C-isotope composition of CO2 respired by shoots and roots: fractionation during dark respiration?


Hans Schnyder. Fax: + 49 8161 713243; e-mail: schnyder@wzw.tum.de


The CO2 respired by leaves is 13C-enriched relative to leaf biomass and putative respiratory substrates (Ghashghaie et al., Phytochemistry Reviews 2, 145–161, 2003), but how this relates to the 13C content of root, or whole plant respiratory CO2 is unknown. The C isotope composition of respiratory CO2 (δR) from shoots and roots of sunflower (Helianthus annuus L.), alfalfa (Medicago sativa L.), and perennial ryegrass (Lolium perenne L.) growing in a range of conditions was analysed. In all instances plants were grown in controlled environments with CO2 of constant concentration and δ13C. Respiration of roots and shoots of individual plants was measured with an open CO2 exchange system interfaced with a mass spectrometer. Respiratory CO2 from shoots was always 13C-enriched relative to that of roots. Conversely, shoot biomass was always 13C-depleted relative to root biomass. The δ-difference between shoot and root respiratory CO2 was variable, and negatively correlated with the δ-difference between shoot and root biomass (r2 = 0.52, P = 0.023), suggesting isotope effects during biosynthesis. 13C discrimination in respiration (R) of shoots, roots and whole plants (eShoot, eRoot, ePlant) was assessed as e = (δSubstrate − δR)/(1 + δR/1000), where root and shoot substrate is defined as imported C, and plant substrate is total photosynthate. Estimates were obtained from C isotope balances of shoots, roots and whole plants of sunflower and alfalfa using growth and respiration data collected at intervals of 1 to 2 weeks. eplant and eShoot differed significantly from zero. eplant ranged between −0.4 and −0.9‰, whereas eShoot was much greater (−0.6 to −1.9‰). eRoot was not significantly different from zero. The present results help to resolve the apparent conflict between leaf- and ecosystem-level 13C discrimination in respiration.


The biomass of C3 plants is depleted in 13C relative to atmospheric CO2. This effect is caused mainly by the slower diffusion of 13CO2 relative to 12CO2 through stomata, and a strong preference of Rubisco for the light CO2 species (Farquhar, O’Leary & Berry 1982). However, the RuBP carboxylase reaction discriminates much more than CO2 diffusion. The overall discrimination during photosynthesis varies as a function of the relative limitations imposed by diffusion and photosynthetic capacity (Farquhar, Ehleringer & Hubick 1989). Therefore, 13C discrimination is a useful indicator of the nature of photosynthetic limitations.

However, fractionation during post-photosynthetic processes can also affect δ13C values of plant biomass. Thus, it has been shown that different chemical compounds differ in their C isotope composition: for instance, lipids and lignin are generally 13C-depleted, whereas non-structural carbohydrates, such as starch, are usually 13C-enriched relative to whole plant biomass (Park & Epstein 1961; Deines 1980; Ghashghaie et al. 2003). Differences in the C isotope composition of organs may result from differences in their chemical composition (Hobbie & Werner 2004).

13C fractionation during allocation and partitioning can also have an effect on the δ13C of respiratory CO2 (e.g. Jacobson et al. 1970). Large differences between the δ13C of biomass or putative substrates (δSubstrate) and respiratory CO2 (δR) have been reported (range: +10.3 to −8.1‰, cf. Ghashghaie et al. 2003), whereas absence of any effect has also been demonstrated (Lin & Ehleringer 1997). The δ-difference between biomass or substrate and respiratory CO2 was termed apparent fractionation (e′) by Ghashghaie et al. (2003), and is a wide-sense definition of 13C discrimination in respiration. It includes all isotope effects during allocation and partitioning which lead to a difference between the δ13C of photosynthetically fixed CO2 and that of respiratory CO2. Such effects may result from the non-statistical distribution of 13C in the glucose molecule (Rossmann, Butzenlechner & Schmidt 1991), isotope effects during the pyruvate dehydrogenase reaction, and/or the chemical identity of the substrate for respiration (e.g. lipids or sugars) (Ghashghaie et al. 2003). Tcherkez et al. (2003) observed a strong relationship between the δ of CO2 respired by leaves of Phaseolus vulgaris and the respiratory quotient (RQ; CO2 produced per O2 consumed), which was in accordance with switches between substrates of different chemical identity and δ13C. Similar observations were made by Jacobson et al. (1970) in studies of cyanide effects on respiration of potato tuber slices.

Knowledge of 13C discrimination in respiration is important, since it could modify the interpretation of biomass δ13C in terms of the nature of photosynthetic limitations. Further, if there is a systematic difference between the δ13C of CO2 fixed in photosynthesis and that of respiratory CO2, it could be used to partition the photosynthetic and respiratory component of net CO2 exchange (Yakir & Wang 1996; Bowling, Tans & Monson 2001). The finding of a substantial and systematic 13C-enrichment in respiratory CO2 relative to biomass and different putative substrates in leaves of several herbaceous C3 dicots reported by Ghashghaie and co-workers (Duranceau et al. 1999; Duranceau, Ghashghaie & Brugnoli 2001; Ghashghaie et al. 2001) is therefore of great interest. Yet, it is currently unknown if similar or divergent effects occur in roots, or at the whole plant level. Such information is essential for the interpretation of natural abundance of 13C signatures of respiratory CO2 emitted by plants and ecosystems.

