δ13C of CO2 respired in the dark in relation to δ13C of leaf carbohydrates in Phaseolus vulgaris L. under progressive drought

Authors


Jaleh Ghashghaie E-mail: jaleh.ghashghaie@eco.u-psud.fr

Abstract


A, leaf net CO2 assimilation
a, fractionation against 13C for CO2 diffusion through air
b, net fractionation against 13C during CO2 fixation by Rubisco and PEPc
δ13C, carbon isotopic composition
Δ, discrimination against 13C during CO2 assimilation
d, the term including the fractionation due to CO2 dissolution, liquid phase diffusion and also discrimination during both respiration and photorespiration DW, leaf dry weight
dδ13C, the difference between CO2 respired in the dark and plant material in their carbon isotope composition
dΔ, variation in modelled discrimination at a given pi/pa relative to a reference value at pi/pa = 0·7
FW, leaf fresh weight
gc, leaf conductance to CO2 diffusion
HPLC, high-performance liquid chromatography
LMA, leaf mass per area
pa, ambient partial pressure of CO2
pi, intercellular partial pressure of CO2
PEPc, phosphoenolpyruvate carboxylase
PPFD, photosynthetic photon flux density
RPDB, 13C/12C ratio of standard PDB
RS, 13C/12C ratio of sample
Rubisco, ribulose 1,5 bisphosphate carboxylase-oxygenase
RWC, leaf relative water content
SW, leaf saturated weight
VPD, vapour pressure deficit

The variations in δ13C in both leaf carbohydrates (starch and sucrose) and CO2 respired in the dark from the cotyledonary leaves of Phaseolus vulgaris L. were investigated during a progressive drought. As expected, sucrose and starch became heavier (enriched in 13C) with decreasing stomatal conductance and decreasing pi/pa during the first half (15 d) of the dehydration cycle. Thereafter, when stomata remained closed and leaf net photosynthesis was near zero, the tendency was reversed: the carbohydrates became lighter (depleted in 13C). This may be explained by increased pi/pa but other possible explanations are also discussed. Interestingly, the variations in δ13C of CO2 respired in the dark were correlated with those of sucrose for both well-watered and dehydrated plants. A linear relationship was obtained between δ13C of CO2 respired in the dark and sucrose, respired CO2 always being enriched in 13C compared with sucrose by ≈ 6‰. The whole leaf organic matter was depleted in 13C compared with leaf carbohydrates by at least 1‰. These results suggest that: (i) a discrimination by ≈ 6‰ occurs during dark respiration processes releasing 13C-enriched CO2; and that (ii) this leads to 13C depletion in the remaining leaf material.

INTRODUCTION

Plants discriminate against 13C during photosynthetic CO2 fixation. The carbon isotope discrimination (Δ) by plants involves both physical and biochemical processes. Farquhar, O’Leary & Berry (1982b) developed a theoretical model which predicts a linear relationship between Δ and the ratio of intercellular to atmospheric partial pressures of CO2 (pi/pa) in C3 plants:

image

where a (= 4·4‰) is the fractionation due to diffusion of CO2 from ambient air into the leaf and b (≈ 28‰) is the net fractionation during carboxylation by both the primary carboxylating enzymes Ribulose 1,5 bisphosphate carboxylase-oxygenase (Rubisco) and phosphoenolpyruvate carboxylase (PEPc) (O’Leary 1981).

According to this model, all factors decreasing pi/pa should also decrease Δ. This has been validated for many species by: (i) instantaneous ‘on-line’ measurements of carbon isotope discrimination in an open gas-exchange system during leaf photosynthesis (Evans et al. 1986); and (ii) time-integrated net discrimination determined using carbon isotope compositions in leaf dry matter (Farquhar et al. 1982a) or in leaf carbohydrates (Brugnoli et al. 1988). In all these experiments, pi/pa has been varied using different photosynthetic photon flux densities (PPFD) or vapour pressure deficits (VPD). Lauteri, Brugnoli & Spaccino (1993) also obtained a similar relationship between carbon isotope discrimination (determined on leaf soluble sugars) and pi/pa in four sunflower genotypes subjected to a progressive drought. Indeed, under water deficit, when stomata close, pi/pa and, thus, Δ, decrease leading to the formation of heavier carbohydrates (enriched in 13C).

