Allocation of reserve-derived and currently assimilated carbon and nitrogen in seedlings of Helianthus annuus under subambient and elevated CO2 growth conditions


  • Christoph A. Lehmeier,

    1. Lehrstuhl für Grünlandlehre, Technische Universität München, Am Hochanger 1, D−85350 Freising, Germany
    Search for more papers by this author
  • Rudi Schäufele,

    1. Lehrstuhl für Grünlandlehre, Technische Universität München, Am Hochanger 1, D−85350 Freising, Germany
    Search for more papers by this author
  • Hans Schnyder

    Corresponding author
    1. Lehrstuhl für Grünlandlehre, Technische Universität München, Am Hochanger 1, D−85350 Freising, Germany
      Author for correspondence: Hans Schnyder Tel: +49 (0) 8161 713242 Fax: +49 (0) 8161 713243 Email:
    Search for more papers by this author

Author for correspondence: Hans Schnyder Tel: +49 (0) 8161 713242 Fax: +49 (0) 8161 713243 Email:


  • • Here, we analysed the transition from heterotrophic to autotrophic growth of the epigeal species sunflower (Helianthus annuus), and how transition is affected by CO2.
  • • Growth analysis and steady-state 13CO2/12CO2 and 15NO3/14NO3 labelling were used to quantify reserve- and current assimilation-derived carbon (C) and nitrogen (N) allocation to shoots and roots in the presence of 200 and 1000 µmol CO2 mol−1 air.
  • • Growth was not influenced by CO2 until cotyledons unfolded. Then, C accumulation at elevated CO2 increased to a rate 2–2.5 times higher than in subambient CO2 due to increased unit leaf rate (+ 120%) and leaf expansion (+ 60%). CO2 had no effect on mobilization and allocation of reserve-derived C and N, even during the transition period. Export of autotrophic C from cotyledons began immediately following the onset of photosynthetic activity, serving roots and shoots near-simultaneously. Allocation of autotrophic C to shoots was increased at subambient CO2.
  • • The synchrony in transition from heterotrophic to autotrophic supply for different sinks in sunflower contrasts with the sequential transition reported for species with hypogeal germination.


The size and growth rate of a seedling is critical for its competitiveness and survival (Harper, 1977; Schwinning & Weiner, 1998; Walters & Reich, 2000). Initial seedling growth is completely dependent on reserves stored in the endosperm or cotyledons. Although knowledge of the developmental and biochemical events in germinating seed plants is substantial (Ashton, 1976; Bewley & Black, 1994), there is little quantitative knowledge about the importance of reserve- and current assimilation-derived substrates for root and shoot growth of seed plants during the successive phases of seedling growth.

Identification and quantification of the contributions of reserves and current assimilates to growth requires differential labelling of the two substrate sources with a known, constant isotopic composition (Geiger & Shieh, 1988; Schnyder, 1992; Deléens et al., 1994). For studies of the transition from heterotrophic to autotrophic growth, this means that either the current assimilate must be labelled continuously over the entire experimental period (long-term, steady-state labelling), or that the seed reserves must be homogeneously labelled by growing the mother plants in an atmosphere with altered, constant carbon (C) isotope composition of CO2. To date, such studies have been performed with maize (Zea mays; Deléens et al., 1984) and walnut (Juglans regia; Maillard et al., 1994a,b), two species with hypogeal germination, in which the cotyledons do not develop any (walnut) or hardly any (maize) photosynthetic activity. To our knowledge, there are no reports of reserve- and current assimilation-derived C and N (nitrogen) allocation during germination and seedling growth of species with epigeal germination.

The work with walnut and maize indicated that early shoot and root growth were entirely dependent on reserves, but that the reserve-dependence of roots was much longer than that of shoots. This was particularly true for walnut where an approx. 3-wk delay between the onset of photosynthesis and first detection of autotrophic C in roots was observed (Maillard et al., 1994b). In addition, the work with maize (Deléens & Brulfert, 1983) and walnut (Maillard et al., 1994a) indicated that transition from reserve-dependence to current assimilation supply was faster for the respiratory activity of plants than for growth. Thus, it appears, that the transition from heterotrophic to autotrophic substrate supply is asynchronous for different functions. Different kinetics in the transition from reserve to current assimilation dependence suggest that the substrate pools serving root and shoot growth and respiration are at least partially separated. Such a separation could be biochemical – the different functions using chemically distinct substrates (e.g. carbohydrates or amino acids) – or physical, but is not well characterized and understood. Also, it is unknown if this transition pattern is the same in species with epigeal germination, in which the transition is initiated by the photosynthetic activity of the cotyledons.