One great difficulty for the proof of ‘true’13C discrimination in respiration is the natural and wide-spread occurrence of isotopic disequilibria in plants. These are caused by fluctuations in the isotopic composition of photosynthate, which arise from changes in photosynthetic 13C discrimination (ΔP; Farquhar et al. 1989), or alterations of the δ of source CO2 (δCΟ2). Changes in the δ13C of photosynthate (δP) are translated into a δ-change of respiratory CO2 only with a delay, because of the time needed for the new assimilate to arrive in metabolic pools (e.g. Ekblad & Högberg 2001, Pataki et al. 2003; Schnyder et al. 2003). Therefore, fluctuations of δP are not in phase with δR, causing the isotopic non-equilibrium. True 13C discrimination in respiration can only be verified and quantified if such isotopic disequilibria are suppressed in the experiment.

The aims of this study were: (1) to examine whether true 13C discrimination occurs during respiration of plants growing in controlled environment with a constant concentration and δ13C of CO2; (2) to analyse the relationship between the δ13C of shoot and root respiratory CO2; and (3) to explore the consequences of respiratory 13C discrimination for the interpretation of overall 13C discrimination at the whole plant level. To this end we took advantage of a series of experiments (n = 4) that had been conducted with different objectives, but which all incorporated measurements of the δ13C of root and shoot respiration. These included studies with perennial ryegrass (Lolium perenne L.), sunflower (Helianthus annuus L.), and alfalfa (Medicago sativa L.) grown in diverse, but constant environmental and C isotopic conditions. Thus, the generality (or variability) of the above relationships [(1)–(3)] could be assessed.


Plant material and growth conditions

The study reports on data from four independent experiments. These experiments were conducted with different specific objectives (see below), but all investigated features of the C economy of plants, including respiration of whole shoots (including stems and leaves) and roots.

In all experiments the plants were grown in constant conditions in controlled environment chambers (E15; Conviron, Winnipeg, Canada) which formed part of a mesocosm 13CO2/12CO2 gas exchange system described by Schnyder et al. (2003). The experiments included sunflower cv. Optisol, alfalfa cv. Planet, and perennial ryegrass cv. Racolta. All plants were grown in stands, formed by a usually dense (but see below) arrangement of plastic pots (made of polyvinylchloride) holding single plants. The pots (5 cm diameter, 35 cm depth) were filled with washed quartz sand. The plants received a modified Hoagland-type nutrient solution with nitrate as the sole nitrogen source several times per day (for modifications of the nutrient solution, see below). The pots were held in plastic containers (76 cm long, 56 cm wide, 32 cm high). Two containers were placed in each growth chamber.

In all experiments the air supplied to the growth chambers was generated from CO2-free air, which was mixed with CO2 of known δ13C using mass flow controllers (MFC). The rate of air supply to each chamber was 250–400 standard litres per minute. Thus, air flow into the chamber was equal to five to eight times the chamber volume per hour. Chambers received CO2 from one of two CO2 sources, either of mineral (relatively 13C-rich) or fossil-organic origin (relatively 13C-depleted; all from Messer Griesheim, Frankfurt, Germany). In all experiments the δ13C and concentration of CO2 inside the chambers was kept near constant by periodic adjustments of airflow and of the CO2 concentration at the chamber air inlet. Experiments I, II and III are described in Lötscher, Klumpp & Schnyder (2004). Details of experiment IV are given in Lattanzi, Schnyder & Thornton (2004).

Experiment I (sunflower)

This study investigated effects of plant–plant competition for light on C allocation, partitioning and respiration of sunflower plants. To this end stands of sunflower were established and grown at densities of 400 (high density) and 50 plants m−2 (low density). The high density stand was a 50 : 50 mix of two sub-populations: plants sown early and plants sown with a delay of 1 week. The former grew taller and, hence, dominant as a result of the head start, whereas the latter became subordinate. Only results from the former are reported here, as the subordinate plants exhibited very small respiration. Throughout the experiment a drip irrigation system supplied a half-strength Hoagland-type solution with 5.0 mmol L−1 nitrate-N to every pot every 4 h. Light was supplied by cool white fluorescent tubes and incandescent lamps, providing a 16 h light period with a photosynthetic photon flux density (PPFD) of 700 µmol m−2 s−1 at the top of the canopy. The temperature was set at 20/16 °C and relative humidity at a constant 70% during the light and dark periods. The concentration and δ13C of CO2 in the chambers was kept constant near 350 µL L−1 and −1.8‰. Respiration was measured and plants sampled at weekly intervals, between 3 and 5 weeks from germination.