Equation 1 is an approximation of the more complete theory described by Farquhar et al. (1982b) and Evans et al. (1986) and abbreviated by Hubick, Farquhar & Shorter (1986) as follows:

image

where d includes the fractionation due to CO2 dissolution, liquid phase diffusion and also discrimination which may occur during respiration and photorespiration. The magnitudes of the discrimination during both photorespiration and respiration are uncertain. Nevertheless, d has been shown to be very small in wheat leaves (Evans et al. 1986) suggesting that a possible discrimination during photorespiration and respiration cannot significantly modify the net discrimination during on-line measurements in the light. Moreover, the discrimination during both photorespiration and respiration is difficult to measure. Thus, the simple model is usually used.

However, the contribution of both photorespiration and respiration to net discrimination will depend not only on the fractionation during decarboxylation, but also on the rates of decarboxylation relative to CO2 fixation. There could also be a possible effect of respired CO2 refixation. Because a relatively high CO2 drawdown in the assimilation chamber is needed for accurate on-line measurements of Δ, leaf net photosynthesis was always maintained high enough, never reaching values near zero. Recently, however, Gillon & Griffiths (1997) showed that the difference between Δ observed on-line and Δ predicted by the simple model of Farquhar (Eqn 1) increases when photorespiration or respiration makes a larger contribution to net CO2 exchange, i.e. at low light intensities or at high oxygen concentration. They suggested that the carbon isotope composition of respired CO2 was different from that of fixed CO2 and that this difference was due to a fractionation during decarboxylation and/or to the carbon isotope composition of the substrates used for photorespiration or respiration which might be different from that of total leaf material.

There are only a few old and contradictory data in the literature about the carbon isotope composition (δ13C) of respired CO2 by plants. Respiratory CO2 (collected in a CO2-free atmosphere in the dark) has been reported to be 13C-enriched (1–8‰) or 13C-depleted (1–4‰) compared with leaf or whole plant dry matter, varying within species [see O’Leary (1981) for a review]. O’Leary (1981) also suggested that at least some of these variations may be due to the source of CO2 used for respiration which might have a carbon isotope composition different from that of whole plant material. One can also suggest that the enrichment or depletion of respired CO2 in 13C compared with leaf dry matter or leaf carbohydrates may be due to a fractionation during dark respiration. However, a recent work on mesophyll protoplasts (isolated from mature leaves of common bean, a C3 plant, and corn, a C4 plant), incubated in the dark in culture medium containing different carbohydrate substrates (fructose, glucose or sucrose) with known δ13C values, showed no fractionation associated with metabolism of any of these substrates (Lin & Ehleringer 1997).

Moreover, there are no data in the literature about the carbon isotope composition of leaf material in advanced stages of dehydration cycles. Indeed, under such conditions, respiration being of the same order of magnitude as gross assimilation, it may be suggested that any discrimination during dark respiration and a higher refixation of respired CO2 compared with assimilation of ambient CO2 could modify the 13C content of carbohydrates formed under these conditions.

The main objectives of the present work were: (i) to examine whether any discrimination occurs during dark respiration in intact leaves; (ii) to compare the carbon isotope composition of respired CO2 with that of leaf carbohydrates; and (iii) to determine the carbon isotope composition of leaf carbohydrates synthesized under water deficit. We investigated the variations in δ13C in both leaf carbohydrates (starch and sucrose) and CO2 respired in the dark from the cotyledonary leaves of Phaseolus vulgaris L. during plant dehydration.

MATERIALS AND METHODS

Plant material

Seeds of French bean (P. vulgaris L., cv Contender) were germinated, before transplanting in 2 dm3 baked clay pots containing a peat-based compost. The plants were grown in a controlled-environment cabinet: the temperature was 21 °C and 17 °C during the day and the night, respectively, and the relative humidity was ≈ 70%. The PPFD at the top of the plants was 220 μmol m–2 s–1 during the 16 h photoperiod. Initially, for a 15 d period, the plants were watered every 2 d to field capacity, then dehydration of the leaves was produced by withholding water from the pots. All the measurements were made on the cotyledonary leaves. At the beginning of the experiments, the cotyledonary leaves were not mature, i.e. they had three-quarters of their final leaf area.

