Changes in the acquisition and partitioning of carbon and nitrogen in the gibberellin-deficient mutants A70 and W335 of tomato (Solanum lycopersicum L.)

Authors

  • O. W. Nagel,

    1. Plant Ecophysiology, Utrecht University, PO BOX 800·84, 3508 TB Utrecht, The Netherlands and
    Search for more papers by this author
  • H. Lambers

    1. Plant Ecophysiology, Utrecht University, PO BOX 800·84, 3508 TB Utrecht, The Netherlands and
    2. Plant Sciences, Faculty of Agriculture, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia
    Search for more papers by this author

Correspondence: Oscar W. Nagel, Division of Environmental and Applied Biology, School of Life Sciences, University of Dundee, Dundee DD1 4HN, UK. Fax: + 44 (0)1382 344275; e-mail: o.w.nagel@dundee.ac.uk

Abstract

Even though the growth-promoting effects of gibberellins (GAs) in plants are well established, little is known about GA action on carbon metabolism and the available reports seem contradictory. We studied the effects of GA deficiency in mutants of tomato (Solanum lycopersicum L.) on rates of carbon acquisition and the allocation of acquired carbon to growth and respiration of leaves, stems and roots. Carbon budgets were calculated from 24 h measurements of photosynthesis and respiration. The partitioning of nitrogen compounds to leaves, stems and roots, which strongly influences carbon budgets, was also studied. The GA-deficient mutants acquired less carbon per unit plant mass per day than did the wild type and used a larger fraction of it for root growth and root respiration. To find out to what extent these changes were just consequences of restriction of growth, the experiment was repeated at a low exponential nitrate addition rate, which forced all genotypes to grow at the same rate. Under these conditions, the low-GA mutants still photosynthesized and respired faster and partitioned more carbon to root growth than the wild type did. The reasons for the observed differences in carbon economies between the wild type and the low-GA mutants are discussed.

Introduction

Gibberellins strongly promote growth of plant stems and, to a lesser extent, leaves and roots (e.g. Brian & Hemming 1955; Briant 1974; Tanimoto 1987; Nagel, Konings & Lambers 2001a). Remarkably little information is available about the effects of GAs on photosynthesis and respiration, which supply the carbon skeletons and energy for growth. The few reports that are available do not show a clear relationship between GA concentration and rates of photosynthesis, respiration and growth. Some studies have reported that application of GA enhanced photosynthesis and growth (Khan 1996; Hayat et al. 2001; Yuan & Xu 2001), whereas others found that applied GA stimulated growth, but decreased the rate of photosynthesis (Dijkstra, ter Reegen & Kuiper 1990). On the other hand, Thetford et al. (1995) found that the application of inhibitors of GA biosynthesis stimulated photosynthesis, while reducing growth, whereas others found a reduction of both growth and photosynthetic rate (Bode & Wild 1984; Heide, Bush & Evans 1985). Cramer et al. (1995) found no difference in photosynthesis rates between wild-type tomato plants and low-GA mutants. Among the possible causes of these contradictory results are the different ways of measuring and expressing growth and photosynthesis, the different time scales over which effects on growth and photosynthesis have been measured, and the different methods of application of GA or GA inhibitors. To accurately quantify the effects of GA on the relationships between photosynthesis, respiration and growth, we performed 24 h measurements of photosynthesis and respiration of wild-type tomato (Solanum lycopersicum L. cv Moneymaker) plants and the near-isogenic GA-deficient mutants A70 and W335. Rates of photosynthesis and respiration were integrated over the whole 24 h period to calculate the carbon budgets of the wild type and the low-GA mutants. The allocation of reduced nitrogen compounds to leaves, stems and roots strongly influences a plant's carbon balance. Generally, plant species with a high relative growth rate (RGR) allocate a large fraction of nitrogen to their leaves, and have higher rates of photosynthesis per unit leaf nitrogen (Poorter, Remkes & Lambers 1990; Atkin, Botman & Lambers 1996a; Westbeek et al. 1999). To find out whether changes in C economy are associated with differences in distribution of N, reduced nitrogen concentrations of leaves, stems and roots were measured.

