Respiratory metabolism of illuminated leaves depends on CO₂ and O₂ conditions

Abstract

Day respiration is the process by which nonphotorespiratory CO₂ is produced by illuminated leaves. The biological function of day respiratory metabolism is a major conundrum of plant photosynthesis research: because the rate of CO₂ evolution is partly inhibited in the light, it is viewed as either detrimental to plant carbon balance or necessary for photosynthesis operation (e.g., in providing cytoplasmic ATP for sucrose synthesis). Systematic variations in the rate of day respiration under contrasting environmental conditions have been used to elucidate the metabolic rationale of respiration in the light. Using isotopic techniques, we show that both glycolysis and the tricarboxylic acid cycle activities are inversely related to the ambient CO₂/O₂ ratio: day respiratory metabolism is enhanced under high photorespiratory (low CO₂) conditions. Such a relationship also correlates with the dihydroxyacetone phosphate/Glc-6-P ratio, suggesting that photosynthetic products exert a control on day respiration. Thus, day respiration is normally inhibited by phosphoryl (ATP/ADP) and reductive (NADH/NAD) poise but is up-regulated by photorespiration. Such an effect may be related to the need for NH₂ transfers during the recovery of photorespiratory cycle intermediates.

It has been 70 years since Krebs and Johnson (1, 2) proposed the mechanism by which pyruvic acid is oxidized to CO₂, which is now called the “Krebs cycle” or tricarboxylic acid (TCA) cycle. Whereas the basics of the metabolic reactions involved in leaf respiration are known, intense efforts are still currently devoted to elucidating the regulation of the TCA cycle (and, more generally, of day respiration) in illuminated leaves (for a recent review, see ref. 3).

Leaf day respiration (nonphotorespiratory CO₂ evolution in the light) is an essential metabolic pathway that accompanies photosynthetic CO₂ assimilation and photorespiration. It is widely accepted that leaf respiration is partly inhibited in the light when compared with darkness (4). This acceptance is based on several strong lines of evidence, ranging from gas-exchange to molecular studies (for a review, see ref. 4): (i) the inhibition is thought to cause the light-enhanced dark respiration (5); (ii) the pyruvate dehydrogenase (PDH) is down-regulated in the light (6, 7); (iii) the metabolic flux through the TCA cycle in the light is reduced in both extracted mitochondria (8) and intact leaves (9, 10); (iv) mitochondria experience high ATP/ADP and NADH/NAD⁺ ratios in the light that inhibit NAD-dependent isocitrate dehydrogenase (11); and (v) carbohydrate molecules such as sucrose (Suc) and glucose (Glc) are prevented from entering glycolysis (9), because of a modification of phosphofructokinase activity by the allosteric effector fructose (Fru)-2,6-bisphosphate (12). Nevertheless, not all leaf cells are photosynthetic (e.g., most epidermal cells, phloem, and xylem) so that some “heterotrophic” background respiration in the light is expected, but its contribution is minor.

Although inhibited by light, day respiration is critical for plant growth and leaf N assimilation, as it provides ATP for Suc synthesis and TCA cycle intermediates (e.g., 2-oxoglutarate and oxaloacetate) for ammonium assimilation and amino acid synthesis (13). Thus for many years, the down-regulation of day respiration and the TCA cycle in the light has been viewed as a perplexing phenomenon. It may be argued that the partial inhibition of day respiration comes from a balance between two metabolic constraints: the energy requirement for Suc synthesis and the minimal competition between glycolysis and Suc synthesis for improved carbon gain. However, day respiration is additionally affected by other metabolic processes such as O₂ assimilation (photorespiration) (14), the rate of which depends on the internal CO₂/O₂ ratio. The photorespiratory cycle leads to Gly oxidative decarboxylation in the mitochondrion that supposedly gives rise to a large NADH/NAD⁺ ratio, which in turn inhibits certain respiratory mitochondrial enzymes in vivo (11). Photorespiration is thus assumed to down-regulate day respiratory CO₂ evolution. Nevertheless, increased photorespiration rates could require more glutamate (Glu) cycling to provide amino groups for glycine (Gly) synthesis in the peroxisomes. This higher demand might in turn require an increase in 2-oxoglutarate and Glu synthesis, and thus a higher day respiratory rate. The rationale of the metabolic homeostasis between day respiration and photorespiration is therefore currently uncertain.

