Metabolic fluxes, carbon isotope fractionation and respiration – lessons to be learned from plant biochemistry



Post-carboxylation carbon kinetic isotope fractionation (KIF) is in the focus of interest because it alters the isotopic signal imprinted on newly assimilated organic matter by photosynthetic isotope discrimination in downstream metabolic processes (Badeck et al., 2005; Gessler et al., 2009). KIF during respiration and the dynamic change of the isotopic composition (i.e. δ13C) of respired CO2 is providing insight into the carbon flow through metabolic pathways (Ghashghaie et al., 2003; Pataki, 2005) and thus serves as an important tool to characterise plant physiological adaptations to environmental conditions.

The isotopic signature of respired CO213Cres) is determined by the following factors: (1) carbon source and its δ13C with (2) a heterogeneous intramolecular 13C-distribution and ‘molecule fragmentation’ (fragmentation fractionation; Tcherkez et al., 2004), (3) a possible 13C kinetic isotope effect (KIE) of respiratory enzymes in connection with (4) specific metabolic turnover rates (‘commitment values’) of respiratory substrates. The 13C-enrichment in respired CO2 as often observed in leaves is in general attributed to the release of CO2 from the ‘heavy’ (+ 4.1‰ relative to glucose mean δ13C-value) C3 and C4 positions of glucose (Rossmann et al., 1991; Gleixner & Schmidt, 1997) by the pyruvate dehydrogenase (PDH) reaction. In addition, the PDH reaction has a 13C-KIE on the C1 atom of pyruvate (former C3/C4 of glucose; Melzer & Schmidt, 1987).

Recently, an intensive and interesting discussion in the Forum of New Phytologist (Tcherkez, 2010; Werner, 2010) took place about the interpretation of respiratory isotope signals formed by metabolic fluxes and pathways and how to identify the factors determining δ13Cres.

In this correspondence the following main issues were discussed: KIF during respiration (cf. Schmidt, 2003; Tcherkez et al., 2005); and the phenomenon of light enhanced dark respiration (LEDR), both affecting δ13Cres. With LEDR it is assumed that a change in δ13C is due to transient substrate switches and reorganisation of the tricarboxylic acid (TCA, citric acid or Krebs) cycle (Barbour et al., 2007; Gessler et al., 2009) when light-acclimated leaves experience darkness.

This discussion shows the potential of respiratory isotope signals to identify plant metabolic responses, but also the need for a careful interpretation of such data combined with further specific research. Here, we want to draw the attention to the following biochemical facts, which have not been taken into account sufficiently in the general discussion of respiratory isotope fractionation: the impermeability of the inner mitochondrial membrane for acetyl-coenzyme A (acetyl-CoA), which influences the KIF of citrate synthase (CS) during respiration, and the interplay between mitochondrial malate dehydrogenase (mtMDH) and mitochondrial malic enzyme (mtME) during the light–dark transition, and its effects on LEDR.