Because GA deficiency reduced growth rates, and plants with a low growth rate allocate a smaller fraction of carbon to leaf growth and a much larger fraction to root respiration (Poorter et al. 1990; Van der Werf et al. 1993a; Atkin, Botman & Lambers 1996b), the decreased allocation of C to leaf growth and increased allocation to root respiration of the low-GA mutants may have been brought about by their lower growth potential. To rule out any interference due to differences between the genotypes in RGR or ability to take up nitrate we repeated the experiment at a low relative addition rate of nitrate. Reasons for the observed changes in C budgets due to GA deficiency and the role of uptake and allocation of N herein are discussed. Using our results, we will assess the possibility, as put forward by several authors (Dijkstra et al. 1990; Rood, Zanewitch & Bray 1990) that gibberellins have a role in the inherent variation in RGR amongst plant species.

Materials and methods

Plant growth

Seeds of wild-type tomato (Solanum lycopersicum L. cv Moneymaker) plants and the near-isogenic GA-deficient mutants A70 (mutated at the gib3 locus; moderately deficient) and W335 (gib1 locus; extremely low GA; Koornneef et al. 1990) were germinated on moist filter paper in a growth cabinet. The germination conditions were 25 °C day, 15 °C night, 70% relative humidity, 12 h light, irradiance level 100 µmol m−2 s−1 (PAR). To promote germination of the non-germinating mutant seeds, they were first soaked for 2 h in a solution of 10 µm GA4+7 in demineralized water. The wild-type seeds were soaked in demineralized water for 2 h. Subsequently, the seeds were sterilized in 2% (m/v) sodium hypochlorite and rinsed with demineralized water. After 8 d, the seedlings were transferred to 0·032 m3 containers with aerated nutrient solution in a growth room (20 °C day/night; 70% relative humidity; 14 h day; 500 ± 20 µmol m−2 s−1 PAR, provided by 400 W, Philips HPI-T bulbs; Philips B.V., Eindhoven, The Netherlands). The plants were grown in a nutrient solution (see Nagel, Konings & Lambers 1994 for details) to which nitrate was either supplied at a high concentration (10 mol m−3 NO3) or at a growth-limiting rate, that was achieved using a modified technique of exponential addition (Van der Werf et al. 1993b; after Ingestad 1981). The relative nitrate addition rate (RAR) was chosen to be 100 mg N (g plant N)−1 d−1, which is below the RGR of the most extreme mutant when grown with free access to nitrate. The nutrient solution was replaced every week and destructive growth analysis was used to verify that RGR equalled the nitrate addition rate.

Gas exchange measurements

Photosynthesis and shoot respiration

Intact plants were placed inside a transparent, airtight cuvette with sealed shoot and root compartments, and their net exchange of CO2 was determined during 24 h, using ADC 255 MKIII infra-red gas analysers (Analytical Development Company, Hoddesdon, UK). The open system that was used has been described by Poorter & Welschen (1993). The irradiance during measurements was 450 µmol m−2 s−1 (PAR) and the temperature in the cuvettes was maintained at 22 °C, which was the lowest temperature that could be achieved in this system at the present irradiance. The CO2 readings of the gas analysers were corrected for dilution by water vapour.

Root respiration rates

The root respiration rates were measured as rates of CO2 release, using infrared gas analysis of the air bubbled through the root compartment during the shoot gas exchange measurements (open system; Poorter & Welschen 1993). Immediately after the gas exchange measurements, leaf areas were measured with a Li-Cor LI 3100 leaf area meter (Lambda Instruments Co., Lincoln, NE, USA) and fresh masses were determined. Dry masses were measured after 48 h drying in an oven at 70 °C.

Carbon budgets

The carbon budgets were calculated from rates of net photosynthesis, shoot dark respiration and root respiration, assuming that respiration rates were the same in the light and in the dark.

Chemical analyses

Carbon and total nitrogen concentrations were measured of oven-dried plant material using an automatic C–H–N analyser (Model 1106; Carlo Erba, Milan, Italy) with gas chromatography. Nitrate concentrations were determined colorimetrically using the salicylic acid method (Cataldo et al. 1975). Reduced nitrogen concentrations were derived by subtracting nitrate-N concentrations from total N concentrations.