To clarify the regulation of day respiration in illuminated leaves and its interactions with photorespiration, we have investigated the effect of the carboxylation-to-oxygenation ratio on day respiratory metabolic fluxes by using isotopic ¹²C/¹³C spectrometry and ¹³C and ³¹P NMR. The results show that, although respiratory CO₂ evolution is always inhibited in the light when compared with the dark, the metabolic flux associated with the TCA cycle is inversely related to net CO₂ assimilation, which correlates with changes in phosphorylated metabolites levels. In addition, ¹³C distribution after labeling shows a larger commitment toward TCA intermediates and Glu as photorespiration increases. These findings, which are consistent with a role of day respiration in sustaining photorespiratory N cycling and perhaps nitrate assimilation, have important implications, ranging from the improvement of nitrogen use efficiency to the understanding of leaf and global ecosystem carbon budgets.

Results

To determine the amplitude and the steps of the leaf respiratory pathway that are inhibited in the light, detached leaves were fed with positionally labeled ¹³C-enriched substrates [pyruvate (Pyr) or Glc] and decarboxylation rates were measured by gas exchange coupled to isotopic spectrometry. This method allowed us to calculate the decarboxylation rate in the light and, by comparing with the rate in darkness, the inhibition of decarboxylation in illuminated leaves. The positional labeling in Pyr allowed us to discriminate between the CO₂ produced by either the PDH ([1-¹³C]Pyr, [3-¹³C]Glc) or the TCA cycle ([2-¹³C]Pyr, [1-¹³C]Glc).

TCA-Mediated Decarboxylations Are Enhanced Under Low CO₂/O₂.

When leaves were fed ¹³C-enriched Pyr, the apparent carbon isotope discrimination Δ_obs increased, showing that ¹³CO₂ was produced (Fig. 1A). Interestingly, the decarboxylation of [2-¹³C]Pyr was low compared with that of [1-¹³C]Pyr, showing the predominance of CO₂ produced by PDH as opposed to that produced by the TCA cycle. In addition, the smaller the CO₂/O₂ ratio, the larger the Δ_obs associated with [1-¹³C]Pyr. A similar but modest trend occurred with [2-¹³C]Pyr (Fig. 1A).

Fig. 1.

Steady isotope discrimination and carbon isotope composition of CO₂. (A) Carbon isotope discrimination (Δ_obs) associated with photosynthesis of detached leaves [at 21°C and 400 μmol·m⁻²·s⁻¹ photosynthetic photon flux density (PPFD)] fed with either ¹³C-1- or ¹³C-2-enriched Pyr under four CO₂/O₂ conditions: 140, 400, or 1,000 μl·liter⁻¹ CO₂ in 21% O₂ or 400 μl·liter⁻¹ CO₂ in 2% O₂. (B) Carbon isotope composition (δ¹³C) of respired CO₂ in darkness after the corresponding light periods. When leaves experienced a light period under 2% O₂, the carbon isotope composition was measured in either 21% (indicated as 400 μl·liter⁻¹, 2–21%) or 2% O₂ (indicated as 400 μl·liter⁻¹, 2–2%). Each value is the mean ± SE of three measurements. The control δ¹³C value of respired CO₂ was −22.1 ± 0.5‰.

Dark-respired CO₂ was ¹³C-enriched compared with the natural abundance (δ¹³C of −22.1 ± 0.5‰) with both [1-¹³C]- and [2-¹³C]Pyr, indicating that decarboxylation of Pyr was substantial (Fig. 1B). The decarboxylation of [1-¹³C]Pyr in the dark was larger after a light period at low CO₂ (140 μl·liter⁻¹) compared with a high CO₂. Both [1-¹³C]Pyr and [2-¹³C]Pyr were more decarboxylated in the dark after leaves had been exposed to light under low O₂ conditions (Fig. 1B), and this low O₂ effect was sensitive to O₂ conditions in the dark.

The isotopic data shown in Fig. 1 were used to calculate the decarboxylation rates (Fig. 2) associated with PDH and the TCA cycle by using mass-balance equations (see Materials and Methods). Although the decarboxylation rate associated with PDH was inhibited by only 30% in the light (Fig. 2), TCA cycle-mediated decarboxylations were much lower in the light than in the dark, with an inhibition of ≈80% under typical atmospheric conditions (400 μl·liter⁻¹ CO₂ in 21% O₂). The inhibition of the PDH-mediated decarboxylation was relatively constant under the CO₂ and O₂ conditions investigated. In contrast, TCA-mediated decarboxylations were much more sensitive to low CO₂ conditions, with 35% inhibition only at 140 μl·liter⁻¹ CO₂ in 21% O₂ (Fig. 2C). It should be noted that although the inhibition value associated with the TCA cycle was similar in 400 μl·liter⁻¹ CO₂ 2% O₂ and 400 μl·liter⁻¹ CO₂ 21% O₂, the absolute decarboxylation value was larger in 2% O₂ both in the dark and in the light (Fig. 2). This effect was caused by a higher stomatal conductance in 2% O₂ and, subsequently, larger transpiration rates that induced a higher absorption of labeled compounds (data not shown). Such a positive effect of low O₂ conditions on stomatal conductance has already been observed in Xanthium strumarium (15).