A related question, which has not been addressed experimentally, is whether the transition from reserve to current assimilation supply to different functions is influenced by environmental conditions, such as the concentration of CO2. In studies with cotyledons of sunflower, Pfeiffer & Kutschera (1996) observed a faster lipid mobilization in light than in the dark. This was interpreted in terms of increased substrate demand by light-induced cotyledon expansion. Eastmond et al. (2000) found that addition of sucrose to seeds of Arabidopsis thaliana, germinating on agar plates caused a delay of storage lipid breakdown, suggesting that the breakdown process is subject to metabolic control.

To shed light on these questions, we analysed (1) the timing and rate of seed reserve-C and -N mobilization, and of their allocation to root and shoot tissue, (2) the time-course and rate of autotrophic C and N acquisition and allocation and (3) how these processes are modified by contrasting CO2 concentrations in sunflower, a species with epigeal germination. Plants were germinated and grown in controlled environment with CO2 concentrations of either 200 or 1000 µmol mol−1 and sampled at intervals over the first 15 d after imbibition (DAI) of seeds. Steady-state 13CO2/12CO2 and 15NO3/14NO3 labelling was used to distinguish and quantify heterotrophic and autotrophic C and N (net) fluxes.

Materials and Methods

Plant material and growth conditions

Seeds of sunflower (Helianthus annuus L., cv. Sanluca) were sown individually in plastic pots (35 cm high, 5 cm diameter) filled with 800 g of washed quartz sand (0.3–0.8 mm grain size). Seeds were preselected to a weight of 60 mg ± 10 mg. Pots were placed in four growth chambers (Conviron E15; Conviron, Winnipeg, Canada) and arranged at a density of 169 plants m−2. Twenty millilitres of modified half-strength Hoagland nutrient solution (2.5 mm KNO3, 2.5 mm Ca(NO3)2, 0.5 mm KH2PO4, 1 mm MgSO4, 0.5 mm NaCl, 0.125 mm iron (Fe) as ethylenediaminetetraacetic acid (EDTA) and micronutrients) were supplied three times per day to every individual plant by automatic drip irrigation. This was true except for 15N-labelled plants, which were irrigated and fertilized manually (see later). Light was supplied by cool white fluorescent tubes and incandescent lamps. Irradiance during the 16 h photoperiod was maintained at 520 µmol m−2 s−1 photosynthetic photon fluence rate (PPFR) at the top of the canopy. Temperature was controlled at 20°C and 16°C and relative humidity at 75% and 80% during the light the dark periods, respectively.

Carbon dioxide control and 13C and 15N labelling

The four growth chambers formed part of the 13CO2/12CO2 exchange and labelling system described by Schnyder et al. (2003). The CO2 concentration inside the chambers was held constant at 200 µmol mol−1 in two chambers and at 1000 µmol mol−1 in the other two chambers. The air supplied to the chambers was generated by mixing CO2-free air and CO2 with known C isotope composition using mass flow controllers. Two sources of CO2 (both from Linde AG, Höllriegelskreuth, Germany) with different isotopic composition were used: CO2 with δ13C −44 and −4 (where δ13C = (13C/12C in the sample/13C/12C in the international VPDB standard) − 1). Each CO2 was supplied to one of the two chambers held at a given CO2 concentration.

To label current uptake-derived N 21 plants in each chamber received a nutrient solution in which the nitrate was enriched to 1 atom%15N.