The inner mitochondrial membrane, kinetic isotope effects and citrate synthase

Both Werner (2010) and Tcherkez (2010) account for the KIE on the CS reaction (calculated to be 1.025; Tcherkez & Farquhar, 2005) in their discussion. When estimating δ13Cres, they calculate a KIF, depending on the commitment of acetyl-CoA to the TCA cycle. Acetyl-CoA produced by the mitochondrial pyruvate dehydrogenase (mtPDH) reaction is one of the key substrates of the amphibole TCA cycle and is used either for oxidation to CO2 or for biosynthesis of secondary compounds. The irreversibility of the CS, the mitochondrial isocitrate dehydrogenase (mtIDH) and the 2-oxoglutarate dehydrogenase (2-OGDH) reactions make the whole TCA cycle unidirectional in the dark. All acetyl-CoA molecules produced by the PDH reaction inside the mitochondrion are transformed to citrate (via a condensation of acetyl-CoA to oxaloacetate) by the CS reaction, as the inner mitochondrial membrane is not permeable to acetyl-CoA (Voet & Voet, 1995; Bowsher et al., 2008), which is not used otherwise within the mitochondrion (according to Voet & Voet, 1995; Bowsher et al., 2008). If acetyl-CoA is directed to anabolic metabolism outside the mitochondrion, it is transported via a specific shuttle in the form of citrate and not as acetyl-CoA (e.g. Evans et al., 1983) and is regenerated by cytosolic ATP-citrate lyase (ctCL, Fig. 1) (Voet & Voet, 1995; Bowsher et al., 2008). In short, the quantitative conversion of acetyl-CoA by CS will not lead to an in vivo KIF in the ‘acetyl-part’ of citrate (cf. Schmidt, 2003; Hobbie & Werner, 2004). Yet we cannot exclude that a KIF is visible in the ‘oxaloacetate part’ of the citrate (oxaloacetate can be precursor for aspartic acid and descendants, Fig. 1). Tcherkez (2010) and Werner (2010) also discuss a KIF caused by 2-OGDH. However, such a KIF might be of minor importance in vivo. The TCA cycle works as an organised enzyme complex (Srere et al., 1996). The proposed channelling of bulk substrates at reduced concentrations will avoid divergence to other enzymatic reactions (Srere et al., 1996). With such an avoidance of metabolic branching, possible KIEs on IDH and 2-OGDH reactions will not be expressed in vivo.

Figure 1.

Simplified scheme of the complete unidirectional TCA cycle with anapleurotic reactions (dotted lines a–c) known to supply the TCA cycle with carbon skeletons in cases when intermediates have been removed for anabolic reactions (dotted-dashed lines). Scheme modified after Bowsher et al. (2008). Phosphoenolpyruvate and pyruvate are also starting points of important biosynthesis pathways.

As a consequence, mainly the fragmentation fractionation and the KIE on the mtPDH reaction can become effective. However, Tcherkez (2010) and Werner (2010) have shown that neither of these two reactions can fully explain the observed variability in δ13Cres in leaves. On the one hand, the maximum isotope variation between 0% and 100% commitment of acetyl-CoA to CO2 production will be the difference between δ13C of C1 and the global δ13C of pyruvate, that is, 4‰. However, differences in δ13Cres of up to 14.8‰ have been observed (Wegener et al., 2010; Werner, 2010). On the other hand, an incomplete commitment of pyruvate to the TCA cycle would lead to more negative δ13C values of the CO2 released by PDH (KIE, Melzer & Schmidt, 1987), which is in contrast to experimental observations. Thus, factors other than just the PDH reaction or fragmentation must interact with the whole process.

LEDR and the malate catabolising enzymes

Tcherkez (2010) and Werner (2010) indicate that other respiratory substrates – especially malate – could be responsible for the observed high temporal variations in δ13Cres in autotrophic tissues. After darkening of light-acclimated leaves, an increased release (Atkin et al., 2000) of 13C-enriched CO2 has been observed (Barbour et al., 2007; Gessler et al., 2009; Werner et al., 2009) following typical temporal dynamics. Carbon dioxide with high 13C-enrichment is released at first, followed by a strong decline in δ13Cres of up to 5‰ within 30 min. This temporal pattern cannot be explained by KIF on mtPDH and on decarboxylation reactions during the TCA cycle or by a contribution of mtME alone (Werner et al., 2009; see earlier). The malate originating via oxaloacetate from phosphoenolpyruvate carboxylase (PEPc) is 13C-enriched in the C4 atom (Melzer & O’Leary, 1987, 1991). The decarboxylation of this malate by mtME would explain the extent of 13C-enrichment of respired CO2 (Gessler et al., 2009), but would start with relatively 13C-depleted CO2 (due to the probable 13C-KIE of NAD-ME; Grissom et al., 1987), thus producing the opposite temporal dynamics as observed (Gessler et al., 2009; Werner et al., 2009; Werner, 2010). To solve these discrepancies and to find a metabolic explanation for the observed temporal dynamics of δ13Cres during LEDR, again the TCA cycle (Fig. 1) is a good starting point.