Fig. 2.

Decarboxylation rates and inhibition of decarboxylation by light calculated from data of Fig. 1, using the method of ref. 9. (A and B) The decarboxylation rates by PDH (empty bars) and the TCA cycle (filled bars) are given in the light (A, denoted as r_light) and in the dark (B, denoted as r_night). (C) Inhibition by light (calculated as 1 − r_light/r_night, in %) is indicated. Conditions experienced by leaves during the light period are indicated on the x axis as in Fig. 1: 140, 400, and 1,000 μl·liter⁻¹ CO₂ in 21% O₂ and 400 μl·liter⁻¹ CO₂ in 2% O₂. The δ¹³C values of dark-respired CO₂ obtained in 2% O₂ (Fig. 1B Right) were used to calculate the inhibition value after a light period in 2% O₂.

The Commitment to Glycolysis Is Enhanced Under Low CO₂/O₂.

Similar experiments were carried out with ¹³C-enriched Glc to determine whether the glycolytic carbon flow changes under varying CO₂/O₂ conditions. Under typical conditions (400 μl·liter⁻¹ CO₂ in 21% O₂), leaves fed with [1-¹³C]Glc or [3-¹³C]Glc did not produce significant amounts of ¹³CO₂ in the light, as indicated by the very small deviation of the apparent carbon isotope discrimination Δ_obs [supporting information (SI) Table 1]. The same applied to high CO₂-to-O₂ conditions (1,000 μl·liter⁻¹ CO₂ in 21% O₂), with a Δ_obs value of 24.5 ± 0.5‰. In contrast, the decarboxylation of [1-¹³C]Glc and [3-¹³C]Glc became apparent in the light under low CO₂ conditions (140 μl·liter⁻¹ in 21% O₂), with Δ_obs values increasing up to 106.4 ± 13.8‰ (with [3-¹³C]Glc; SI Table 1). This finding indicated that [¹³C]Glc could be oxidized in the light by glycolysis and that PDH and TCA activities were both at the origin of the respired ¹³CO₂ from [¹³C]Glc. It should be noted that this increase was not an artifact caused by the low CO₂ mole fraction in the chamber (making decarboxylated CO₂ proportionally larger), because the CO₂ mole fraction was taken into account in the mass-balance-based calculations. In darkness, the δ¹³C value of respired CO₂ increased to 365‰ (with [3-¹³C]Glc) after photosynthesis at 400 μl·liter⁻¹ CO₂ and 719‰ (with [3-¹³C]Glc) after photosynthesis at 140 μl·liter⁻¹ CO₂ in 21% O₂ (SI Table 1). When these values were used to calculate decarboxylation values, it was found that both TCA- and PDH-mediated decarboxylations were inhibited by nearly 90% at 400 μl·liter⁻¹ CO₂ and 60% at 140 μl·liter⁻¹ CO₂ in the light (SI Table 2). These data show that leaf CO₂ levels modulate the entry of Glc molecules into the glycolytic and respiratory pathways in both the light and the dark; Glc is a better respiratory substrate at low CO₂ levels.

Distribution of the ¹³C Label in Metabolites.

To gain information on the changes in metabolic pathways under the different CO₂/O₂ conditions in the light, the fate of the ¹³C atoms (from ¹³C-substrate feeding) was determined in leaf metabolites by ¹³C NMR analyses. Leaves were fed with positionally enriched (99% ¹³C) substrates under either 140, 400, or 1,000 μl·liter⁻¹ CO₂ in 21% O₂ and the positional isotopic abundances of identified metabolites (in % of ¹³C), measured by NMR, are displayed as an isotopomics array (Fig. 3). As expected, hexoses and Glc or Fru moieties of Suc were ¹³C-labeled when [¹³C]Glc was supplied to leaves so that several C-1 and C-6 positions formed clusters (these positions are redistributed by aldolase and triose-phosphates isomerase reactions). The C-4 and C-5 positions in hexoses clustered near the C-3 positions, indicating that a redistribution of the ¹³C label occurred in the light through the pentose-phosphate pathway.

Fig. 3.