Net CO2 uptake

Leaf net CO2 assimilation was measured using an open gas-exchange system as previously described (Cornic & Ghashghaie 1991). The assimilation chamber had a gas volume of 190 cm3 allowing measurements to be made on a leaf area of ≈ 50 cm2. The air flow rate was 40 dm3 h–1 and the leaf boundary-layer resistance for water vapour was 0·83 m2 s mol–1. The leaf was illuminated using a slide projector (Leitz, Wetzlar, Germany). The leaf gas-exchange parameters were calculated according to von Caemmerer & Farquhar (1981). Measurements started after net CO2 uptake and leaf transpiration reached a steady-state rate under a limited PPFD of 220 μmol m–2 s–1 (saturating PPFD was ≈ 500 μmol m–2 s–1) in air containing 350±10 μmol mol–1 CO2. Leaf temperature during the experiment was 22·5 ± 0·5 °C. VPD was 0·8 ± 0·1 kPa.

Leaf relative water content (RWC) and leaf mass per area (LMA)

RWC and LMA were determined on discs (4 cm2) sampled from the leaves opposite to those used for gas-exchange measurements. One disc per leaf was sampled and RWC was calculated using:

image

where FW, SW and DW are the leaf fresh, water saturated and dry weights, respectively. Saturated weight was determined by weighing the leaf after a 12 h immersion in distilled water in a closed Petri dish at 4 °C in the dark. LMA was calculated as leaf dry weight (g) per leaf area (m2).

Respired CO2 sampling

Leaf respired CO2 was collected at the beginning and at the end of the night using a tight closed system as follows: an attached leaf was first placed in a 600 cm3 darkened respiratory chamber included in the closed system. The air of the whole system was then passed, using the pump of an infrared gas analyser (CI 301, CID, WA, USA) also included in the system, through a soda lime column to remove all CO2 before respired CO2 sampling. A CO2-free atmosphere in the whole system was obtained ≈ 30 min after enclosing the leaf inside the chamber. This ensured that the CO2 in the finally collected air only came from leaf respiration. A magnesium perchlorate column allowed the air in the system to be desiccated before passing through the infrared gas analyser. The whole closed system had a gas volume of ≈ 3·8 dm3 and it took 1·5–2 h to collect ≈ 200 μmol mol–1 CO2 in a 3 dm3 sampling vessel. The sampling vessel was then isolated from the system and its air was passed through a series of alcohol–dry ice and liquid nitrogen traps to remove water vapour and to freeze CO2, respectively. The carbon isotope composition of the purified CO2 was then determined. Preliminary experiments using air samples with known δ13C showed that no fractionation occurred during CO2 purification through the traps. The tightness of the closed system was also tested by experiments with no leaf in the chamber.

Starch and soluble fraction extraction

Leaf discs of 0·5 cm2 were harvested at mid-afternoon (9 h into the photoperiod), at the beginning and at the end of the night (2 h and 8 h into the dark period, respectively, just after respired CO2 sampling). They were rapidly frozen in liquid nitrogen and stored at – 80 °C before sugar extraction. The ribs were avoided at sampling.

Three discs (1·5 cm2) sampled on the same leaf and at the same time were ground in 300 mm3 distilled water using a chilled pestle. The resulting extract was maintained for 20 min at 0 °C before centrifugation at 12 000 g for 10 min. The supernatant containing the water soluble fraction was boiled for 3 min and centrifuged as described above. The supernatant was stored at – 20 °C for further purification of soluble sugars by high-performance liquid chromatography (HPLC). The pellet containing starch and cellular residues was washed with 80% (v/v) ethanol three times in order to remove the pigments; 1 cm3 of ethanol was added to the pellet and boiled for 10 min before centrifugation at 12 000 g. The white pellet was then suspended in 1 cm3 2·5 N NaOH; 100 mm3 was used for starch content measurements (see below) and 900 mm3 was suspended in 250 mm3 12 N HCl to solubilize the starch and was maintained for 1 h at 5 °C. After 10 min centrifugation at 12 000 g the supernatant was retained and the residue was again suspended with 500 mm3 6 N HCl. Methanol [80% (v/v)] was added to the supernatant containing HCl-solubilized starch and maintained at 5 °C overnight to precipitate the starch. The starch powder obtained after desiccation of the precipitate was stored for carbon isotope analysis. In order to have enough material for δ13C analysis, extracts obtained using six discs (3 cm2) sampled on the same leaf were mixed before carbon isotope analysis.