Statistical analyses

Two-way anova was carried out to determine effects of treatments and interactions between genotype and nitrate supply. A Tukey's a posteriori test was used to indicate differences among means in figures and tables. All statistical analyses were made using the SPSS statistical package (Norusis 1993)

Results

Carbon and nitrogen economy at a high supply of nitrate

When plants were grown with free access to nitrate, the low-GA mutants acquired less carbon per unit plant mass per day than did the wild type. The carbon assimilation rate per unit plant mass per day of the moderately GA-deficient A70 mutant was 82% of that of the wild type, and that of the extremely GA-deficient W335 mutant was only 52% of that of the wild type (Fig. 1). The mutants’ lower rates of carbon gain were not associated with variation in photosynthesis per unit leaf area (Fig. 2a) or the fraction of biomass that was present in leaves (leaf mass ratio, LMR), but were merely due to their lower specific leaf area (SLA; Table 1). Because A70 and W335 assimilated less carbon per unit plant mass per day, and used a larger proportion of it for respiration (Fig. 1), their RGR, as derived from the carbon budgets, were reduced to 81 and 46%, respectively, of that of the wild type (242 mg g−1 d−1). The mutants used a larger proportion of total carbon gain for root growth, at the expense of leaf and stem growth (Fig. 1).

Figure 1.

Carbon budgets, as derived from gas exchange measurements, of wild-type (wt) tomato plants and low gibberellin mutants A70 and W335. The numbers above the pies indicate the gross rates of carbon assimilation per unit total plant mass per day, expressed as a percentage of that of the wild type [10·2 mmol C (g plant)−1 d−1]. The size of the segments indicates the percentage of total carbon gain. Different letters in corresponding segments indicate significant difference between genotypes or N supply. Values are means of four plants.

Figure 2.

Photosynthesis per unit leaf area (Aa; a), nitrogen concentration per unit leaf area ([N]a; b) and photosynthesis per unit leaf nitrogen (A/N; c) of wild-type (circles) A70 (squares) and W335 (triangles) plants at high N supply (filled symbols) and low N supply (open symbols), plotted against plant relative growth rate. Each symbol represents the mean of four measurements ± SE.

Table 1.  Specific leaf area (SLA), leaf mass ratio (LMR), stem mass ratio (SMR), and root mass ratio (RMR) of wild type (Wt) tomato plants and low-GA mutants measured immediately after the 24 h measurements of gas exchange
 SLA
(m2 kg−1)
LMR
(g g−1)
SMR
(g g−1)
RMR
(g g−1)
  1. Values are means ±SE of four plants. Different superscript letter marks in a column indicate significant differences at P < 0·05.

High N
Wt22c ± 1·10·73c ± 0·010·20c ± 0·010·07a ± 0·01
A7017b ± 0·40·68b ± 0·010·21c ± 0·010·11b ± 0·01
W33513a ± 0·60·69b ± 0·010·14b ± 0·010·17c ± 0·01
Low N
Wt17b ± 0·20·52a ± 0·020·19c ± 0·020·29d ± 0·01
A7016ab ± 0·80·50a ± 0·010·14b ± 0·010·37e ± 0·01
W33514ab ± 0·60·49a ± 0·010·10a ± 0·010·42f ± 0·01

Because rates of shoot respiration decreased proportionally with RGR (Fig. 3a), the fraction of carbon that was used for shoot respiration was the same for the three genotypes (Fig. 1). Although root respiration rates were also strongly positively correlated with RGR (R2 = 0·997), the root respiration rates of the mutants decreased to a lesser extent than did their RGR (Fig. 3b), increasing the proportion of C used for root respiration. The fraction of C spent on root respiration was further increased by the higher root mass ratio (RMR) of the mutants.

Figure 3.

Rates of shoot respiration per unit shoot mass (a) and root respiration per unit root mass (b) and net nitrate uptake rate (c) of wild-type (circles) A70 (squares) and W335 (triangles) plants at high N supply (filled symbols) and low N supply (open symbols), plotted against plant relative growth rate. Each symbol represents the mean of four measurements ± SE. NNURs were derived from changes in plant mass, total N concentration and RMR (Garnier 1991).

The net nitrate uptake rates [NNUR; mmol NO3 (g root dry mass)−1 d−1] of W335 and A70 were only 25 and 50%, respectively, of that of the wild type (Fig. 3c); The NNUR of the low-GA mutants therefore decreased more than did their RGR. The GA deficiency increased the total reduced nitrogen concentration per unit plant mass. The nitrogen concentration of W335 was 21% higher than that of the wild type, whereas that of A70 was not significantly different from that of the wild type. In all genotypes, the leaves contained about 75% of the total reduced nitrogen in the plant, but the mutants had a smaller fraction of total nitrogen in their stem and a larger fraction in their roots (Fig. 4). The nitrogen concentration per unit leaf area ([N]a; mmol N m−2) of the mutants was lower than that of the wild type (Fig. 2b), but the rate of photosynthesis per unit leaf area (Aa) was the same for all genotypes (Fig. 2a). Hence, the rate of photosynthesis per unit of leaf nitrogen (A/N; µmol CO2 mol−1 N s−1) of the mutants was lower than that of the wild type (Fig. 2c).