Isotopomics array representation of ¹³C abundance in the carbon atom positions of major metabolites in detached leaves incubated with ¹³C-substrates for 2 h at 21°C, 21% O₂, and 400 μmol·m⁻²·s⁻¹ PPFD. CO₂ mole fraction was 140 μl·liter⁻¹, 400 μl·liter⁻¹, and 1,000 μl·liter⁻¹. At t = 2 h, leaves were immediately frozen in liquid nitrogen for perchloric acid extraction. Perchloric extracts were analyzed for positional ¹³C abundances by NMR. Each column is a separate set of experimental conditions. Cit, citrate; Ido/Gal, uncertain d-hexofuranose belonging to the idose-galactose group; Obt, oxobutyrate; SF and SG, fructosyl and glucosyl moieties of sucrose, respectively. Red and green cells indicate ¹³C abundances above and below the natural abundance (which is 1.1%). Below-natural abundance cells appear dark green because the ¹³C abundance is still very close to 1.1%.

The flux through the pentose–phosphate cycle may be estimated with both NMR and gas-exchange data. At 140 μl·liter⁻¹ CO₂, the ¹³C amount in the C-atom positions redistributed by the pentose–phosphate cycle (i.e., the ¹³C amount in C-2, C-3, C-4, and C-5 after [1-¹³C]Glc labeling and the ¹³C amount in C-1, C-2, C-5, and C-6 after [3-¹³C]Glc labeling), as found by NMR after labeling, corresponds to a ¹³C flux of 0.03 μmol·m⁻²·s⁻¹. With the gas-exchange data, the flux through the pentose-phosphate cycle can be estimated by the excess of calculated CO₂ production from the TCA cycle with respect to the PDH, because the pentose-phosphate cycle involves the decarboxylation of the C-1 atom of Glc. With [1-¹³C]Glc labeling, the TCA cycle-mediated decarboxylation rate was 0.11 μmol·m⁻²·s⁻¹, whereas the PDH-mediated decarboxylation was 0.07 μmol·m⁻²·s⁻¹ (SI Table 2). The CO₂ production by the pentose–phosphate cycle was thus 0.04 μmol·m⁻²·s⁻¹, a value that is very close to that obtained with NMR. Such a value is nevertheless not enough to explain the entire ¹³CO₂ production from [¹³C]Glc at low CO₂, so that the enhancement of the commitment to glycolysis still holds under this condition.

The commitment to the TCA cycle can be assessed with ¹³C labeling in organic acids. Some organic and amino acids clustered on the upper part of Fig. 3, with positional ¹³C enrichment generally <10%. However, the C-2 atom of malate (Mal) and the C-2 and C-3 atoms of Glu were clearly labeled when [¹³C]Pyr was supplied. This labeling was lower at high CO₂, and the same applied to the C-2 and C-3 atoms of fumarate (Fum). Because these C-atom positions in Mal, Glu, and Fum can only be labeled by the interplay (redistribution) of the TCA cycle, this labeling trend is indicative of an increase in TCA cycle activity. This view agrees with the increase of the ¹³C abundance in Glu, succinate (Succ), and Fum from high to low CO₂/O₂ ratios as shown in Fig. 4(where ¹³C abundances are relative to those found in Mal to take into account variations in refixed decarboxylated ¹³CO₂ by phosphoenolpyruvate carboxylase (PEPC) activity, as discussed below).

Fig. 4.

¹³C abundance in Glu (black bars), Succ (light gray bars), and Fum (dark gray bars) relative to that in Mal. Values are from the data of Fig. 3. The three different CO₂ mole fractions (in μmol·mol⁻¹) used in the experiment are indicated on the x axis. (Inset) Data of Fig. 3 replotted to show the relationship between the positional ¹³C abundance (in percentage of ¹³C) in Mal C-2 (▾), Fum C-2/3 (●), and Glu C-2 (□) and the quantity of metabolite (in μmol per gram of fresh weight). Short dashed lines indicate exponential decay (Fum and Mal) and linear (Glu) regressions; both are significant: F = 7.26 (P < 0.005) and F = 16.38 (P < 0.003), respectively.

Nevertheless, the lower ¹³C abundance in organic acids under high CO₂ conditions may also partly be caused by the diluting effect of assimilated ¹²C-enriched carbon (inlet CO₂ has a δ¹³C value near −50‰). This effect is expected in Mal and Fum, which are two major metabolites accumulated by Xanthium leaves. There was indeed a significant and negative correlation between the isotopic enrichment in Fum C-2/3 and Mal C-2 and the quantity of these organic acids (Fig. 4 Inset). However, the correlation was reversed in the case of Glu C-2 (Fig. 4 Inset), whereas there was no correlation at all with other organic acids or amino acids (data not shown). Therefore, we conclude that the lower ¹³C enrichment in organic and amino acids at high CO₂ conditions was not only caused by an isotopic dilution but also by the decrease of the commitment of [¹³C]Pyr to the TCA cycle.