Sucrose and starch contents

The water soluble fraction obtained as described above was filtered (filter HV 0·45 μm type, Nihon Millipore Kogyo K.K, Japan). The sucrose content in filtered extracts was determined by HPLC analysis on 20 mm3 aliquots applied to a Sugar-Pak1 column (6·5 mm diameter and 300 mm length, Waters). The flow rate was maintained at 0·5 cm3 min–1, the pressure at 1700 psi and the temperature of the column at 90 °C. The sugar peaks were evidenced by refractometry (R410 refractometer, Waters). The same process was used for preparing isolated pure fractions of the different sugars for isotope analysis. In this case, in order to obtain enough material, 200 mm3 of filtered extracts were applied to the sugar column and another refractometer was used (RI detector, Jobin Yvon, Iota Model, France). The purity of the collected peaks was checked by cochromatography. Similar analyses using analytical grade sucrose solution showed that no fractionation occurred during its passage through the HPLC column.

The starch content was determined on 100 mm3 of the NaOH-suspended fraction using a colorimetric method, as previously described by Séne, Thévenot & Prioul (1997).

Carbon isotope analysis

Carbon isotope composition (δ13C) with respect to PDB is:

image

where, RS and RPDB are the molar abundance ratios of carbon isotopes,13C/12C, of the sample and the standard PDB, respectively (Farquhar et al. 1982b).

δ13C of respired CO2, whole leaf dry matter and carbohydrates purified as described above, were determined using a stable isotope ratio mass spectometer (VG Optima, Fison, Villeurbane, France) with a high precision (± 0·2‰). The isotope abundance of the samples were compared with that of a working standard reference gas having a δ13C of – 44·45‰ with respect to PDB. δ13C of carbohydrates was determined on CO2 coming from the combustion of sucrose and starch samples in a Carlo Erba elemental analyser. Analytical grade proline with a known δ13C of – 13·45‰ with respect to PDB, was periodically analysed to test the variation in isotopic composition due to sample combustion.

Discrimination during carbohydrate synthesis

δ13C of the air in the culture room (source) was ≈– 9‰ at the beginning of the photoperiod, but decreased during the day, reaching minimum values between – 13 and – 15‰ in the afternoon and increased then at the end of the day. These variations were not continually monitored. Thus, the isotopic composition of the early product of photosynthesis (sucrose) could not be exactly calculated from experimental data.

The variation in modelled discrimination with changes in pi/pa relative to a reference value of pi/pa (= 0·7) was calculated as follows:

image

A decreased dΔ calculated for a given instantaneous pi/pa will thus indicate lower discrimination by the carboxylating enzymes than in the reference case. An increased dΔ which will result from pi/pa higher than the reference value, will indicate an increased discrimination.

RESULTS

Leaf net CO2 assimilation and leaf conductance to CO2

The variations in leaf net CO2 assimilation (A), leaf conductance to CO2 (gc), leaf RWC and LMA during the experiment are shown in Fig. 1 for both well-watered (closed symbols) and dehydrated (open symbols) plants. Time zero corresponds to the last watering for dehydrated plants. During the first 11 d, RWC remained constant (around 90%) in both treatments (Fig. 1c). It then decreased in dehydrated plants reaching 75% after 21 d without watering. Leaf net CO2 assimilation rate (A) rapidly declined in non-watered plants and was near zero 16 d after withholding water from the pots (Fig. 1a, open symbols). In control plants, it remained constant during 11 d and then decreased, reaching values between 1 and 3 μmol CO2 m–2 s–1 at the end of the experiment. In both control and dehydrated plants, the decline in net CO2 uptake was accompanied by a parallel decline in leaf conductance to CO2 (Fig. 1b). The decrease in gc and A in control plants in spite of a constant leaf RWC may indicate leaf ageing in these plants. LMA was low (Fig. 1d) for both treatments at the beginning of the experiment (≈ 15–20 g m–2). It then increased for both treatments, stabilizing around 30 g m–2 in control plants. This increase can be attributed to leaf maturation at the beginning of the experiment. In dehydrated plants, LMA increased more rapidly than in controls during the first half of the dehydration cycle, but decreased during the second half, stabilizing around 25 g m–2 at the end of the experiment.

Figure 1.