Experiment II (alfalfa)

This experiment was performed to assess effects of competition on C allocation, partitioning and respiration. This was done with singly potted individuals of alfalfa growing in dense pure stands (400 plants m−2). Stands contained a 50 : 50 mixture of plants originating from two pretreatments: plants established at a moderate irradiance (350 µmol m−2 s−1 PPFD at pot height during a 16 h day for 4 weeks from sowing) and cut to a stubble height of 8 cm, and plants established at low irradiance (100 µmol m−2 s−1 PPFD) and cut to a stubble height of 3 cm, at the time of assembling the stands. The containers holding stands were flooded twice a day for 1 h with a half-strength modified Hoagland solution containing 7.5 mmol L−1 nitrate-N. Once a week the pots were flushed with water. Growth chambers were adjusted to 22/18 °C day/night temperatures, 75% relative humidity, and 350–500 µmol m−2 s−1 PPFD at canopy height during the 16 h photoperiod. It was expected that the pretreatments would result in plants of different height. However, all plants reached the top of the canopy. The CO2 concentration in all chambers was held constant near 350 µL L−1. Two chambers were supplied with CO2 having a δ13C of −2.0‰, whereas the other two received CO2 with a δ13C near −43.7‰. Respiration was measured and plants sampled at weekly intervals starting at 2 until 4 weeks after assembling the stands.

Experiment III (alfalfa)

This study was conducted to simulate effects of light and N competition on C allocation, partitioning and respiration. To this end 3-week-old plants of alfalfa were cut to a 3-cm stubble height and assigned and exposed to one of four combinations of high or low irradiance (I+/I–: 610 or 90 µmol m−2 s−1 PFD at canopy height), and high or low nitrogen supply (N+/N–: 7.5 or 1.5 m m L−1 of nitrate-N in the nutrient solution) for 5 weeks. In all treatments the concentration and δ of CO2 in the chambers was kept constant near 350 µL L−1 and −2.4‰. Respiration was measured and plants sampled at 5 weeks after imposition of treatments.

Experiment IV (ryegrass)

In this study we analysed effects of light competition on growth and recovery from defoliation in ryegrass and Paspalum dilatatum Poir. Briefly, a 50 : 50 mixed stand with plants of the two species was grown at 15/14 °C day/night, a vapour pressure deficit of 0.5/0.3 kPa, and 550 µmol m−2  s−1 PPFD at the top of the canopy during the 12 h photoperiod. All plants grew singly in pots. Stands were flooded four times a day for 30 min with a half-strength modified Hoagland solution containing 7.5 mmol L−1 nitrate-N. Industrial CO2 with δ13C −47.0‰ was used as a CO2 source. The concentration and δ13C of CO2 inside the chamber was kept near 300 µL L−1 and −47.0‰ by periodic adjustments of airflow and of the CO2 concentration in the inlet air. Respiration measurements were performed and plants sampled 9 week after sowing. Paspalum plants were too small and respired very little. Thus, only the ryegrass respiration data are presented.

Respiration measurements and isotope analysis

Respiratory 13CO2/ 12CO2 exchange system

A detailed description of the respiration measurement system was recently given by Lötscher et al. (2004). Briefly, root and shoot respiration as well as the δ13C of root and shoot respiratory CO2 of individual plants was measured in an open gas exchange system. The system included four cuvettes, each of which consisted of a tube and a top and bottom plate (all made of polyvinylchloride) (Fig. 1), which could be opened and closed quickly to insert a pot. A bushing that matched exactly the size of the pot was glued into the bottom plate. A similar system was used to seal the bottom of the pot. Rubber seals and vacuum grease were used to ascertain that the cuvettes were gas tight. The system was connected to an infrared gas analyser (IRGA) and a continuous-flow stable isotope-ratio mass spectrometer (CF-IRMS) via Teflon tubes. Air with known constant δ13C and concentration of CO2 (200 µL L−1) was supplied to cuvettes at a rate of 0.3–1.0 L min−1 after passage of a humidifier. Air flow was controlled by MFCs. Cuvettes were arranged in a plant growth cabinet held at the same temperature as plant growth chambers. Each cuvette had two outlets: one in the shoot section on the opposite end of the inlet, and the other at the bottom of the pot which enclosed the root compartment. Air in the shoot compartment was ventilated by a fan. About 0.15–0.45 L min−1 of the cuvette air was drawn through the root compartment with a gas-tight Teflon-lined peristaltic membrane pump. The air was then dried and the flow to a multiway valve block (sample air selector, SAS) controlled by a MFC. The remaining air from the shoot compartment was directly conveyed to the SAS. A reference air line (0.9 L min−1) was also connected to the SAS. The SAS sequentially sampled the reference air line and the different sample air lines and fed the air to the IRGA and IRMS as described in Schnyder et al. (2003). A full measurement cycle of all four cuvettes was completed in about 30 min. Readings of the IRGA and IRMS were continuously stored in files. The standard deviation associated with the measurement of δ13C and CO2 concentration of the reference air CO2 was typically < 0.15‰ and < 1 µL L−1.