Sampling, sample preparation, and elemental and isotopic analysis

Sampling dates were arranged according to defined stages of plant development: DAI 3, radicle emerged; DAI 4, erection of the hypocotyl; DAI 4.5, cotyledons unfolded; DAI 6, appearance of the first pair of leaves; DAI 7, first pair of leaves > 1 cm2; DAI 9, appearance of the second pair of leaves; DAI 11, second pair of leaves > 1 cm2; DAI 13, appearance of the third pair of leaves; DAI 15, third pair of leaves > 1 cm2. On every sampling date, four seedlings were sampled in each chamber. Seedlings were removed from pots, cleaned from adhering sand by washing with distilled water, and then dissected into cotyledons, remaining shoot tissue (termed ‘shoot’ in the following, and including the hypocotyl, internodes and leaves) and roots. After determination of the leaf area (LI3100; Licor, Lincoln, NE, USA) the plant parts were frozen in liquid nitrogen and stored in a freezer at −20°C. Samples were dried at 105°C for 1 h followed by 48 h at 60°C, and then ground in a ball mill to flour mesh quality. Aliquots of 0.75 mg ± 0.05 mg of each sample were weighed into tin cups (IVA Analysentechnik e.K., Meerbusch, Germany) for isotope analysis. The tin cups were combusted in an elemental analyser (NA 1110; CarloErba Instruments, Milan, Italy) interfaced (ConfloIII; Finnigan MAT, Bremen, Germany) to an isotope ratio mass spectrometer (Delta plus; Finnigan MAT) for determination of C and N elemental contents and isotopic composition. Each sample was measured against a working gas standard which was previously calibrated against secondary isotope standards (IAEA-CH6 for 13C, accuracy ± 0.06 SD; IAEA-NO3 for 15N, accuracy ± 0.19 SD). A laboratory standard (a fine ground wheat flour) was run regularly after every tenth sample to estimate the precision of the isotope analyses (on average over all measurements ± 0.08 SD for 13C and ± 0.18 SD for 15N).

Data analysis

The fractions of seed reserve- and current assimilation-derived C in samples (fold and fnew) were estimated using procedures analogous to Schnyder & de Visser (1999):

image(Eqn 1a )
image(Eqn 1b )

P, δold, and δnew denote the C isotope composition of the sample, and of the reserve- and current assimilation-derived C; δold was taken as the δ13C of C in the decorticated seed before germination (δold = −26.6 ± 0.8 SE, n = 10)).

A value for δnew could not be determined directly, but was obtained from (Farquhar & Richards, 1984):

image(Eqn 2 )

S is the δ of CO2 inside the chamber, and Δ the 13C discrimination, as expressed in the current assimilation-derived C in biomass; Schnyder & de Visser, 1999).

A value for Δ was obtained as:

image(Eqn 3 )

(H and L denote the chambers supplied with CO2 of ‘high’ (−4) and ‘low’ (−44) δ13C). Derivation of Eqn 3 was based on the assumptions that Δ, and fold (and fnew) were the same in the two chambers maintained at the same CO2 concentration (thus, ΔH = ΔL, and fold H = fold L). These assumptions are justified, since growth conditions were the same in the two chambers, except for the δ13C of CO2. Analysis of variance of samples collected from the two chambers of the same CO2 treatment revealed no difference (P < 0.05), except for their C isotopic composition. Also, it is established that Δ is independent of the C isotope composition of CO2 (Farquhar & Richards, 1984).

The value for Δ obtained by Eqn 3 was in the expected range (27.6 ± 0.2 SE and 23.2 ± 0.2 SE for autotrophic C in the whole plant at elevated and subambient CO2) and did not change during the experiment.

Treatment-, tissue-, sampling date- and size-specific estimates of Δ were calculated to account for all eventual effects of CO2 concentration, allocation, partitioning (including respiration) and ontogeny on Δ, as expressed in a particular sample. To this end the replicates of a particular plant fraction collected from one chamber on a given sampling date were first ranked according to dry mass of the whole plant; Δ was then calculated for every pair of samples with the same rank from the ‘H’ and ‘L’ growth chambers using Eqn 3.

In all calculations we assumed that δold was the same for all tissues and all times. This was probably not exactly true, since the C isotope composition of seed reserves may be heterogeneous (lipids are usually 13C-depleted relative to proteins and nonstructural carbohydrates) (Deines, 1980; Gleixner et al., 1993) and allocation patterns for the different substances could differ. Thus, an error in δold could cause an error in the estimate of fold and fnew. A sensitivity analysis indicated that a variation of ±2 for δold could have caused a relative error in the estimate of fnew and fold of up to ±6%.

The masses of reserve- and current assimilation-derived C in a sample (Cold and Cnew) were obtained as:

image(Eqn 4 )

(C is the total mass of C in a sample).

The fractions and masses of reserve- and current assimilation-derived N in samples were determined by analogous procedures, except that possible fractionation of N during uptake and post-uptake processes was neglected, since any likely fractionation was insignificant relative to the isotopic enrichment in the (labelling) nutrient solution (cf. Handley & Raven, 1992).