In the light, only a partial TCA ‘cycle’ (Fig. 2a) operates in autotrophic tissues (Bowsher et al., 2008; Tcherkez et al., 2009) because the mtIDH (Igamberdiev & Gardeström, 2003), the succinate dehydrogenase (SDH, Popov et al., 2009), and the 2-OGDH (Gessler et al., 2009) are inhibited, and the decarboxylating activity of the TCA ‘cycle’ is generally reduced (Tcherkez et al., 2009). Moreover, reactions from fumarate to citrate have been identified (Tcherkez et al., 2009; Sweetlove et al., 2010), mtPDH is down-regulated (Budde & Randall, 1990), fumarase activity can be increased (Lee et al., 2010), and the surplus of citrate produced is transported to the cytosol (Tcherkez et al., 2009; Sweetlove et al., 2010). The PEPc reaction in leaves of C3-plants is activated under illumination (Duff & Chollet, 1995) to replenish the carbon skeletons used for biosyntheses. The concentration of oxaloacetate (from PEPc) and NADH (from photorespiration) rise and shift the mtMDH reaction towards malate production (Graham & Walker, 1962). MtME in turn is down-regulated (Hill & Bryce, 1992; Igamberdiev et al., 2001) and thus malate levels in leaves are highest at the end of the light period. From seven C-atoms (four malate + three pyruvate C-atoms per one conversion cycle, Fig. 2a) at maximum two CO2 molecules will be produced during illumination by mtPDH and cytosolic ctIDH. The C-atom released as CO2 by ctIDH is derived from C1 of malate (Voet & Voet, 1995; Supporting Information Fig. S1) and acetyl-CoA will not be directly oxidized to CO2.

Figure 2.

Simplified metabolic scheme of TCA (+ anapleurotic reactions) in dependence of the light situation. (a) Noncyclic TCA ‘cycle’ during the day, (b) TCA cycle with increased decarboxylating activity during light–dark transition in light-acclimated plants, (c) steady-state TCA metabolism during the night. Dashed lines, down-regulated or even inhibited reactions. Differences in line strength: increase or decrease in metabolic flux rates for a given reaction between (a), (b) and/or (c) (qualitatively shown). OAA, oxaloacetate; Pyr, pyruvate; Mal, malate; Asp, aspartate; Cit, citrate; mtME, mitochondrial malic enzyme; mtMDH, mitochondrial malate dehydrogenase; mtPDH, mitochondrial pyruvate dehydrogenase. C1, C3, C4, specific carbon position in corresponding molecule (see also Supporting Information Fig. S1). Under particular circumstances, there might also be back transport of Asp and/or Cit from cytosol.

In the dark, the whole TCA cycle will operate as unidirectional cycle (Fig. 2b,c) and during the reorganization in the light–dark transition, the malate pool accumulated during the day will be degraded (Gessler et al., 2009). This decrease of malate concentration immediately after darkening (Igamberdiev et al., 2001; Gessler et al., 2009) cannot be explained by the ‘catabolic’ property of the TCA cycle alone, but implies the interplay of mtME, mtMDH and mtPDH (Fig. 2b part α).

MtME has a lower affinity towards malate as compared with mtMDH (Wedding et al., 1976). A high malate concentration at the light–dark transition triggers the mtME to deliver additional pyruvate for mtPDH (Hill & Bryce, 1992), resulting in an increased activity directly after darkening (Igamberdiev et al., 2001). This should lead to a higher turnover of pyruvate to acetyl-CoA, which in turn would trigger CS activity and TCA decarboxylation. As the malate concentration is diminishing (falling under the ‘activity level’ of mtME) during LEDR, a rapid decrease of the activity to steady-state nocturnal conditions follows and exactly this decrease of mtPDH activity from LEDR to steady-state dark conditions was observed by Werner et al. (2009).