¹³CO₂ Refixation.

Whereas the trend of ¹³C labeling in TCA cycle intermediates is clear, the rate of decarboxylation of ¹³C-enriched substrates in the light and the ¹³C labeling of different metabolites may have been adulterated by ¹³CO₂ refixation by either PEPC or Rubisco. Indeed, ¹³CO₂ fixation by PEPC occurred under each CO₂ condition investigated in the present study, as revealed by the ¹³C enrichment in the C-4 of Mal when [1-¹³C]Pyr or [2-¹³C]Pyr was supplied to leaves. However, the C-2 and C-3 positions in Mal were also clearly ¹³C-enriched under low CO₂ conditions when fed with [2-¹³C]Pyr (Fig. 3), and this observation is consistent with increased TCA cycle activity.

Metabolites could have been also ¹³C-labeled via photosynthetic ¹³CO₂ refixation. The labeling of hexoses and Suc in the C-3 position after Pyr feeding (refixation of ¹³CO₂ decarboxylated from [1-¹³C]Pyr by Rubisco) did occur but the maximum positional ¹³C abundance in C-3 indicates that the proportion of refixed ¹³C in the whole molecule was only ≈0.7%. Refixation of respired CO₂ into starch was also assessed after gas exchange and on-line apparent Δ_obs measurements (Fig. 1): the carbon isotope composition (δ¹³C) of starch was between −36.5 (minimum value obtained at 1,000 μl·liter⁻¹ CO₂, with [1-¹³C]Pyr feeding) and −12.7‰ (maximum value obtained at 140 μl·liter⁻¹ CO₂ with [1-¹³C]Pyr feeding), indicating that the proportion of starch containing refixed CO₂ was only between 0% and 0.4% (see SI Table 3). Therefore, although Rubisco-mediated refixation of decarboxylated ¹³CO₂ did occur, it was always very low.

Correlations with Phosphorylated Metabolites.

Phosphorylated metabolites, such as dihydroxyacetone phosphate (DHAP), are known to be important regulators of the primary carbon metabolism of plant leaves (12, 16). Therefore, we measured the amounts of phosphorylated compounds by ³¹P NMR spectroscopy on the same samples used for the ¹³C labeling and ¹³C NMR analyses. There was a clear linear and positive correlation between the activity of the TCA cycle, as witnessed by gas exchange (values from Fig. 2), and the DHAP-to-Glc-6-phosphate ratio (Fig. 5, dashed line). In contrast, there was no statistically significant correlation with the DHAP/P_i ratio although a positive trend was apparent (Fig. 5, dotted line).

Fig. 5.

Relationship between the inhibition of the TCA cycle in the light (in %; data from Fig. 2) and the DHAP to inorganic phosphate (P_i) (○) or Glc-6-phosphate (●) ratio. Phosphorylated compounds were measured by ³¹P NMR on the same samples used for ¹³C NMR after ¹³C labeling. Lines stand for linear regressions. The regression with DHAP/Glc-6-P is significant (F = 61.7, P < 0.08).

Discussion

Based on data from gas-exchange analyses to enzymatic activities (6–9), it has become widely accepted that leaf day respiration (nonphotorespiratory CO₂ evolution in the light) is inhibited in the light (for a review, see ref. 3). However, the rationale and the effect of environmental conditions on day respiration and its inhibition are uncertain. Solving such an issue is critical for understanding how leaves adapt carbon partitioning between export of photosynthates and respiration or N economy under varying natural conditions. CO₂ and O₂ levels are two parameters of fundamental importance because the leaf internal CO₂/O₂ ratio changes when environmental conditions alter stomatal closure (e.g., drought). Here, we have developed isotopic methods to provide evidence that respiratory metabolism is up-regulated when the CO₂/O₂ ratio decreases, and we argue that day respiration is an exquisite example of metabolic compromise between feedback inhibition by NADH and ATP and 2-oxoglutarate precursor requirement for N metabolism.

The Regulation of the TCA Cycle.

Although no major effect on PDH decarboxylation was observed, the carbon flow through the TCA cycle increased under low CO₂/O₂ conditions, as evidenced by the larger decarboxylation rate of [2-¹³C]Pyr, so that the inhibition by light of the TCA cycle was 40% at 140 μl·liter⁻¹ CO₂ as compared with nearly 90% under typical conditions (400 μl·liter⁻¹ CO₂; Figs. 1 and 2). Furthermore, the ¹³C labeling of Succ, citrate, and Glu, as revealed by NMR tracing after [2-¹³C]Pyr feeding, increased as the CO₂ mole fraction decreased (Fig. 4). In this context, the observed labeling in Mal (Fig. 3), which indicated the simultaneous increase in PEPC activity, came as no surprise. It is in agreement with the anapleurotic role of this enzyme, which compensates for 2-oxoglutarate consumption (for Glu synthesis) by feeding the TCA cycling with oxaloacetate molecules (17). The whole picture is thus consistent with an increased commitment to the TCA cycle and Glu production under low CO₂ conditions.