Figure 4.

Nitrogen allocation to leaves, stems and roots of wild-type (wt) tomato plants and low-GA mutants A70 and W335. The numbers above the pies indicate the plant N concentrations, expressed as a percentage of that of the Wt. The size of the segments indicate percentage of total nitrogen. Different letters in corresponding segments indicate significant difference between genotypes or N supply. Values are means of four plants.

Carbon and nitrogen economy at a low nitrate addition rate

When nitrate was supplied at a low exponential rate, the carbon gain per unit biomass of all the genotypes was reduced to about 30% of that of the wild type at high N supply, and the differences in total C gain amongst the genotypes disappeared (Fig. 1). In all genotypes, the reduction in total carbon gain by low N supply was associated with a lower rate of photosynthesis per unit leaf area and a lower LMR. In addition, the wild type's lower SLA also contributed to its lower C gain. The low nitrate addition rate drastically reduced the carbon use for leaf growth and stem growth, but increased that for root growth (Fig. 1). At low N supply, the wild type used a three times larger fraction of total carbon for root growth than it did at high N supply, whereas the fraction of carbon used for root growth only doubled for the mutants (Fig. 1). Although the low N-addition rate reduced the magnitude of the differences in C partitioning to stems and roots amongst the genotypes, these differences remained significant. Therefore, the change in carbon partitioning between stems and roots due to low GA was independent of the effect of the low supply of nitrate.

When grown at a low N-addition rate, all genotypes spent about 50% of their assimilated carbon on respiration (Fig. 1). This high proportion of respiration was completely accounted for by the greatly increased carbon use for root respiration (Fig. 1), which was strongest for the wild type (Fig. 1). The rates of shoot respiration decreased proportionally with the RGR of the genotypes (Fig. 3a); hence, the fraction of carbon used for shoot respiration was unchanged in wild type and A70. The SMR of W335 was even further reduced by low N supply, however (Table 1), so that it used a smaller fraction of carbon for shoot respiration than did the others (Fig. 1). The specific root respiration rates of the three genotypes were only one third of those at high N supply (Fig. 3b), but all RMRs increased threefold (Table 1). The total carbon use for root respiration was therefore the same, whereas the total carbon gain was lower, increasing the fraction of carbon used for root respiration. At the low nitrate addition rate, the rates of net nitrate uptake were extremely low, i.e. only 2% of that of the wild type at high N supply, and the differences amongst the genotypes disappeared (Fig. 3c).

Even at limiting supply of nitrate, the nitrogen concentration per unit total plant dry mass of the mutants was higher than that of the wild type (Fig. 5). A much larger fraction, i.e. 40%, of total nitrogen was allocated to the roots, at the expense of allocation to the leaves. The differences amongst the genotypes in N allocation to their stems remained (Fig. 5). The nitrogen concentrations per unit leaf area were much lower than those at high N supply, but were still higher for the mutants (Fig. 2b). For all genotypes, rates of photosynthesis per unit leaf area were also lower than those at high N supply. Because the concentration of reduced N per unit leaf area decreased to a much lesser extent in the wild type than in the mutants, the rate of photosynthesis per unit leaf nitrogen (A/N) of W335 increased, whereas that of the wild type decreased and that of A70 remained the same as at high N-supply (Fig. 2c).

Figure 5.

A summary of the effects of the influence of GA deficiency on various components of a plant's carbon budget. Arrows indicate utilization of carbon or ATP and plus and minus signs indicate stimulation and inhibition due to GA deficiency, respectively.