Rates of reserve- and current assimilation-derived C (or N) deposition (mg per plant per day) in cotyledons, roots and shoots were calculated from the net change in mass of reserve- and current assimilation-derived C (or N) in the respective organs according to Schnyder & de Visser (1999). The relative contribution (fraction) of autotrophic C and N in the (net) substrate flux feeding roots and shoots was then obtained by relating the net rate of current assimilation-derived C (or N) deposition to the total C deposition in the organ.

Allometric analysis as described by Niklas (1994) was used to distinguish ontogenetic (i.e. size-dependent) and direct CO2 effects on allocation of current assimilation-derived biomass. Because roots and shoots are interdependent variables and both have similar error, model II regression (here reduced major axis, RMA) was used with the linear transformed power function:

image(Eqn 5 )

(Y1 and Y2 denote autotrophic C in shoot and root, respectively; βRMA represents the scaling coefficient and αRMA the scaling exponent).


General growth pattern

Seedlings germinated rapidly following imbibition and grew uniformly in all chambers. Their C balance was initially negative: in both CO2 treatments seedlings lost 16% of total decorticated seed C mass until 4.5 DAI (Fig. 1). Cotyledons started to expand rapidly at 4 DAI, when they shed the seed hull (Fig. 2). Expansion was 20% faster at elevated CO2, but stopped at about the same time (approx. 11 DAI) in both treatments. The first pair of leaves started to expand after 6 DAI, and again expansion was significantly enhanced by elevated CO2 (Fig. 2). At 9 DAI the area of the first leaf pair had reached that of the cotyledons in both CO2 treatments. At 15 DAI – when the present investigation stopped – expansion of the first leaves was still not complete. At that time, leaf area in elevated CO2 was 60% higher than at subambient CO2 (Fig. 2).

Figure 1.

Total carbon (C) content of sunflower (Helianthus annuus) seedlings up to 15 d after imbibition of seeds, grown in 200 µmol mol−1 CO2 (closed circles) and 1000 µmol mol−1 CO2 (open circles). Vertical bars indicate 95% confidence intervals (n = 6–8). Note the logarithmic scaling of the y-axis.

Figure 2.

Leaf area development of cotyledons (circles) and the first pair of leaves (squares) of sunflower (Helianthus annuus) seedlings, grown in 200 µmol mol−1 CO2 (closed symbols) and 1000 µmol mol−1 CO2 (open symbols). Vertical bars indicate 95% confidence intervals (n = 6–8).

The C balance became positive, and growth rate in the two CO2 treatments started to diverge, near the time when cotyledons unfolded (cf. Figs 1 and 2). After that stage seedling growth was near exponential, with a 2–2.5 times higher rate at elevated CO2. At 15 DAI the C mass of seedlings was 4 and 10 times that of the decorticated seed at germination in subambient and elevated CO2, respectively (Fig. 1).

Seed reserve C mobilization and allocation

The timing and rates of reserve C mobilization and allocation were virtually identical in the two CO2 treatments (Fig. 3). Mobilization in cotyledons was rapid until 7 DAI, when cotyledons had lost c. 60% of their initial C mass. Thereafter, mobilization of reserve C in cotyledons continued at a slow rate until the end of the experiment (Fig. 3a).

Figure 3.

Reserve carbon (C) content in different parts of sunflower (Helianthus annuus) seedlings up to 15 d after imbibition, grown in 200 µmol mol−1 CO2 (closed symbols) and 1000 µmol mol−1 CO2 (open symbols); (a) whole plant (circles), cotyledons (squares); (b) root; (c) shoot. Vertical bars indicate 95% confidence intervals (n = 6–8).

A significant fraction of reserve C was lost from the plants. This loss accounted for 30% of the initial C mass and was most active between 3 and 6 DAI (Fig. 3a).

Roots incorporated reserve C at an almost constant rate until 7 DAI, when it stopped abruptly (Fig. 3b). Interestingly, incorporation in the shoot continued for longer, and only stopped at c. 11 DAI (Fig. 3c). Reserve C incorporation in the shoot occurred in two distinct phases. The first lasted until 4.5 DAI and was associated with hypocotyl growth (data not shown). The second started at 7 DAI, stopped at around 11 DAI (Fig. 3c), and was mainly related to incorporation in the first leaves (data not shown).