. Changes in (a) leaf net CO2 assimilation rate (A), (b) leaf conductance to CO2 (gc), (c) leaf relative water content (RWC%) and (d) leaf mass per area (LMA) as a function of time for both well-watered (closed symbols) and dehydrated plants (open symbols). Day zero corresponds to the last watering for dehydrated plants which occurred 15 d after transplanting. Leaf gas-exchange measurements were made under the ambient growth conditions (photosynthetic photon flux density (PPFD) of 220 μmol m–2 s–1, atmospheric CO2 molar ratio of 350 ± 10 μmol mol–1 and 22·5 ± 0·5 °C). Gas exchange was measured on one cotyledonary leaf for each plant and RWC and LMA were measured on the same plant but on the opposite cotyledonary leaf. The opposite leaves were also used for both carbohydrate content and δ13C shown in Fig. 2. Each data point represents a measurement carried out on an individual plant.

Starch and sucrose contents

The variations in leaf carbohydrate contents are illustrated in Fig. 2(a & b) for starch and sucrose, respectively. Leaf sucrose content increased, reaching a maximum on day 11 in both control and dehydrated plants and then decreased, stabilizing around 1 g m–2 and 0·5 g m–2 in control and dehydrated plants, respectively (Fig. 2b). During the first 11 d, the starch and sucrose contents were similar in both dehydrated and control plants, in spite of a lower leaf net photosynthesis in dehydrated plants. This may indicate a larger carbohydrate exportation from the cotyledonary leaves of control plants compared with dehydrated plants during this period. Thereafter, the tendency reversed; the cotyledonary leaves of control plants contained a large quantity of carbohydrates (particularly starch) while those of dehydrated plants showed a substantial decrease in carbohydrate content. These variations were consistent with those of LMA. The starch content was always at least twice the sucrose content. The variation in the starch content was parallel to the variation in the sucrose content except for control plants after day 11 where the starch content increased continuously, reaching 6 g m–2 at the end of the experiment, while the sucrose content decreased slightly (Fig. 2a). Because glucose and fructose contents were negligible in both control and dehydrated plants (data not shown), carbon isotope analysis was carried out on sucrose and starch.

Figure 2.

. Changes in (a, b) starch and sucrose contents and (c, d) δ13C of starch and sucrose as a function of time for both well-watered (closed symbols) and dehydrated plants (open symbols). Samples were taken at mid-afternoon (9 h into the photoperiod) on the same cotyledonary leaves used for leaf relative water content (RWC) measurements. For carbohydrate contents, each data point is the mean ± standard error (SE) for two measurements on the same leaf. Three leaf discs (1·5 cm2) were used for each measurement. In order to have enough material for δ13C analysis, extracts obtained using six discs (3 cm2) sampled on the same leaf were mixed before measurements. Each point is one measurement on an individual plant.

Isotopic composition of starch and sucrose

Changes in carbon isotope composition (δ13C) in both leaf starch and sucrose during the experiment are shown in Fig. 2(c & d), respectively. The samples were taken at mid-afternoon (9 h into the photoperiod) on the cotyledonary leaf opposite the one used for gas-exchange measurements. Each point is one measurement on an individual plant. The observed scattering may be due to the variability between plants. δ13C of both starch and sucrose was ≈– 26‰ at the beginning of the experiment (day 4) for both treatments, which is similar to values normally observed in C3 plants. δ13C of sucrose increased until day 16 by ≈ 8‰ (from – 26‰ to – 18‰) in dehydrated plants and by ≈ 6‰ (from – 26‰ to – 20‰) in controls, and then decreased in both treatments, reaching the initial level for control plants at the end of the experiment. After day 5, δ13C of starch and sucrose was higher in desiccated plants than in controls. In dehydrated plants, δ13C of starch increased during the first half of the dehydration cycle and then decreased during the second half, as observed for sucrose. However, the amplitude of these variations was less for starch than for sucrose, sucrose remaining heavier than starch in dehydrated plants during the experiment. In well-watered plants, δ13C of starch seemed to be constant and even slightly decreased at the end of the experiment. The same trends have been observed in two other replications (data not shown).

Variation in pi/pa and the expected discrimination

The variation in pi/pa is shown as a function of the leaf net CO2 assimilation (Fig. 3). At high leaf photosynthesis, which corresponds to the beginning of the experiment, pi/pa was high (≈ 0·7) for both treatments. It decreased during the experiment, reaching its lowest values with decreasing stomatal conductance and then increased again. For each value of pi/pa, a corresponding variation in the expected discrimination (dΔ) relative to the reference value (pi/pa = 0·7) was calculated (see Materials and Methods). The changes in dΔ showed that when pi/pa decreased from 0·7 to 0·55 during the first phase of the experiment in control plants, the discrimination should decrease by ≈ 4‰. In dehydrated plants, a decrease in pi/pa from 0·7 to ≈ 0·45, should also decrease the discrimination by ≈ 6‰. Conversely, during the second phase of the experiment, the discrimination should increase with increasing pi/pa.