Discussion

Effects of GA deficiency on the carbon budget of tomato plants

Figure 5 summarizes the effects of GA deficiency on various aspects of the mutants’ carbon economies. Even though GA deficiency did not affect the mutants’ photosynthesis rates per unit leaf area (Fig. 2a), it did reduce gross carbon acquisition rates via a decrease in SLA (Fig. 5; Table 1). GA deficiency decreased partitioning of carbon to growth, but root growth was not as strongly reduced as shoot growth (Fig. 1), resulting in a smaller fraction of C used for shoot growth and a larger fraction for root growth (Fig. 5). Because shoot growth was reduced, carbon use for shoot respiration decreased proportionally (Figs 5 & 3a). The fraction of net carbon use for root respiration was much larger than could be accounted for by the larger proportion of C used for root growth (Figs 5 & 3b) for reasons which will be discussed below. When forced to grow at the same rate, all genotypes partitioned a much smaller fraction of carbon to biomass, but the differences in allocation to leaves, stem and roots persisted. Therefore, these differences in biomass allocation must be direct consequences of GA deficiency, rather than indirect effects of a reduced growth rate. Partitioning of carbon to respiration massively increased in all genotypes, but the extreme W335 mutant still used a larger fraction of its carbon for root respiration and allocated more carbon to its roots. Therefore, it seems that reduction of growth per se increased partitioning of C to respiration, but that the effect of GA deficiency on respiration further contributes to the effect of the low RAR.

The RGRs calculated from the difference between gross photosynthesis and respiration were lower than those obtained from a destructive growth analysis of plants in the same batch, especially at the low nitrate addition rate. The lower RGRs as calculated from gas exchange measurements may be due to stress resulting from the stronger air movement inside the shoot cuvettes than in the growth room or from transfer of the plants to the cuvette system. This does, however, not affect our conclusions, because at both nitrate regimes the differences in RGR among the genotypes were proportionally the same with either method.

The possible role of gibberellins in inherent variation in RGR amongst plant species

Because species with a low maximum RGR have a low SLA (Poorter & Remkes 1990) and application of GA increases both RGR and SLA (Dijkstra et al. 1990), it has been postulated that GA may be a determinant of interspecific differences in RGR (Dijkstra et al. 1990; Rood et al. 1990). Indeed, the C economies of the slow-growing low-GA mutants very much resemble those of slow-growing species from nutrient-poor habitats (Poorter et al. 1990; Van der Werf, Welschen & Lambers 1992) or alpine habitats (Atkin et al. 1996b). As with slow-growing species from nutrient-poor or alpine environments, the lower total photosynthate production of the mutants is not associated with a lower rate of photosynthesis per unit leaf area, but with a lower SLA. The positive correlation between A/N and RGR (Fig. 2c) is the same as that found for species from sites of different nutrient supply (Poorter et al. 1990; Van der Werf et al. 1993a; Pons, Van der Werf & Lambers 1994; Garnier, Gobin & Poorter 1995), but not for alpine versus lowland species (Atkin et al. 1996a). The low-GA mutants’ smaller partitioning of carbon to leaf growth is also similar to what has been found for slow-growing species from nutrient-poor and alpine habitats, but the lower partitioning to stem growth of W335 is exactly opposite (Poorter & Remkes 1990; Van der Werf et al. 1993b; Atkin et al. 1996b). The negative trend between plant nitrogen concentration and RGR is also opposite to that found when comparing different species. This higher concentration in the low-GA mutants is due to less accumulation of starch, soluble sugars and cell-wall material (Nagel et al. 2001a), causing a lower dry matter percentage. When expressed per unit fresh mass, the nitrogen concentrations of the three genotypes are the same (results not shown). The mutants partitioned a similar fraction of their nitrogen to leaves as the wild type did, but they partitioned a larger fraction to their roots at the expense of that to their stems (Fig. 4). Slow-growing species from nutrient-poor habitats, however, partition more N to their stems and roots, at the expense of partitioning to the leaves (Poorter et al. 1990; Garnier et al. 1995). We conclude that GA deficiency causes different symptoms than those found in slow-growing species, but that nevertheless the carbon budgets of low-GA mutants and slow-growing species are remarkably similar.

Mechanistic relations between growth rate and C budget

The slow-growing low-GA mutants partitioned less carbon to their shoot and more to their roots than the wild type. Because low endogenous GA concentrations reduce expansion and division of cells in leaves and stems (e.g. Rood et al. 1990; Nagel, Konings & Lambers 2001b), the reduction in shoot growth rate is probably a direct effect of the mutants’ low endogenous GA concentration. Low endogenous GA concentrations also reduce the growth rate of isolated roots (Tanimoto 1987; Butcher, Clark & Lenton 1990; Barlow, Brain & Parker 1991; Yaxley et al. 2001), but to a much lesser extent. Therefore, the greater partitioning of carbon to root growth in the A70 and W335 mutants is probably the consequence of a stronger inhibition of leaf and stem expansion than of root expansion (Nagel et al. 2001b). Thus, it appears that the lower growth rate of the slow-growing mutants is not the consequence of the greater carbon use by the roots, but that the larger partitioning of carbon to the roots is the consequence of the strongly reduced shoot growth rate.