Figure 1.

Scheme of respiration cuvette. For further details see Materials and methods section.


Plants were removed from the growth chambers at the end of the light period, pots flushed with 0.5 L water and then rinsed with nutrient solution, which was previously aerated with CO2-free air for 3 d. Thereafter, the pots were enclosed in the respiration cuvettes. Air injection in the shoot compartment was initially set at 3 L min−1. When the CO2 concentration in the exit of the shoot compartment had decreased to a near-constant concentration, the drainage of the root compartment was opened and the air fitting installed. Air was then drawn through the root compartment at 1 L min−1 until the CO2 concentration at the root exit became near constant. These procedures aimed at removing all extraneous CO2 from the shoot and root compartment as quickly as possible, and thus minimize the time lag between insertion of a pot and collection of first reliable δR data. System tests demonstrated that up to 1 h was required to purge the system from all extraneous CO2 (Fig. 2). Tests also revealed that the system did not fractionate 13C: when pots with clean wet sand were inserted in the system, the δ13C of CO2 at the root outlet (following purging of the system) did not differ from that at the cuvette inlet (Fig. 2). When CO2 concentration readings had become near stable the flow rates were set corresponding to the plant size. First measurements of shoot and root respiration and δ were obtained at about 1 h after the plants were removed from the stands. Respiration was measured throughout the regular duration of the dark period. The C isotope composition of CO2 respired by shoots and roots was relatively constant during a dark period (Fig. 3). That is, any trends over time of the δ of CO2 respired by shoots and roots were statistically non-significant. Flux-weighted mean δ13C of respiratory CO2 collected throughout the dark period are presented in this paper.

Figure 2.

Time course of the concentration (a) and δ13C (b) of CO2 entering the respiration cuvette (•), and leaving it at the root outlet (○,□) of two pots. Measurements were done with pots containing clean sand. Time refers to the closure of the system when the top plate (see Fig. 1) was fixed following insertion of the pots. Pots were flushed either with CO2-free nutrient solution (□), or normal nutrient solution (○) before inserting them into the respiration chamber. The δ13C of CO2 at the chamber inlet and outlets did not differ after 1 h: •−49.03‰ (± 0.13), ○−48.95‰ (± 0.10), □−48.83‰ (± 0.09).

Figure 3.

Time course of the δ13C of CO2 respired by shoots (○, □) and roots (•, ▪) of sunflower during the regular dark period of the day. Plants were grown at low (○, •) and high (□, ▪) density (experiment I). Means ± SE.

Plant sampling, and elemental and isotope analysis

Plants were separated into shoot [including all leaf and stem tissue (reproductive organs had not formed)] and root fractions immediately after the termination of respiration measurements. The two fractions were dried (30–60 min at 100 °C, and then at 65 °C for 48 h) in a forced draught oven, weighed, and ground to a fine powder in a ball mill. The C content and δ13C of 1 mg aliquots of samples was determined by a elemental analyzer (Carlo Erba NA1110, Milano, Italy) interfaced to the IRMS.

All isotope data were expressed in the conventional δ13C -notation (in ‰) relative to a VPDB-gauged standard as δsample = 1000(RSample/Rstandard − 1) (McKinney et al. 1950), where R is the 13C/12C ratio in the sample or the standard.

Data analysis: gas exchange parameters, growth analysis, allocation model, 13C fractionation

Respiration of roots (RRoot) and shoots (RShoot) and the δ13C of the respiratory CO2 evolved from roots (δR Root) and shoots (δR Shoot) were obtained as

Rshoot = Fin(Cin − CS out)((1a))
Rroot = FR out(CS out − CR out)((1b))
δR Shoot = [(δS out Fin CS out) − (δin Fin Cin)]/RShoot((2a))
δR Root = [(δR out FR out CR out) − (δS out FR out CS out)]/RRoot((2b))

with F the flux of air at standard humidity (mol s−1), C the concentration of CO2 in air (mol mol−1), and subscripts ‘in’, ‘R out’ and ‘S out’ designating the respiration chamber inlet, root compartment outlet and shoot compartment outlet, respectively.

The C isotope composition of C accumulating in roots (δRoot) and shoots (δShoot), of C imported in roots (δI Root) and shoots (δI Shoot), as well as of the total photosynthate (δP) produced by individual plants during a given interval i − 1 to i was estimated using the data from the periodic sampling and C elemental and isotope analysis of roots and shoots, and of the concurrent respiration measurements in experiments I and II. Thus, for δRoot and δShoot

δShoot = ( δShoot ti Cshoot ti − δShoot ti − 1 Cshoot ti − 1)/ (Cshoot ti − Cshoot ti − 1) and((3a))
δRoot = ( δRoot ti Croot ti − δRoot ti − 1 Croot ti − 1)/(Croot ti − Croot ti − 1)((3b))

with δShoot ti, Cshoot ti, δRoot ti, and Croot ti representing the C isotope composition and C mass of shoots and roots at sampling date i, and δShoot ti−1, Cshoot ti−1, δRoot ti−1, and Croot ti−1 the same parameters for the previous sampling date.