Figure 3.

. Changes in pi/pa during the experiment as a function of leaf net CO2 assimilation for both control (closed symbols) and dehydrated plants (open symbols). According to Farquhar's simple model, for each observed value of pi/pa an expected variation in the discrimination (dΔ) relative to a reference value when pi/pa = 0·7 was determined (see Materials and Methods).

Carbon isotope composition of whole leaf organic matter

The carbon isotope composition of the whole leaf, measured on the remaining leaf material after sampling the discs for carbohydrate isotope analyses and RWC and LMA, was ≈– 27·5‰ at the beginning of the experiment for both treatments (Fig. 4). As observed for δ13C of sucrose, δ13C of whole leaf increased first during dehydration, reaching – 24·5‰, and then decreased, reaching – 27‰ at the end of the experiment, but always remained lighter than both sucrose and starch. It changed less in controls than in dehydrated plants.

Figure 4.

. Changes in δ13C of whole leaf organic matter for both control (closed symbols) and dehydrated plants (open symbols) during the experiment. Analyses were carried out on the remaining leaf material after sampling the discs for leaf mass per area (LMA), leaf relative water content (RWC) and carbohydrate content and isotopic analysis. Each point corresponds to one leaf.

Isotopic composition of CO2 respired in the dark

The carbon isotope composition of CO2 respired in the dark by the leaves is shown in Fig. 5 at the beginning (circles) and at the end of the night (triangles). Similarly to δ13C of sucrose, two phases were observed for δ13C of respired CO2. First, an enrichment in 13C of respired CO2 was observed, with the highest value on day 16 for non-watered plants: it increased from – 17‰ on day 5 to – 13·5‰ on day 16 at the beginning of the night (Fig. 5, open circles) and from – 19·5‰ to – 14‰ at the end of the night (Fig. 5, open triangles). Control plants showed similar variations but with a maximum value on day 19: δ13C increased from – 20·5‰ to – 16‰ both at the beginning and at the end of the night (Fig. 5, closed symbols). Then, a decrease in δ13C of respired CO2 occurred in both well- and non-watered plants, reaching values close to those observed at the beginning of the experiments. Respired CO2 from the leaves of dehydrated plants was always heavier (richer in 13C) at the beginning of the night compared with that measured at the end of the night.

Figure 5.

. Changes in the isotopic composition of CO2 respired in the dark by the cotyledonary leaves of both control (closed symbols) and dehydrated plants (open symbols) as a function of the time from withholding water. Respired CO2 was sampled using one cotyledonary leaf of a plant at the beginning of the night (circles) and the opposite one of the same plant at the end of the night (triangles). For each date, one control and one dehydrated plant were used. Respired CO2 collected in a sampling vessel (included in a closed system) was purified before isotopic analysis (see Materials and Methods).

Relationship between δ13C of respired CO2 in the dark and δ13C of sucrose and starch

δ13C of respired CO2 plotted versus δ13C of starch and sucrose is shown in Fig. 6. Leaf discs for starch and sucrose carbon isotope analysis were sampled on the same leaves just after respired CO2 sampling. A good correlation was found between δ13C of respired CO2 and δ13C of sucrose for both treatments. The correlation was higher at the beginning compared with the end of the night (Fig. 6c & d, respectively). This line is parallel to a 1:1 relationship, respiratory CO2 always being enriched in 13C compared with sucrose by ≈ 6‰. No correlation was obtained between δ13C of respired CO2 and δ13C of starch (Fig. 6a & b).

Figure 6.

. Relationship between the carbon isotope composition (δ13C) of CO2 respired in the dark and of sucrose and starch measured on the same leaves (a, c) at the beginning of the night and (b, d) at the end of the night. After respired CO2 trapping, six discs (3 cm2) were sampled on each leaf for starch and sucrose δ13C analysis. The values of δ13C of CO2 respired in the dark from Fig. 5 are used for these relationships. The dashed line corresponds to a 1:1 relationship.