Despite their higher concentration of reduced N per unit leaf area, the low-GA mutants had the same rate of photosynthesis per unit leaf area as did the wild type. A number of factors may contribute to this low efficiency of nitrogen use for photosynthesis. The slower carbon acquisition per unit plant mass per day of the mutants is associated with a lower SLA. The lower SLA may limit exposure of the photosynthetic machinery to light, so that a smaller fraction of the total photosynthetic capacity is used (Pons et al. 1994). Another possibility is that photosynthesis is feedback inhibited when the supply of carbohydrates exceeds the demand (Stitt 1997). As the low-GA mutants had lower soluble sugar and starch concentrations in their dried leaf material than did the wild type, the latter option does not seem plausible. The mutants’ much larger use of C for root respiration, on the other hand, seems to contradict a limitation of C acquisition due to their lower SLA. The high use of C for root respiration in the low-GA mutants and in plants grown at the low nitrate addition rate may be partly explained by higher respiratory costs for the maintenance of the larger fraction of root biomass, but, assuming that specific costs for maintenance of biomass do not differ considerably, this cannot account for the threefold difference in respiratory C use (Fig. 1). The root respiration rate of the mutants decreased to a lesser extent than root growth rate (Fig. 3b), which could be due to less efficient energy generation, or higher specific costs for growth, maintenance of biomass, or ion uptake (Van der Werf et al. 1988; Poorter et al. 1991). Higher specific respiratory costs for root growth would be expected if the root biomass of the mutants contained a larger proportion of ‘expensive’ compounds, such as lipids or lignin, but as differences in chemical composition were very small (Niemann et al. 1993), substantial differences in specific costs for growth are unlikely (Poorter 1994). Therefore, the most probable cause for the high root respiration rate of the mutants is a higher energy cost associated with anion uptake, as has been found before by several authors for slow-growing plant species (Van der Werf et al. 1988; Poorter et al. 1991; Scheurwater et al. 1999; Mata et al. 2000). As nitrate uptake rates decreased even more than proportionally with RGR (Fig. 3c), higher energy use for ion uptake can only be explained by much higher energy costs per net mole of anions absorbed. Increased specific respiratory costs for ion uptake in slow-growing plants have been associated with a relatively large efflux of nitrogen from roots to their surrounding medium (Scheurwater et al. 1999; Mata et al. 2000), suggesting that the relatively high root respiration rate of the slow-growing low-GA mutants was also caused by inefficient net nitrate uptake, due to a larger efflux of N (Lambers et al. 1998). We attempted to measure influx and net uptake rate of nitrate, using the 15N stable isotope method as described in Scheurwater et al. (1999), but were unsuccessful, probably due to disturbance of nitrate uptake during transfer of the plants to the solution containing 15NO3 (see Bloom & Sukrapanna 1990).

Final remarks

It follows from the discussion above that under various circumstances that restrict growth (low temperature, low nutrient supply or low endogenous GA), plants reduce their C acquisition rates and allocate more carbon to root respiration. In the present study, the low nitrate addition rate and GA deficiency had very different effects on SLA, photosynthesis per unit leaf area, root respiration and nitrogen allocation, but the overall effects of both factors on C economy were remarkably similar. Variation in RGR is often regarded to be the consequence of variation in rates of photosynthesis and respiration, but our results support an alternative view: the RGR and SLA of the low-GA mutants are under the direct control of gibberellins (see Nagel et al. 2001b), and the rates of photosynthesis per unit leaf area and root respiration seem to be adjusted to the C requirement for growth. The large use of C for root respiration of the slow-growing GA-deficient mutants is probably the consequence of an enhanced efflux of N (Fig. 5), due to the limitation of shoot growth.

Acknowledgments

Professor Maarten Koornneef kindly provided the seeds of wild type and mutants. Ingeborg Scheurwater and Adrie van der Werf are thanked for their constructive criticism on previous versions of this manuscript. This research was supported by the Foundation of Life Sciences (SLW), which is subsidized by the Netherlands Organization for Scientific Research (NWO).

Received 29 November 2001;received inrevised form 15 February 2002;accepted for publication 20 February 2002

Ancillary