereafter (Maule et al., 1995). Here, the S:R and leaf soluble protein concentration increased sharply with increased irradiance to around 10 % Ir, then decreased steadily with increased irradiance to 50 % Ir. A linear regression model utilizing leaf soluble protein concentration and plant dry weight could explain 87 % of the variation in S:R across irradiance levels. Similarly, for Himalayan balsam supplied with 1 or 5 mol m⁻³ , at a range of relative irradiance levels (1–55 % Ir), plant dry weight increased with irradiance from 1 % to 8 % Ir and from 1 % to 28 % Ir at the lower and higher concentrations, respectively (Maule, 2000). In general, S:R decreased with increased irradiance throughout and at similar irradiance levels was greater at 5 than 1 mol m⁻³ . The S:R was not significantly related to plant dry weight but was significantly related to leaf soluble protein concentration. A linear regression model incorporating leaf soluble protein concentration and plant dry weight could explain 60 % of the variation in S:R within and across treatments and 92 % of the variation across treatment means.

Generally, S:R changes little or decreases with increased growth associated with increased CO₂ concentration (Stulen et al., 1998; Poorter and Nagel, 2000). We have not tested the relationships between CO₂ supply, growth, S:R and leaf protein concentration but the available data are consistent with our proposal that CO₂ affects S:R through effects on shoot protein concentration. Specifically, where tested, decreased S:R with increased CO₂ was usually associated with decreased leaf N and/or protein concentration. It has been argued by several workers that CO₂ affects S:R via its effect on N status and if nutrient supply is maintained at optimal level then S:R is little affected by CO₂ supply (Stulen et al., 1998; Poorter and Nagel, 2000).

Increased S:R associated with decreased irradiance is likely to be associated with increased leaf concentration but the magnitude of the increase in leaf concentration in shade is often much greater than that associated with high supply (Maule, 2000; Andrews et al., 2005; Table 1). The relationship between leaf concentration and S:R does not hold for NR transformants of tobacco at twice ambient CO₂ concentration (Kruse et al., 2002).

The evidence: new data for tobacco

Table 1 shows S:R, total plant dry weight and leaf soluble protein and concentrations for tobacco under different nutrient deficiencies, low irradiance and when in the complete nutrient solution was replaced with other N forms; all measurements varied greatly depending on treatment (P < 0·001). In relation to the different N form treatments, the nutrient solutions used were not sterilized and there is likely to have been a degree of N transformation within the pots. Nevertheless, the data shown in Table 1 indicate major differences in relationships between total plant dry weight and leaf soluble protein and concentrations across N-form treatment which provides evidence that there were differences in the major form of N taken up and assimilated. For example, leaf soluble protein concentration was almost three times greater with glutamine than with low , but total plant dry weight was around 50 % greater with low . Across all treatments, S:R was not significantly correlated with plant dry weight or leaf concentration but there was a strong positive relationship between S:R and leaf soluble protein concentration (Fig. 2). The linear component could explain 82 % of the variation in S:R within and across the treatments. Only the values for the low P treatment obviously fell outside this line. When the low P treatment values were omitted from the analysis, the linear component could explain 91 % of the variation in S:R within and across treatments, although there is an indication that the curve is ‘flattening off’ and a quadratic model gave an R² value of 96 %. Such a strong relationship between S:R and leaf protein concentration over such a wide range of conditions is further evidence that leaf protein concentration often plays an important role in the regulation of S:R.

Fig. 2.

The relationship between S:R and leaf soluble protein concentration for tobacco under different nutrient and low irradiance conditions (see Table 1): replicate values are presented. Filled squares indicate values for P-deficient plants.

Scheible et al. (1997) reported that on high supply, tobacco transformants with very low NR activity (note this is with high S:R and leaf concentrations), had leaf protein concentrations similar to the -limited wild type. This at first appears to be inconsistent with our proposal that affects S:R through changes in shoot protein concentration. However, Scheible et al. (1997) presented protein concentrations on a fresh weight basis and it is likely that water content and hence protein per unit dry weight were substantially greater in the high NR transformants than in the low wild type due to the osmotic effect of accumulation (Andrews et al., 2005). The finding of Scheible et al. (1997), that roots of tobacco transformants on high supply with lower accumulation than in shoots contained high levels of protein, support this proposal. Also, although the tobacco transformants resembled the N-deficient wild type with respect to starch content, starch turnover and sugar levels when grown on low supply, they behaved differently on high supply. Here, when accumulated to high levels, the leaves contained much less starch and greater sugar concentrations than expected in an N-deficient plant. The potential magnitude of the effect of accumulation on the difference between protein levels per unit fresh weight, or dry weight, is highlighted using the data obtained for tobacco here. At low N and low irradiance, respectively, protein concentrations were 4·13±0·15 and 5·25 ± 0·19 mg g⁻¹ f. wt leaf but, on a dry weight basis, values were almost four times greater with the low irradiance treatment (Table 1). Recent work that examined starch mobilization induced by resupply to N-starvedArabidopsis plants found that this process was blocked in an NR-null mutant (Wang et al., 2004). As and glutamine induced starch mobilization in the wild type and mutant, it was concluded that reduction was necessary for this process to occur. Wang et al. (2004) highlighted that their findings were not consistent with respect to the results reported by Scheible et al. (1997) for NR deficient tobacco, where starch mobilization was similar in the wild-type and mutant lines. It was proposed that residual reduction in the NR-deficient tobacco accounted for the mobilization of starch in these experiments.

CONCLUSIONS

Results from the literature and new data for tobacco show that S:R and leaf concentrations are not significantly correlated over a wide range of conditions. A mechanism involving the relative availability of C and N substrates for growth in shoots can explain how shoot protein concentration affects shoot growth and, hence, root growth and S:R. Generally, results in the literature are compatible with the hypothesis that macro-nutrients, water, irradiance and CO₂ affect S:R through effects on shoot protein concentration. In detailed studies on several species, including tobacco, a linear regression model incorporating leaf soluble protein concentration and plant dry weight could explain the greater proportion of the variation in S:R within and between treatments over a wide range of conditions. It is concluded that evidence is strong that environmental effects on S:R are often related mechanistically to their effects on leaf protein concentration and not leaf concentration. It is recommended that leaf protein concentration is measured in studies where environmental effects on dry matter partitioning are investigated.