The δ of C imported in roots and shoots was calculated as the weighted average δ13C of C accumulated and respired in root and shoots

δI Shoot = [δShoot (dCShoot/dt) + δR Shoot RShoot)]/IShoot and((4a))
δI Root = [δRoot (dCRoot/dt) + δR Root RRoot)]/IRoot((4b))

with IShoot and IRoot, the rate of C import in shoots and roots, as obtained from the rates of C accumulation and respiration: Ishoot = dCShoot/dt + RShoot, and Iroot = dCRoot/dt + RShoot, respectively. dCShoot/dt was calculated as dCShoot/dt = (Cshoot ti − Cshoot ti−1)/[ti − (ti − 1)]. dCRoot/dt was obtained similarly.

The estimate of R included respiration during darkness and daylight hours, where the latter was estimated from respiration in darkness and temperature, assuming a Q10 = 2, for short-term, that is, diurnal fluctuations of R. This was verified for sunflower and alfalfa, by continuous measurements of root respiration (data not shown). It was further assumed (a) that the δ13C of respiratory CO2 was the same in light and darkness, and (b) that losses of C in the form of exudates and volatile organic C were negligible. Dead tissue (mainly leaves) was included in the sampling.

The δ13C of total photosynthate (δP) produced by the plant during the interval i − 1 to i, was obtained as

δP = (δI Shoot Ishoot + δI Root IRoot)/Pn,(5)

with Pn = Ishoot + IRoot.

Apparent fractionation (e, ‰) during respiration in the shoot (eShoot) and root (eRoot) was estimated according to the relationship e = (δSubstrate − δProduct)/(1 + δProduct/1000), where the δ13C of the substrate in the root or shoot was equated with that of total C import. Thus

eShoot = (δI Shoot − δR Shoot)/(1 + δR Shoot/1000) and((6a))
eRoot = (δI Root − δR Root)/(1 + δR Root/1000)((6b))

Accordingly, apparent fractionation during respiration at the whole plant level (ePlant) was obtained as

ePlant = (δP − δR Plant)/(1 + δR Plant/1000)(7)

with δR Plant the respiration-weighted average δ13C of root and shoot respiration.

Statistical analysis

Analyses of variance and t-test of paired samples of δ13C bulk and respired C of shoot and root, respectively, were performed using procedure anova and t-test for paired samples of STATISTICA (Statistica 5; Statsoft, Inc., Tulsa, OK, USA).


Relationships between δ13C of shoot and root respiratory CO2 and δ13C of biomass

Shoot respiratory CO2 was always 13C-enriched relative to shoot biomass (+2.6‰ ± 0.4) (Table 1). In contrast, root respiratory CO2 was always 13C-depleted relative to root biomass (−2.7‰ ± 0.4). Because of the opposing relationships in shoots and roots, the discrepancy was much smaller at the whole plant level: respiratory CO2 evolved by the whole plant was on average 1.1‰ enriched in 13C relative to total plant C.

Table 1. δ-differences among respiration and biomass components of sunflower, alfalfa and perennial ryegrass: difference between shoot biomass-δ13C and shoot respiratory CO2-δ13C (δShoot − δR Shoot), difference between root biomass-δ13C and root respiratory CO2-δ13C (δRoot − δR Root), difference between shoot and root biomass-δ13C (δShoot –δRoot), difference between shoot respiratory CO2-δ13C and root respiratory CO2-δ13C (δR Shoot − δR Root), and difference between whole plant biomass-δ13C and whole plant respiratory CO2-δ13C
 δShoot − δR ShootδRoot − δR ΡοοτδShoot − δRootδR Shoot − δR RootδPlant − δR Plant
  1. Means and (SE). Differences are statistically different from 0 at P < 0.05, except where indicated by n.s.

 Low density stand−2.70 (0.25)0.46 (0.37)−0.80 (0.19)2.38 (0.46)−1.83 (0.16)
 High density stand−2.02 (0.28)2.00 (0.30)−0.91 (0.09)3.11 (0.42)−1.24 (0.21)
 Low light pretreatment−2.34 (0.56)2.95 (0.91)−1.59 (0.34)3.70 (0.79) 0.03 (0.59) n.s.
 High light pretreatment−1.90 (1.02)1.52 (0.29) n.s.−1.00 (0.22)2.42 (1.20)−0.81 (0.49) n.s.
 High light/high nitrogen−1.06 (0.20)3.73 (0.42)−1.03 (0.15)3.76 (0.18) 0.16 (0.28) n.s.
 High light/low nitrogen−1.60 (0.41)2.84 (0.38)−0.95 (0.38)3.48 (0.10)−0.18 (0.29) n.s.
 Low light/high nitrogen−5.64 (1.12)2.39 (1.15)−1.38 (0.10)6.65 (2.07)−3.94 (0.79)
 Low light/low nitrogen−2.18 (0.58)2.68 (0.64)−1.19 (0.06)3.68 (1.15)−0.76 (0.32) n.s.
 Perennial ryegrass−3.74 (0.37)5.39 (0.19)−1.53 (0.53)7.60 (0.82)−0.97 (0.32)