DISCUSSION

Sucrose and starch had a similar carbon isotope composition at the beginning of the experiment, but thereafter, sucrose became heavier than starch in both control and dehydrated plants during leaf ageing and dehydration. This is in contrast with the results of Deléens-Provent & Schwebel-Dugué (1987) on maize seedlings, Brugnoli et al. (1988) on some C3 species (bean, poplar and cotton) and Gleixner et al. (1993) on sugar beet, where starch has been reported to be heavier than soluble sugars. Brugnoli et al. (1988) suggested that the branch points at triose phosphate and at fructose-6-phosphate levels provide possibilities for discrimination leading to the formation of heavy starch, while the lighter triose phosphates preferentially move out of the chloroplasts and form sucrose in the cytosol. This may be the case when comparing the starch and sucrose synthesized during a short period and/or under relatively constant conditions. But during our experiment, the carbon isotope composition was measured for the total starch pool being formed during 4 weeks under changing stomatal conductance. The turnover of sucrose being more rapid than that of starch, it is suggested that δ13C of the starch pool represents a time-integrated discrimination, while the sucrose pool reflects the carbon isotope composition of more recently formed carbohydrates. This may explain a relatively lower 13C content in the starch pool compared with sucrose during stomatal closure in the present study. Moreover, in cotyledonary leaves, the starch pool may have some fraction deriving from the seed reserves with an isotopic signature of the mother plant, i.e. more negative. This may explain why the starch, contrary to the observations mentioned above, was not heavier than the sucrose before stomatal closure at the beginning of the experiment.

As expected, leaf carbohydrates became heavier during the first phase of plant dehydration. This enrichment in 13C is more marked in sucrose than in starch, probably due to their different turnover rates. This 13C enrichment can be attributed to the decrease in stomatal conductance decreasing pi/pa and, thus, decreasing discrimination by Rubisco. For control plants, this may also be explained by stomatal closure, but only due to leaf ageing and not due to dehydration (RWC was constant in control plants). These results are in agreement with the simple model of Farquhar. Interestingly, during the second phase, when stomata were closed and net photosynthesis very low (near zero in dehydrated plants and ≈ 1–3 μmol CO2 m–2 s–1 in control ones) the tendency was reversed; sucrose became lighter (depleted in 13C). A similar decrease in 13C content was also observed for starch, but only in dehydrated plants. In control plants, δ13C of starch remained constant during the experiment. This can be explained by starch accumulation in these plants preventing discernible variation in starch pool 13C content.

The most simple explanation for the reversion in 13C content in the carbohydrates during the second half of the experiment, could be the increase in pi/pa. Indeed, the variation in δ13C of sucrose observed during the experiment (≈ 6‰ and 8‰ for control and dehydrated plants, respectively) roughly corresponds to the expected variation in discrimination relative to a reference value (≈ 4‰ and 6‰ for control and dehydrated plants, respectively) calculated using the simple model of Farquhar, indicating that the observed variation in sucrose δ13C could be mostly attributed to the variation in pi/pa due to stomatal closure during both phases of the experiment and for both treatments. The increase in pi/pa at the end of the experiment may be due to a possible non-stomatal decrease in leaf photosynthetic capacity. However, the values of pi/pa under severe water deficit cannot be estimated with precision and should be taken with caution because of: (i) a possible stomatal patchiness; and (ii) increasing imprecision in the determination of pi/pa when stomatal and cuticular conductances are of the same order of magnitude [see Meyer & Genty (1998)].

Other possible explanations for this depletion in 13C content of leaf carbohydrates during the second phase of the experiment are: (i) leaf net CO2 assimilation being very low: the proportion of recently formed carbohydrates (enriched in 13C as predicted by the model) compared with those synthesized at the beginning of the experiment (depleted in 13C) decreased, resulting in a more negative total carbohydrate pool; (ii) the dark respiration being of the same order of magnitude as the gross assimilation: a possible discrimination during substrate utilization releasing 13C-enriched CO2 during the night should lead to 13C depletion in the remaining leaf material; (iii) under such conditions, the proportion of refixation of respired CO2 increased relative to net CO2 uptake. Respired CO2 being depleted in 13C compared with ambient air, the carbohydrates are progressively depleted in 13C.

δ13C of CO2 respired in the dark was parallel to δ13C of sucrose for both treatments, respired CO2 always being heavier than sucrose. One may suggest that sucrose or a closely linked substance is the main substrate for dark respiration and that the variation in δ13C of respired CO2 observed during the experiment was due to the variation in δ13C of this substrate. Because of a fast turnover, sucrose should carry the signature of recently formed carbohydrates, as long as the daily carbon balance is positive, or at least during the day and at the beginning of the night, as long as the net photosynthesis is positive. It is representative for all assimilates derived from photosynthesis and forming potential substrates of respiration. In dehydrated plants, different pools seem to be used at the beginning and at the end of the night: the recently formed carbohydrates (enriched in 13C) were used at the beginning of the night, but their amount was limited in these plants, the previously formed reserves (depleted in 13C) were used at the end of the night. In control plants, the same pool seems to be used at the beginning and at the end of the night.