Additionally to normal TCA activity (Fig. 2b part β) during LEDR, a maximum of 50% of the eight C-atoms of two malate molecules can be released as CO2 (Fig. 2b part α) in one conversion cycle. The CO2 released by mtME and mtPDH originates from C1/C4 of malate. The CO2 molecules produced during a first turn of the TCA cycle by mtIDH and 2-OGDH are also from C1/C4 of malate (Voet & Voet, 1995; Fig. S1). The C4 of malate is 13C-enriched (Melzer & O’Leary, 1987, 1991) and since C1 and C4 of malate are equilibrated by fumarase (Gout et al., 1993), which has an enhanced activity when malate accumulates during the afternoon (Lee et al., 2010), both C1/C4 atoms are likely to share comparable 13C-enrichment. The CO2 released during the second catabolic cycling of the TCA will originate from C1 of acetyl-CoA added during the first cycle and from C3 of oxaloacetate also from the first cycle (Voet & Voet, 1995; Fig. S1). We thus postulate that the 13C-KIF on the mtME reaction does not necessarily lead to a continuous increase in δ13Cres. The KIF of the mtME reaction is balanced when mtMDH consumes the remaining 13C-enriched malate, and the C1 and C4 atoms of this malate are subsequently released as CO2 by mtIDH and 2-OGDH. During LEDR, the (mitochondrial) malate pool is decreasing as long as mtME is active, later on, the PEPc-derived ‘malate’ is rapidly diluted with 13C-depleted carbon from glycolysis (complete unidirectional TCA cycle in the dark). All this would be reflected in the δ13Cres starting with strong 13C-enrichment after darkening and – with increasing amounts of carbon from glycolysis released as CO2– a decrease in δ13C with time, exactly as experimentally observed.

The mtME reaction is only of major importance in leaves during the light–dark transition – in agreement with flux literature, assigning only a fraction of < 10% of pyruvate to this pathway (Sweetlove et al., 2010) and measured enzyme activities (Merlo et al., 1993). However, LEDR is not simply a measurement artefact that occurs when light-acclimated leaves are darkened under experimental conditions. Barbour et al. (2011) have clearly shown under ‘real world conditions’ that an increase in δ13C of leaf- and ecosystem-respired CO2 occurs after sunset. The authors estimated the amount of carbon released by LEDR to be up to 175 mg C m−2 ground area per night, depending on the amount of cumulatively fixed carbon in the light period before. There are not enough experimental results yet to estimate the importance of LEDR-driven isofluxes for the temporal dynamics of δ13C of organic matter and respired CO2 on plant, canopy and ecosystem scales. There are, however, first indications that experimental data on tissue and metabolite levels (e.g. from 13C-labelling) can be used in combination with field studies to better address this issue (cf. discussion in Barbour et al., 2011).

After ‘deactivation’ of mtME in the course of darkening, the anapleurotic TCA activity can lead to a decreasing malate and an increasing citrate concentration in autotrophic cells (Graham & Walker, 1962). The maximum CO2 production of the reconstituted TCA cycle per conversion cycle (without mtME) is then three CO2 molecules (one each from mtPDH, mtIDH and mt2-OGDH) out of seven carbon atoms (four malate- and three pyruvate-carbon; Fig. 2c). The real balance depends on the actual fraction of the removed carbon skeletons for biosynthesis and storage (citrate) during darkness.


If we assume the impermeability of the inner mitochondrial membrane for acetyl-CoA and thus a 100% turnover of acetyl-CoA, the CS reaction cannot contribute to isotope fractionation under physiological conditions. The discussed KIF on the IDH and 2-OGDH reactions is at least questionable due to the channelling principle of the TCA cycle enzymes. We propose that the processes associated with LEDR can fully explain not only the 13C-enrichment of the respiratory CO2 during light–dark transition but also its temporal course. The closure of the TCA cycle (i.e. completing the full cycle again), which occurs immediately after the light–dark transition, in connection with the interplay of the malate catabolising enzymes (mtME and mtMDH) facilitates the degradation of the 13C-enriched malate pool (synthesised via oxaloacetate by ctPEPc) accumulated under illumination. The gradual decrease of malate concentration is in agreement with the decrease in δ13C of CO2, explaining the initially high 13C-enrichment of LEDR and the observed temporal dynamics.