The relationship between the δ13C value of respiratory CO2 and biomass varied among species and treatments (Table 1). Since the different species were grown in different conditions (light, N nutrition, temperature), it is unclear whether species’ differences had a genetic component. However, the large variation in the relationship between the δ13C of biomass and respiratory CO2 in the different treatments of alfalfa suggested a strong influence of growth conditions.

There was also a striking difference between the δ13C values of CO2 respired by shoots and roots, with shoot respiratory CO2 always 13C-enriched relative to root respiratory CO2 (Table 1). On average of species and treatments the δ13C of shoot respiratory CO2 was 4.1‰ enriched in 13C relative to that of roots. Again, there was substantial variation in this relationship (2.4 to 7.6‰). The δ-difference between shoot and root respiratory CO2 was largest in ryegrass, and smallest in sunflower, but almost the entire range of variation was found among the different treatments of alfalfa. The δ13C values of shoot and root biomass also differed, and this difference was opposite to that of the respiratory CO2 (Table 1). That is, the shoot biomass was always 13C-depleted relative to the root. On average, this difference was −1.2‰. Again, the relationship was variable. Importantly, the δ-difference between shoot and root biomass was inversely related to the δ-difference between shoot and root respiratory CO2 (r2 = 0.52, P = 0.023) (Fig. 4).

Figure 4.

δ-difference between CO2 respired by shoots and roots versus δ-difference between biomass C of shoots and roots. Results from experiments with sunflower (□), alfalfa (○) and perennial ryegrass (▵). Means ± SE.

13C discrimination in respiration of shoots, roots and whole plants

In experiments I (sunflower) and II (alfalfa) respiration was measured and plants sampled at intervals, thus allowing construction of C isotope balances for periods of 7–14 d. These data indicated that the δ13C of C concurrently accumulating in roots and shoots did not differ significantly (Table 2). Moreover, the δ13C of C imported into roots and shoots of alfalfa did not differ significantly, although they did in sunflower. In the latter the δ13C of C imported into roots was approximately 1‰ depleted relative to that in the shoot (P < 0.01). In both experiments the respiratory CO2 evolved from roots was consistently 13C-depleted relative to that of shoots.

Table 2.  Mean C isotope composition (δ13C) of total C accumulated, imported and respired during 1- to 2-week intervals in shoots and roots of sunflower and alfalfa, and respiratory 13C discrimination in shoots, roots and whole plants
high density
low density
  1. SE values are given in parentheses. The δ13C of C accumulation, respiration and import were calculated by isotopic mass balance as explained in Materials and Methods. Data from experiments I and II. Data from the two pretreatments in experiment II were merged by δCO2 during growth. **, *and n.s. indicate significance at the P < 0.01, P < 0.05 level and non-significance for the difference between shoot and root δ13C, or for e being different from zero.

δ13C of accumulated CShoot−23.40 (0.48)−24.33 (0.23)−23.33 (0.30)−66.91 (0.31)
Root−22.03 (0.84)−24.39 (0.23)−22.40 (0.26)−67.10 (1.93)
δ13C of imported C (δI)Shoot−22.47 (0.19)−23.19 (0.17)−22.94 (0.14)−65.89 (0.13)
Root−24.05 (0.36)−23.82 (0.05)−23.12 (0.28)−66.86 (0.79)
δ13C of respired C (δR)Shoot−21.71 (0.03)−21.42 (0.08)−22.35 (0.31)−64.05 (0.41)
Root−24.94 (0.05)−23.50 (0.07)−23.70 (0.30)−67.36 (0.52)
FractionationeShoot*−0.77 (0.20)**−1.81 (0.24)−0.61 (0.36)**−1.97 (0.48)
eRoot0.92 (0.35)−0.33 (0.12)*0.60 (0.12)0.30 (0.30)
ePlant−0.42 (0.21)**−1.35 (0.20)−0.04 (0.20)*−0.68 (0.17)

Apparent 13C discrimination in respiration of shoots averaged −1.3‰, meaning that respiratory CO2 was 13C-enriched relative to imported C. The 13C discrimination in respiration of roots was insignificant, with one exception: the small (0.6‰) 13C discrimination in root respiration of alfalfa growing in the presence of CO2 with a δ13C −4‰ was statistically significant. However, plants growing under identical conditions but exposed to CO2 with a δ13C of −47‰ did not demonstrate 13C discrimination in root respiration. Clearly, if 13C discrimination did occur in root respiration, then it was small.