The up-regulation of the TCA metabolism under low CO₂ stems from a larger glycolytic carbon input, as evidenced by the enhancement of [3-¹³C]Glc decarboxylation (see Results and SI Tables 1 and 2) and the slight ¹³C labeling in Mal C-2 after [¹³C]Glc feeding (Fig. 3). In addition, as a consequence of the lower photosynthetic CO₂ fixation rate, the DHAP/Glc-6-P ratio decreased with CO₂ mole fraction, and importantly, this ratio was strongly correlated with the inhibition of the TCA cycle (Fig. 5). The relative abundance of triose phosphates is already known to be a metabolic parameter that controls carbon entry to glycolysis by the interplay of the effector Fru-2,6-bisphosphate (12) and promotes pyruvate kinase activity (18). It thus appears clear that it also controls the commitment of carbon molecules to the respiratory pathway. Uncertainty nevertheless remains about whether it acts indirectly on the TCA cycle (through the enhancement of glycolysis) or not.

Interactions with Photorespiration.

The higher carbon flow through the TCA cycle under low CO₂ may appear somewhat paradoxical as it has often been supposed that large photorespiratory rates inhibit mitochondrial respiratory enzymes (11, 19), such as the NAD-dependent isocitrate dehydrogenase (11). Accordingly, Gly decarboxylase antisense lines of potato (Solanum tuberosum) have lower decarboxylation rates in the light (as revealed by ¹⁴C labeling experiments) and the ATP/ADP and NADH/NAD⁺ ratios are both higher than in the wild type (20). In addition, the predominance of NADH production by Gly oxidation over that by the TCA cycle has been shown by using isolated mitochondria under ADP-limiting conditions (21). This competition might reduce the NAD⁺ available for the mitochondrial dehydrogenase steps of the TCA cycle. Our results show that the inhibition of day respiration occurs whatever the CO₂/O₂ ratio and furthermore, there is a lower inhibition value under very low O₂ conditions (400 μl·liter⁻¹ CO₂, 2% O₂) as compared with high CO₂ conditions (1,000 μl·liter⁻¹ CO₂, 21% O₂) (Fig. 2). This finding would be consistent with a much reduced mitochondrial redox poise caused by the nonphysiological, very low oxygen mole fraction.

However, we show here that the inhibition of the TCA cycle is relaxed under low CO₂ conditions in 21% O₂ (see above). Isocitrate production is probably not influenced by such photorespiratory conditions, as citrate synthase remains active because of the very large K_i(ATP) (5 mmol·liter⁻¹) and the absence of any NADH effect (22). As the mitochondrial NAD-dependent isocitrate dehydrogenase is believed to be inhibited in the light, we suggest that isocitrate is processed by the cytosolic or mitochondrial NADP-dependent isocitrate dehydrogenase (23). This bypass would allow the sustaining of the necessary Glu flow under photorespiratory conditions.

Possible Rationale.

The regulation of the TCA cycle by CO₂/O₂ conditions may be viewed as a side effect of the drop in the DHAP relative quantity on the commitment to glycolysis and therefore respiration (see above and Fig. 5). Unless there are other prevailing imperatives (such as the need of NADH to reduce photorespiration-derived hydroxypyruvate or H₂O₂), it also reflects an increased need for Glu to feed photorespiratory N recycling when conditions shift to low CO₂ mole fractions, simply because the production of Gly from glycolate would require a higher Glu flow. Unsurprisingly then, the production of the Glu precursor 2-oxoglutarate by the TCA cycle is enhanced. The argument that photorespiration is beneficial for Glu synthesis is in agreement with the positive correlation between photorespiration and leaf nitrate reduction (24, 25). This scenario is also consistent with the results obtained in the cytoplasmic male sterile CMSII mutant of tobacco (affected in the mitochondrial respiratory complex I) in which the day respiratory rate is similar or even higher than that of the wild type while both photorespiration and N metabolism (amino acid synthesis) are enhanced (26, 27).