Of the total 18 mg of reserve C that were mobilized until 11 DAI 38% were lost from the plant (via respiration or exudation), 36% were incorporated into the shoot, and 26% were incorporated into the root. Again, these relationships were not modified by the CO2 treatments.

Seed reserve N mobilization and allocation

Carbon dioxide also had no effect on the timing and rate of N mobilization, loss and allocation (Fig. 4). Moreover, the overall pattern of mobilization and allocation of reserve N was similar to that of reserve C with the following distinctions: (1) the rates of mobilization in the cotyledons and rates of loss from plants were more constant with time (compare Figs 3a and 4a), and (2) allocation to the shoot did not show the marked fluctuations which were evident for C (compare Figs 3c and 4c). In both roots and shoots the rate of reserve-derived N incorporation was rapid until 4.5 DAI, and then gradually declined until it stopped at c. 9 DAI. At that time 50% of the mobilized N was contained in the shoot, 30% was located in the root, and 20% had been lost to the environment.

Figure 4.

Reserve nitrogen (N) content in different parts of the sunflower (Helianthus annuus) seedlings up to 15 d after imbibition, grown in 200 µmol mol−1 CO2 (closed symbols) and 1000 µmol mol−1 CO2 (open symbols); (a) whole plant (circles), cotyledons (squares); (b) root; (c) shoot. Vertical bars indicate 95% confidence intervals (n = 4–8).

Autotrophic C assimilation and allocation

No autotrophic C was detected in any plant part until 4.5 DAI, when cotyledons unfolded. But, on the next sampling date (6 DAI) autotrophic C was detected in cotyledons, shoot and root. These effects were qualitatively the same in both CO2 treatments (Fig. 5).

Figure 5.

Content of autotrophic carbon (C) in different parts of sunflower (Helianthus annuus) seedlings grown in 200 µmol mol−1 CO2 (closed symbols) and 1000 µmol mol−1 CO2 (open symbols); (a) cotyledons; (b) root; (c) shoot. Vertical bars indicate 95% confidence intervals (n = 6–8).

Transition to autotrophic growth was fast. In the period between 4.5 and 7 DAI autotrophic C was the only C source for the shoot (Table 1). Also, autotrophic C was already the main C source for root growth at elevated CO2. Even in subambient CO2, autotrophic C was a significant C source for roots during this period, although the rate of autotrophic C acquisition was only one-third of that at elevated CO2 (Fig. 1). In both treatments more than 85% of the C incorporated into shoots and roots after 7 DAI was derived from the seedlings own photosynthetic activity (Table 1).

Table 1.  Per cent contribution of current assimilation-derived carbon (C) and nitrogen (N) to the net rate of C and N deposition in shoot and roots of sunflower (Helianthus annuus) during successive phases of seedling growth at subambient (200 µmol mol−1) and elevated CO2 (1000 µmol mol−1)
  1. 0–4.5 d after imbibition (DAI; I), 4.5–7 DAI (II), 7–11 DAI (III), and 11–15 DAI (IV for C), or 11–13 DAI (IV for N). Data are means ± SD.

Current-assimilation derived C
Shoot0 099 ± 26103 ± 2088 ± 396 ± 5103 ± 8100 ± 3
Root0 034 ± 7 53 ± 696 ± 1997 ± 11 98 ± 4100 ± 4
Current-assimilation derived N
Shoot7 ± 2 8 ± 384 ± 8 73 ± 999 ± 299 ± 2 99 ± 7100 ± 2
Root9 ± 212 ± 477 ± 9 71 ± 596 ± 599 ± 4 99 ± 36102 ± 14

Plants in subambient CO2 had a significantly higher shoot to root ratio for autotrophic C from 11 DAI (data not shown). Allometric analysis demonstrated that this was a true CO2 effect (Fig. 6; P < 0.05), meaning that low CO2 caused a stimulation of autotrophic C allocation to the shoot, which was distinct from any ontogenetic effect of CO2.

Figure 6.