At the whole plant level 13C discrimination in respiration was slightly negative (average −0.6‰). Thus, the total respiratory CO2 produced by plants was slightly 13C-enriched. This effect was significant in two of the data sets (sunflower at high density, and alfalfa growing in δCΟ2−47‰), and over all experiments (−0.7‰, P < 0.01).


Contrasting respiratory fractionation in shoots and roots

This study revealed a contrast in apparent respiratory 13C fractionation in roots and shoots of three herbaceous species: a annual and a perennial dicot, and a monocot. This phenomenon has not been studied and reported before. CO2 respired by roots was always 13C-depleted relative to that of shoots in a large range of environmental conditions, indicating that the effect was not the result of singular experimental conditions, but had a more general validity.

Importantly, the fractionation effects observed in this study were not a result of isotopic disequilibria between current photosynthate and respiratory substrates. Such disequilibria commonly occur in natural situations due to fluctuations in the δ13C of photosynthate and delays in turn over of the respiratory pools by the products of current photosynthesis (Pataki et al. 2003). The δ13C of CO2 and environmental conditions were constant in each experiment, thus minimizing opportunities for the formation of isotopic disequilibria. Isotopic disequilibria between current photosynthate and respiratory pools could still arise in constant conditions, if there were ontogenetic shifts in photosynthetic discrimination. However, we found no indication for such effects in our data: δ13C of biomass and of imported C were constant during the experimental periods. Furthermore, the fractionation effects at the level of the whole plant and shoot persisted, when isotope balances were calculated for discrete periods.

The relative 13C-enrichment of CO2 respired by shoots is in agreement with the findings of Ghashghaie and co-workers (Ghashghaie et al. 2003), which demonstrated that CO2 respired by leaves in the dark was 6‰ enriched compared to leaf sucrose in intact French bean (Phaseolus vulgaris) (Duranceau et al. 1999). Similar results were also obtained in tobacco (Nicotiana sylvestris) and sunflower (Helianthus annuus) though the respiratory CO2 was slightly less enriched in 13C (Ghashghaie et al. 2001). Ghashghaie and co-workers also observed variability in 13C-enrichment of respiratory CO2, which could be induced in the short term and was correlated with changes in the respiratory quotient (RQ). This result suggested switches between different substrates for respiration, and was also consistent with the δ13C of putative substrates extracted from the same leaves (Tcherkez et al. 2003). Our results confirm and extend the finding of Ghashghaie in that they show 13C-enrichment of respiratory CO2 evolved from shoots in other species, and demonstrate its occurrence in tightly controlled constant environmental and isotopic conditions.

The definition of respiratory fractionation (e) used in the present study differs somewhat from that of Ghashghaie and co-workers. They estimated e as the δ-difference between respiratory CO2 and total (leaf) tissue C or putative substrates (e.g. sucrose), which might already be modified somewhat by the respiratory 13C discrimination. Here, we estimate e from the relationship of respiratory CO2 with total photosynthates available for growth, storage and maintenance in the relevant plant part or whole plant. We feel that this definition is appropriate to the scale of this study, particularly since the identity and activity of the different respiratory activities in the shoot and root are uncertain, and the true chemical identity of the substrates and their compartmental distribution is unknown.

The present results suggest a connection between discrimination in respiration and biosynthesis. A corollary of respiratory 13C fractionation is that it changes the 13C signature of the remaining material in the opposite direction (Ghashghaie et al. 2003). In the present study the contrast in the isotopic signature of shoot and root respiratory CO2 was opposite to the contrast of the C isotope signatures of shoot and root biomasses: whereas the respiratory CO2 of shoots was 13C-enriched relative to that of roots, the biomass of shoots was 13C-depleted relative to that of roots. Importantly, the difference between shoot and root respiratory δ13C was negatively correlated with the difference between shoot and root biomass δ (Fig. 4). This relationship passed near the origin and had a slope of approx. −3.6 (if forced through the origin), meaning that for every 1‰ change in the relationship between root and shoot biomass δ13C, the δ-difference between shoot and root respiratory CO2 changed by 3.6‰. This relationship is consistent with the view that respiratory discrimination is associated with biosynthesis, and a growth yield (biomass-C produced per unit C consumed) of approximately 0.78. This was in agreement with experimental estimates of the growth yield of these plants (0.72 and 0.8 for alfalfa I+/N+, I–/N+ and sunflower at high density, respectively) by Lötscher et al. (2004) and theoretical estimates (Penning de Vries et al. 1974; Johnson 1990; Amthor 2000; Cannel & Thornley 2000). The absence of discrimination in respiration observed in studies with protoplasts by Lin & Ehleringer (1997) could be linked to an absence of biosynthetic activity of the protoplasts.