Therefore, we argue that day respiratory homeostasis in leaves is likely to be the result of a compromise between two opposing forces: (i) an inhibition of respiration and glycolysis caused by high mitochondrial NADH levels generated by photorespiration (Gly decarboxylation) and elevated ATP/ADP and DHAP levels generated by photosynthetic activity; and (ii) a stimulation of the TCA cycle to adjust 2-oxoglutarate production to photorespiratory Glu demand. Such a compromise should be very dynamic, adjusting to changes in environmental conditions that modify stomatal closure, thereby altering leaf internal CO₂/O₂ balance. For example, water deficit, which leads to a low internal CO₂ mole fraction, presumably promotes day respiration and photorespiration. Therefore, in the summer months the quantitative significance of these metabolic changes should be evident in many C₃ crops and natural vegetation. However, it is probable that such a promoting effect may disappear on a long-term basis because of acclimation processes (28, 29). Thus the extent to which the regulation of day respiration by CO₂ and O₂ conditions scale up to crop productivity and global carbon sequestration needs further experimental assessment.

Materials and Methods

Plant Material.

Cocklebur (X. strumarium L., Asteraceae) plants were grown in the greenhouse from seed in 100-ml pots of potting mix and transferred to 3-liter pots after 2 weeks. Minimum photosynthetic photon flux density during a 16-h photoperiod was kept at ≈400 μmol·m⁻²·s⁻¹ by supplementary lighting. Temperature and vapor pressure deficit were maintained at ≈25.5/18.5°C and 1.4/1.2 kPa day/night, respectively. The carbon isotope composition (δ¹³C) of CO₂ in the greenhouse air was −9.5 ± 0.3‰. The third or fourth leaves (from the apical bud) were used for all measurements.

Gas Exchange Measurements.

Closed system (dark respiration).

The respiration chamber was placed in a closed system, which was directly coupled to an elemental analyzer (EA) NA-1500 (Carlo-Erba) through a 15-ml loop, as described (30). After decarboxylating the system, respired CO₂ was accumulated until it reached nearly 300 μl·liter⁻¹. The loop was then shunted and the gas inside the loop was introduced into the EA with helium for gas chromatography. The connection valve between the EA and the isotope ratio mass spectrometer (VG Optima; Micromass) was opened when the CO₂ peak emerged from the EA.

Open system (photosynthesis and on-line carbon isotope discrimination).

The photosynthesis system has already been described (9). Briefly, a purpose-built assimilation chamber was connected in parallel to the sample air hose of the LI-6400 system (Li-Cor). Leaf temperature was controlled at 21°C with circulating water from a cooling water bath to the jacket of the leaf chamber and measured with a copper-constantan thermocouple plugged to the thermocouple sensor connector of the LI-6400 chamber/infra-red gas analyzer. Inlet air was adjusted to ≈10 mmol·mol⁻¹ H₂O and passed through the chamber at a rate of 30 liter·h⁻¹, monitored by the LI-6400. Light (400 μmol·m⁻²·s⁻¹) was supplied by a 500-W halogen lamp (Massive). Inlet CO₂ was obtained from a gas cylinder (Alphagaz N48; Air Liquide) with a δ¹³C of −50.2 ± 0.2‰. The outlet air of the chamber was regularly shunted and sent to the loop to measure its ¹²C/¹³C isotope composition and thus the on-line carbon isotopic discrimination (Δ_obs). The gas inside the loop was introduced into the EA for GC as described above. Δ_obs during photosynthesis was measured following the method described (31). Air with 2% oxygen was from a cylinder (crystal gas mixture; Air Liquide). When light was turned off, the leaf was immediately removed from the open system, and one half was frozen in liquid nitrogen. The other half (still attached to the peduncle) was placed in the closed system for dark respiration measurements (see above).

Starch Extraction.

The protocol for starch extraction was similar to that described in ref. 30. The frozen leaf material was lyophilized and powdered. Fifty milligrams of leaf powder was suspended with 1 ml of distilled water in an Eppendorf tube (Eppendorf Scientific). After centrifugation, the pellet was washed four times with 95% ethanol at room temperature, and starch was extracted by HCl solubilization and precipitated with cold methanol. After lyophilization, starch was transferred to tin capsules (Courtage Analyse Service) for isotope analysis.

NMR Analyses.

Leaves used for NMR spectroscopy were fed for 2 h at 21°C and 400 μmol·m⁻²·s⁻¹ with either water (control), ¹²C substrates, or ¹³C substrates in a large Plexiglas chamber (surface area 450 cm²) connected in parallel to the sample air hose of the LI-6400 system (Li-Cor), allowing CO₂ mole fraction monitoring.