Autotrophic carbon (C) content of roots vs. autotrophic C of shoots of sunflower (Helianthus annuus) seedlings grown in 200 µmol mol−1 CO2 (closed circles) and 1000 µmol mol−1 CO2 (open circles). Lines are reduced major axis regression curves of log-log transformed data (solid line for closed circles, n = 46, r2 = 0.97; dashed line for open circles, n = 41, r2 = 0.99). Regression for plants in 200 µmol mol−1 CO2 yielded a scaling coefficient βRMA = 0.47 (SE ± 0.03, 95% confidence interval (CI) = 0.41–0.53) and a scaling exponent αRMA = 1.14 (SE ± 0.03, CI = 1.08–1.20). For plants grown in 1000 µmol mol−1 CO2: βRMA = 0.50 (SE ± 0.03, CI = 0.45–0.55) and αRMA = 1.00 (SE ± 0.02, CI = 0.96–1.04).

Newly assimilated N

In both treatments N absorption from the nutrient solution was insignificant relative to N mobilization from cotyledons until 4.5 DAI (data not shown, but see Table 1). Thereafter N uptake was the dominant N source for shoots and roots, and this was true for both treatments (Table 1).


Transition from reserve-dependent to current assimilation-driven growth

The transition from heterotrophic to autotrophic growth of shoots and roots of sunflower exhibited clear differences to the patterns reported for walnut (Maillard et al., 1994a,b) and, to a lesser degree, maize (Deléens et al., 1984). In sunflower the transition started near-simultaneously (at c. 5 DAI) for the shoot and root system (and, probably, also for respiration), and the change was almost complete within a few days. Also, the kinetics of change were similar for C and N. By contrast, Maillard et al. (1994b) demonstrated a 3-wk delay between the onset of photosynthesis and first detection of autotrophic C in the root system of walnut. Also, in maize, autotrophic C was allocated to roots with a delay of several days (Deléens et al., 1984). These differences are likely due (at least in part) to differences in the type of germination and associated differences in the function of the cotyledons. Maize and walnut have hypogeal germination and their cotyledon(s) have little or no photosynthetic activity, whereas sunflower has epigeal germination and the photosynthetic activity of its cotyledons was substantial. In sunflower the transition from heterotrophic to autotrophic growth of roots and shoots effectively resulted from the photosynthetic activity of the cotyledons: transition occurred between 4.5 and 7 DAI, before the first pair of leaves started to expand. Conversely, in walnut and maize transition to autotrophy could not start before the onset of export from the first true leaf. It is well established that sink to source transition in true leaves occurs at about the time when they reach one-third to two-thirds of their final size (Turgeon, 1989). Assimilate produced until that time is used within the growing leaf, explaining a delay in provision of autotrophic C to other plant parts such as roots. In contrast, reserve assimilate export from cotyledons started much earlier, before the onset of significant expansion (Figs 2, 3a and 4a), showing that the structural requirements for export of assimilate from the cotyledons are established before photosynthesis starts. Indeed, first autotrophic C was detected near-simultaneously in cotyledons, roots, and other shoot parts (Fig. 5, Table 1). At the same time (between 4.5 and 7 DAI), the loss of reserve-derived C from the seedlings decreased rapidly (Fig. 3a), suggesting that also respiration shifted from the consumption of heterotrophic to autotrophic substrates.

Some of the discrepancies between this study and other studies with respect to the times reported for the transition for shoot and root could also arise from differences in the methodology used to assess the transition. Deléens et al. (1984) and Maillard et al. (1994a) estimated the transition on the basis of the kinetics of the isotopic composition of C in the total biomass of shoot and root. These kinetics are sensitive to two factors: (1) the isotopic composition of the substrate entering the organ and (2) the relative growth rate of that organ. Thus, if there is a difference in relative growth rate of roots and shoots, then the kinetics of change of the isotopic composition differs even if the organs draw on isotopically identical substrate. Similarly, a delay in the appearance of autotrophic C in roots could also be related to an arrest of root growth. Here, we assessed the transition on the basis of the isotopic composition of the substrate supplied to the different plant parts in distinct (short) periods (Table 1), which is a measure of the transition that is independent of the relative growth rate of the different organs.

Is there an effect of environmental conditions on the transition from heterotrophic to autotrophic growth? Except for the present report, there are no published quantitative studies of effects of environmental conditions on the transition to autotrophic growth, minimizing opportunities for discussion. However, our results indicate clearly that the timing of first autotrophic C allocation to roots and shoots was hardly affected by CO2, although CO2 had an enormous effect on the rate of autotrophic C acquisition during the time of transition (Table 1, Fig. 5).