Support for a causal relationship between 13C discrimination in respiration and biosynthetic processes is also obtained from the analysis of other data. Thus, others have noted a similar difference between root and shoot biomass δ13C (Bradford, Sharkey & Farquhar 1983; Yoneyama & Ohtani 1983; Viktor & Cramer 2003; Hobbie & Werner 2004). Calculations using the chemical composition of roots and shoots as reported by Poorter & Bergkotte (1992) together with the (average) deviation of C isotope signatures of chemical compound classes relative to bulk C as compiled by Gleixner et al. (1993) and Schmidt & Gleixner (1998) give a δ-difference of 1 and 2‰ between the shoot and root for fast- and slow-growing herbaceous species.

However, other mechanisms than respiratory fractionation in biosynthetic processes could also contribute to the difference in respiratory 13C fractionation between shoots and roots. For instance, re-assimilation of respiratory CO2 in roots by PEPc could add to the contrast of δ13C in respiratory CO2 released by roots and shoots, because the substrate for PEPc (HCO3) is 13C-enriched relative to the CO2 pool from which it is formed. Thus, the respiratory CO2 escaping from the roots would be 13C-depleted relative to the respiratory substrate, increasing the δ-difference between respiratory CO2 evolved from shoots and roots. This would also increase the difference in δ13C between biomass C and respiratory CO2 of roots. Moreover, if the 13C-enriched organic acids were transported to and decarboxylated in the shoot, the contrast between shoot and root respiratory CO2δ13C would be further enhanced. Support for such a mechanism comes from the observation of a stimulated translocation of organic acids from roots to shoots by nitrate nutrition (Cramer & Lips 1995), and a RQ of approximately 1.2 for dark respiration of leaves in a significant fraction of the data reported by Tcherkez et al. (2003). However, more detailed studies of PEPc function in roots and shoots are needed to verify a role of PEPc in determining the δ13C of respiratory CO2 released from shoots and roots.

Allocation also had some effect on the δ13C of respiratory CO2 of roots and shoots: on average of all treatments, C imported into the root was 13C-depleted by 0.8‰ relative to the C imported into the shoot (P < 0.05; Table 2). This effect could result from different allocation patterns of C fixed by well-lit and shaded leaves, and a difference between the two leaf categories in the δ13C of photosynthate produced by them. In a 14C-labelling study with tomato Shishido, Kumakura & Nishizawa (1999) observed that lower leaves allocated relatively more C to roots than upper leaves; whereas upper leaves allocated relatively more C to the shoot apex. Photosynthetic 13C-discrimination in sun leaves is less than in shaded leaves, leading to more negative δ13C of assimilate produced by shaded than by well-lit leaves (LeRoux et al. 2001; Baldocchi & Bowling 2003).

Plant level respiratory 13C discrimination

The results demonstrated that, if the proportions of the CO2 respired by different plant parts and proportions of carbon allocated to these different parts were considered in a mass balance, there was a small plant-level respiratory 13C discrimination. Plant-level respiratory 13C discrimination was small, because of the opposing behaviour of shoots and roots. Still the effect was significant (P < 0.01) and averaged −0.7‰ meaning that respiratory CO2 was somewhat 13C-enriched relative to photosynthetic products. This effect was nearly identical with that of perennial ryegrass stands determined by other means, i.e. by continuous 13CO2/12CO2 gas exchange measurements of whole stands in light and darkness (Schnyder et al. 2003). Since respiration consumed about 45% (± 0.04) of total photosynthate in the present experiments, the effect of respiratory 13C discrimination on the δ13C of total biomass was small: on average of the different experiments it would have shifted the δ13C of biomass by less than 0.3‰ relative to the primary photosynthetic products. One implication of this finding is that neglecting respiratory 13C discrimination leads to only small errors in the interpretation of biomass-δ13C in terms of the contribution of stomatal and photosynthetic capacity limitations on photosynthesis, if it is based on whole plant δ13C. Because of its small magnitude the usefulness of ePlant for partitioning stand-scale respiratory and photosynthetic CO2 fluxes during isotopic equilibrium periods seems greatly limited. So, partitioning based on natural abundance isotope signatures would need to rely on isotopic disequilibria which naturally arise from changing environmental conditions, but which may vary greatly in magnitude. Yet, such environmental perturbations might entrain parallel changes in allocation, and shifts in the relative magnitudes of root and shoot respiration. Such effects could contribute to changes in the δ13C of plant level respiration via their (shoot and root) specific e. But whether or not such effects really occur remains to be tested.

In conclusion, the finding of Ghashghaie and co-workers of a substantial (negative) 13C discrimination in respiration of leaves is validated by this study at the whole shoot level. However, the effect was strongly counterbalanced by roots which tended to produce 13C-depleted respiratory CO2, thus yielding only a small respiratory 13C discrimination at the whole plant level. The present results help to resolve the apparent conflict between leaf- and ecosystem-level 13C discrimination in respiration.


This work was supported by the Deutsche Forschungsgemeinschaft (SFB 607). Jaleh Ghashghaie is thanked for valuable comments on an earlier draft of the paper. We also thank two anonymous reviewers for thoughtful comments.