NMR measurements were carried out as described (9, 32) from perchloric acid extracts prepared from 5 g of frozen leaf material. Spectra were obtained with a Bruker spectrometer (AMX 400) equipped with a 10-mm multinuclear probe tuned at 161.9 and 100.6 MHz for ³¹P and ¹³C NMR, respectively. The assignment of ¹³C resonance peaks was carried out according to ref. 33. Identified compounds were quantified from the area of their resonance peaks by using fully relaxed conditions for spectra acquisition (pulses at 20-s intervals). Peak intensities were normalized to a known amount of the reference compound (maleate for ¹³C and methyl-phosphonate for ³¹P) that was added to the sample (internal standard).

¹³C-Enriched Molecules.

The positional ¹³C-labeled molecules (99% ¹³C in the considered position) were purchased from Eurisotop. Pyr was dissolved in distilled water and the pH was adjusted to 6.7 with NaOH. To obtain nonfully labeled solutions (Δ_obs experiments), the labeled compounds were mixed with industrial Glc (δ¹³C = −9‰) or pyruvate (δ¹³C = −21‰) from Sigma. The resulting overall composition of Glu and Pyr solutions was checked to be 2,500‰ and 1,400‰, respectively. In each case, the final concentration was 0.015 mol·liter⁻¹. The solutions were fed to the leaves through the transpiration stream.

Calculations.

The procedure used to calculate the decarboxylation rates of ¹³C-enriched substrates in the light from apparent Δ_obs values has already been explained in detail (9). Briefly, the difference between apparent Δ_obs values obtained with and without substrate addition is considered to reflect the additional decarboxylation flux in the light. Using mass balance equations, it can be shown that the decarboxylation rate r_day has the following form:

where d is the flow, S is the leaf surface area, V_M is the molar volume at air temperature, and c_e and c_o are the CO₂ mole fractions in inlet and outlet air, respectively. λ values are ¹³C percentages (using delta values is not possible because of large ¹³C enrichments) in inlet CO₂ (subscript e), outlet CO₂ (subscript o), net fixed CO₂ (subscript fixed), and ¹³C-enriched added substrate (subscript s). This equation holds for homogeneously labeled substrates; it is somewhat changed for positional enrichments to take into account the different origin of decarboxylated CO₂. This is the case when Glu or Pyr is added: the C-1 atom of Pyr is decarboxylated by PDH, whereas the C-2 and C-3 positions are decarboxylated by the Krebs cycle. Similarly, the C-3 and C-4 atom positions of Glu are decarboxylated by the PDH reaction, the other being decarboxylated by the Krebs cycle. Taking advantage of positionally ¹³C-enriched substrates: [1-¹³C]pyr would specifically enrich the CO₂ produced by PDH, whereas [2-¹³C]Pyr would specifically enrich the CO₂ that comes from the Krebs cycle. The same applies to positional ¹³C enrichment in Glu. It should be noted that, in contrast to the argument of ref. 3, any isotopic dilution of the substrate is taken into account as the observed carbon isotope discrimination is always a net value that integrates the decarboxylation of natural (that is, not added) Pyr or Glc molecules, both before and after ¹³C substrate addition.

A similar procedure applies to dark-respired CO₂ measurements. In other words, CO₂ that is produced in darkness (¹³C-percentage λ_global) after a light period with ¹³C-enriched substrate feeding comes from respiratory oxidation of new photosynthates (the ¹³C percentage in the net fixed carbon is λ_fixed), photosynthates from the previous light period in the greenhouse (¹³C-percentage λ_previous), and additional C coming from the ¹³C-enriched substrate fed to the leaf (¹³C percentage λ_s). The night-decarboxylation rate has the following form:

where λ_p is a linear combination of λ_previous and λ_fixed. It is equal to 0.6 λ_previous + 0.4 λ_fixed after 2–3 h in the light under ordinary CO₂/O₂ conditions (during which 200–400 mmol C m⁻² has been fixed), and 0.5 λ_previous + 0.5 λ_fixed after 2–3 h in the light under high CO₂ conditions (during which 400–800 mmol C m⁻² has been fixed) (34). It should be noted that possible variations in these coefficients only introduce very slight errors in the estimate of the ¹³C-enriched substrate decarboxylation r_night, because of the strong ¹³C enrichment of the substrate (that is, the λ_p value is always very small compared with λ_global or λ_s and may be neglected). Again, that relationship is somewhat modified with positional enrichments to take into account the different origin of decarboxylated CO₂ (as described in ref. 9).

Clustering Analysis.

The ¹³C NMR data were represented as an isotopomic array as described 35. The positional isotopic abundances (in ¹³C percentage) relative to the natural ¹³C abundance (1.1%) are indicated by colors so that black cells indicate near-natural abundance, and green and red cells indicate lower and larger than natural ¹³C abundance, respectively. The clustering analysis was carried out with Cluster software, and the array was drawn with TreeView software (both from M. Eisen, Stanford University).