Do metabolic fluxes matter for interpreting isotopic respiratory signals?


Intense efforts are currently devoted to disentangling the 12C : 13C isotopic signal of plant respiration as highlighted by papers included in the Virtual Special Issue of New Phytologist, ‘Probing the carbon cycle with 13C’ (; Norby, 2009). Assigning the metabolic origin of the natural 13C-abundance of CO2 evolved by respiration remains difficult, although significant advances have been highlighted recently (Barbour & Hanson, 2009). For example, Barbour et al. (2007) and Gessler et al. (2009) provided the evidence that in Ricinus leaves, 13CO2 evolved by the post-illumination respiratory peak (known as light-enhanced dark respiration, or LEDR) is caused by the decarboxylation of 13C-enriched malate that has just been synthesized from phosphoenolpyruvate carboxylation in the light. By contrast, in the steady state, the isotope composition of CO2 evolved by night respiration shows quite large variations (up to 10‰) between organs or sampling times (see the review of Bowling et al., 2008). Furthermore, the nonsteady isotopic signal of CO2 evolved by darkened, light-acclimated leaves shows substantial variations during the light period and between species (Hymus et al., 2005; Priault et al., 2009). In an attempt to find metabolic explanations for such isotopic variations, several authors used calculations, made qualitative predictions or took advantage of 13C-enrichment experiments (Pataki, 2005; Werner et al., 2007, 2009; Bowling et al., 2008; Priault et al., 2009). However, manipulating isotopic patterns at either natural or enriched abundances requires caution, and in this letter, the importance of considering biochemical pathways and their associated metabolic flux partitioning is shown to be critical for interpreting isotopic mass-balances. That is, basic knowledge of isotopic and metabolic biochemistry should be kept in mind to avoid erroneous conclusions.

First, interpretation should be informed by an understanding of compound-specific signals and their transformation. A common assumption (Pataki, 2005; Bowling et al., 2008) is that isotopic mass-balance accurately describes a process by which 13C-enrichment of CO2 evolved by pyruvate dehydrogenase is naturally compensated for by 13C-depletion in the second product of this enzyme, acetyl-CoA (or in metabolites downstream). Such an argument does not hold, though, because the carbon atom positions in pyruvate involved in CO2 and acetyl-CoA production are not the same (C-1 for CO2, and C-2 and C-3 for acetyl-CoA). Rather, the isotopic mass-balance applies to the specific atom position. To apply the mass-balance approach, one requires knowledge of the 12C : 13C isotope effect of pyruvate dehydrogenase and the amount of the pyruvate pool that is consumed by this enzyme (or the flux value). The known fractionation in C-1 of this enzyme is 23‰ against 13C, thereby enriching in 13C pyruvate molecules left behind (this value is the isotope effect found in baker’s yeast by Melzer & Schmidt, 1987). Importantly, the isotope effect must virtually disappear when the metabolic flux through the enzymatic activity is very large with no branching point, so that there is a full commitment of pyruvate molecules into the reaction.

Such an example highlights the fact that isotope fractionations and metabolic fluxes are intrinsically linked. In mathematical terms, for a given metabolic reaction, the steady 13C-enrichment seen in the substrate is fΔ (where f is the metabolic flux and Δ is the 12C/13C fractionation of the enzyme) and not simply Δ (for a more complete development of mathematics, see Hayes, 2001). Several calculations have assumed constant fractionations, although the parameters used for the calculations are not explicitly given (Werner et al., 2009). However, the prediction of isotopic abundances is extremely sensitive to the architecture of the metabolic network and to associated fluxes.

As such, as suggested by Bowling et al. (2008), it is possible that large fluxes (respiration rates) might correlate with the natural 13C-abundance in evolved CO2, because the large metabolic commitment of decarboxylases would eliminate isotope fractionations against 13C (Rayleigh effect). However, the experimental observation is that fast-growing plants with high respiration rates produce less 13C-enriched (i.e. more 13C-depleted) CO2 than slow-growing plants. The general belief (Bowling et al., 2008; Werner et al., 2009) seems to be that in slow-growing woody plants, a higher proportion of carbon (namely, acetyl-CoA) is diverted from the tricarboxylic acid (TCA) cycle into secondary metabolism. If true, this would have two consequences: first, the metabolic flux through the pyruvate dehydrogenase should be very large, and thereby accompanied by little isotope effect (high commitment); and second, the metabolic commitment into the TCA cycle would be very small (so with a large isotope effect).