There was a linear relationship between carbon isotope composition of respired CO2 and sucrose. Respired CO2 was always enriched in 13C compared with sucrose by ≈ 6‰, suggesting that if sucrose or a closely linked substance was used as the substrate, a discrimination by ≈– 6‰ occurred during dark respiration processes. This is in agreement with some data on intact plants in the literature [see O’Leary (1981)], but in contrast to the recent results of Lin & Ehleringer (1997) who observed no fractionation during dark respiration by protoplasts isolated from bean and maize leaves and incubated with sucrose, glucose or fructose. However, 13C enrichment in CO2 respired in the dark compared with substrate (sugars) can result from a non-statistical 13C distribution within the hexose molecules. Rossmann, Butzenlechner & Schmidt (1991) have indeed demonstrated that C-3 and C-4 of glucose molecules extracted from sugar beet leaves were enriched in 13C compared with other carbon positions. They proposed that, during the course of lipid biosynthesis, the pyruvate issued from glucose by the glycolytic pathway produces 13C-enriched CO2 by decarboxylation of the 13C-enriched carbons (C-3 and C-4 coming from glucose), while more depleted ones (C-1, C-2, C-5 and C-6) form acetyl-CoA. This may explain not only the well-known 13C depletion in lipids by ≈ 8‰ compared with whole plant already reported by Park & Epstein (1961), but also the enrichment of respired CO2 compared with sugars observed in the present study. Besides, the 13C depletion in lipids and, thus, the 13C enrichment in respiratory CO2 may also result from a kinetic isotope effect during decarboxylation of pyruvate. This possibility has already been discussed by Park & Epstein (1961), De Niro & Epstein (1977), Rossmann et al. (1991) and Gleixner et al. (1993). The effect of both the non-statistical 13C distribution within the hexose molecules and the kinetic isotope effect during decarboxylation of pyruvate will depend, however, on the relative importance of the metabolic pathways in plants: a higher rate of lipid biosynthesis (or other substances issued from the Krebs cycle) should enhance the enrichment in 13C of released CO2, conversely a lower rate of synthesis of these substances by the Krebs cycle should lead to the decarboxylation of ‘light’ as well as ‘heavy’ carbons. This may explain the discrepancy in the values of δ13C of respired CO2 reported in the literature on different species.

This may also explain the results of Lin & Ehleringer (1997). The metabolism of isolated protoplasts may indeed be different from that of intact leaves.

The lack of any correlation between δ13C of respired CO2 and starch may be, as mentioned above, due to the presence of previously formed starch persisting during the experiment and preventing discernible variation in δ13C of the starch pool. Similar δ13C values obtained for starch and sucrose at the beginning of the experiment suggest that a similar correlation between δ13C of respiratory CO2 and starch should also be observed if recently formed starch (and not total starch pool) is taken into account. Labelling experiments are needed to confirm this hypothesis.

The time course of δ13C of whole leaf organic matter parallels the time course of sucrose. Yet organic matter was always at least 1‰ lighter than sucrose and starch. This difference was conserved even through the period of ongoing increase in LMA. Therefore, it can be interpreted as an independent indication of an isotopic discrimination which results in a respiratory carbon flux enriched in 13C and a flux impoverished in 13C directed towards incorporation into organic matter.

Our results show that there is: (i) a change in the carbon isotope composition of the CO2 respired in the dark; and (ii) a constant 13C enrichment (≈ 6‰) in respiratory CO2 compared with sucrose during leaf ageing and plant dehydration, both at the beginning and at the end of the night. It is concluded that a discrimination occurs during dark respiration processes in bean cotyledonary intact leaves and that this leads to 13C depletion in the remaining leaf material.

Acknowledgements

We are grateful to Jacqueline Liébert and Marie-Thérèse Adeline for technical assistance during carbohydrate extraction and sucrose purification using HPLC, respectively, to Dr Mamoun Séne for colorimetric analysis of starch and to Alain Thoreux for carbon isotope analysis.

Ancillary