No effect of CO2 on seed reserve mobilization and use

This is the first study investigating the effects of CO2 on reserve mobilization and allocation in germinating plants, and it revealed no effect whatsoever of CO2 concentration on reserve mobilization, allocation and use. This result may not be surprising for the early phase of seedling development, when growth was completely dependent on reserves. However, the latter phase of reserve mobilization coincided with the onset of active photosynthesis in cotyledons, generating an opportunity for interactions/feedbacks between reserve- and photosynthesis-derived C fluxes. It has been suggested that mobilization of lipids in A. thaliana (Eastmond et al., 2000) and sunflower (Pfeiffer & Kutschera, 1996) may be under metabolic control. However, in the present study the timing and rate of C- and N-reserve mobilization was identical in the two treatments, providing no evidence for either negative or positive effects of photosynthesis and related metabolite fluxes on reserve mobilization.

Perhaps, absence of any effect might be related to a relatively large substrate demand for growth, as was evidenced by the strong growth response to CO2, when plants entered the autotrophic phase (see below). Also, the efficiency of mobilized C and N conversion in new biomass was identical in the two treatments. For C it averaged 0.62, which is high compared with the reported range of C use efficiencies (Dewar et al., 1998; Gifford, 2003; van Iersel, 2003), suggesting efficient use of mobilized biomass for seedling growth in both treatments.

Effects of CO2 on growth and allocation of autotrophic C

Carbon dioxide caused a dramatic stimulation of C assimilation (+150% relative to the subambient CO2 treatment). This effect was mainly caused by a higher unit leaf rate at elevated CO2 (+120%). But enhancement of leaf area expansion (up to +60%) also contributed to the effect, since radiation interception by the stands was limited by leaf area in the seedling stage. These responses are consistent with observations of others (for reviews see Pritchard et al., 1999; Poorter & Navas, 2003).

Clearly, seedling growth was C limited at subambient CO2. This was associated with increased C allocation to the shoot. Modifications of the shoot to root ratio by elevated CO2 were often observed in previous studies, but usually resulted from hastened ontogeny, rather than direct effects on allocation (Gebauer et al., 1996; Poorter & Nagel, 2000). Direct responses of allocation to stress or disturbance were often explained in terms of the ‘optimal partitioning’ or ‘functional equilibrium theory’ (McConnaughay & Coleman, 1999; Poorter & Nagel, 2000). These predict that allocation favours the plant part which intercepts the most severely growth-limiting resource. However, the mechanisms controlling allocation are still not well understood and much debated (Farrar & Jones, 2000). But whatever the mechanisms, the present observation of enhanced allocation towards the shoot at subambient CO2 is consistent with optimal allocation theory: plants at subambient CO2 may have stimulated C allocation to shoots because of C starvation. Yet, others have usually not observed a direct CO2 effect on allocation, shedding doubt on the general validity of the ‘optimal partitioning theory’ (Bernacchi et al., 2000; Poorter & Nagel, 2000). One reason why a direct CO2 effect on root vs shoot allocation was manifest in this study, but absent in others may be related to the comparatively weak effect of CO2 on allocation, which may render detection difficult when ‘common’ CO2 scenarios are used. Seedlings, growing close to the ground, may naturally experience both sub- and super-ambient CO2 concentrations.

In conclusion, this study demonstrated a near-simultaneous transition from reserve-dependent to current assimilation-driven growth of shoots and roots in sunflower, effected by the photosynthetic activity of cotyledons. This pattern of transition differs from that reported for species with hypogeal germination, in which transition to autotrophic growth of roots is delayed with respect to the shoot. Notably, although CO2 had a dramatic effect on the photosynthetic activity of cotyledons, this had no effect whatsoever on the concurrent rate of reserve mobilization, allocation and the efficiency of reserve-derived substrate conversion in new biomass.


This research was partially supported by the EU Human Potential Program (contract nr. HPRN-CT-1999–00059, NETCARB). The fine leadership provided by Jaleh Ghashghaie in the NETCARB project is gratefully acknowledged. Clément Piel and Arnoud Boom are thanked for discussion of a early version of this paper. Wolfgang Feneis, Angelika Ernst-Schwärzli, Brigitte Schilling, Anja Schmidt and Monika Breitsameter provided expert technical assistance.