Let us consider a numerical example, in which source pyruvate has a natural isotope composition of −25‰ in C-2 and C-3, and −21‰ in C-1 (values taken from Rossmann et al., 1991). With a commitment of 100% into pyruvate dehydrogenase, CO2 evolved by the enzyme has a signature of −21‰. With a low commitment of 5% into the TCA cycle, we would have a fractionation of c. 20‰ (fractionation by the citrate synthase corrected for a commitment of 5%), giving a TCA-derived CO2 at ((−25−20)+(−25))/2 = −35‰ (the factor 2 comes from the fact that two CO2 are produced by the cycle). This would give a total CO2 of ((−21 × 100) – (35 × 5))/105 = −21.6‰. In the reverse case (large commitment into the TCA cycle), the calculation is simpler, as all the pyruvate molecules committed to respiration are consumed, so that total evolved CO2 is at −21/3 – (25 × 2/3) = −23.6‰. Obviously, the isotopic difference is not large, and cannot account for the observed natural variations. In addition, the present calculation is very crude, as it does not take into account the isotope fractionation by the 2-oxoglutarate dehydrogenase, which is presumably as large as that of the pyruvate dehydrogenase (near 23‰). In other words, the theoretical value of −21.6‰ is more likely c.−22‰ or so. Other reasons have to be found (and explored) to explain the isotopic patterns in respired CO2, such as the involvement of other substrates (13C-depleted lipids, Tcherkez et al., 2003) or other pathways (e.g., the pentose-phosphate pathway, Bathellier et al., 2009) associated with decarboxylations.

In such a framework, are 13C-enrichments (with 13C-pyruvate) simplistic indicators of pyruvate dehydrogenase vs TCA-cycle commitments? Certainly not. Carbon-13 enrichment experiments need very controlled conditions (i.e. photosynthetic source CO2 with a known isotope composition) and must be compared with something relevant. First, as different positional 13C-enrichments are used (e.g. 13C-1-pyruvate and 13C-3-pyruvate), controlling the absorption of the 13C-substrate is critical so as to check that the different positionally labelled molecules are absorbed with the same efficiency. Second, the ratio of decarboxylation rates in the dark and in the light is the only informative parameter, simply because the steady dark conditions give a ‘reference value’ to be compared against. In other words, it is not possible to directly compare the isotope composition of evolved CO2 in different species. In fact, they may exhibit either different transpiration rates or different internal 13C-substrate-transport rates and, as a result, they may have different 13C-substrate absorption through the transpiration stream. Keeping this in mind, when compared with the dark, fast- and slow-growing plants may have a very similar maximal velocity of 13C-pyruvate decarboxylation in darkened light-acclimated leaves, but different internal pyruvate concentration and, consequently, contrasted apparent decarboxylation rates. Third, pyruvate may also be produced or used by other metabolic pathways. Typically, endogenous 12C-pyruvate evolved by malic enzyme may dilute exogenous 13C-pyruvate, and this effect depends upon photosynthesis: the larger the assimilation, the larger the isotopic dilution. With so many critical effects, the conclusions on post-illumination evolved CO2 given by Priault et al. (2009) need reassessment.

It is very likely that the pyruvate dehydrogenase : TCA imbalance hypothesis alone does not explain natural 13C-differences between species or ‘functional groups’. For natural species in which carbon metabolism may be complex, it remains difficult to draw simple conclusions from results of 13C-labelling experiments without a consistent body of metabolic compositions and fluxomics. The recognized involvement of secondary metabolism supposedly originating from acetyl-CoA needs more direct assessment, as many secondary products – such as lignins – do not involve acetyl-CoA as a precursor (in the case of lignins and flavonoids, precursors are phosphoenolpyruvate and erythrose-4-phosphate). Let us consider here a numerical illustration. Lignin is depleted in 13C because of isotope effects associated with the biosynthetic pathway (e.g. phenylalanine ammonia lyase as well as methyl transfers) and this, in turn, enriches in 13C corresponding atom positions in precursor molecules left behind. With a lignin content of c. 0.05–0.15 g C g−1 C (Hatfield & Fukushima, 2005) and a 6‰ depletion compared with cellulose in herbaceous species (Benner et al., 1987), this would cause a 13C-enrichment in precursors by c. 0.6‰. In sclerophyllous species that have a larger lignin content, lignin is less 13C-depleted (by c. 4‰), eventually causing a quite similar 13C-enrichment in precursors. Therefore, the relationship between the commitment into secondary metabolism and the isotopic signal of respired CO2 is not straightforward. Undoubtedly, the way in which metabolic pathways combine to produce variation in natural 13C-abundance of respired CO2 should be addressed more carefully in the future, with appropriate fluxomics techniques. Otherwise, irrelevant arguments may be articulated.