Opposite carbon isotope discrimination during dark respiration in leaves versus roots – a review

Authors

  • Jaleh Ghashghaie,

    Corresponding author
    1. Laboratoire d'Ecologie, Systématique et Evolution (ESE), CNRS UMR8079, Bâtiment 362, Université de Paris-Sud (XI), Orsay Cedex, France
    Search for more papers by this author
  • Franz W. Badeck

    1. Consiglio per la Ricerca e la sperimentazione in Agricoltura, Genomics research centre (CRA - GPG), Fiorenzuola d'Arda (PC), Italy
    2. Potsdam Institute for Climate Impact Research (PIK), Potsdam, Germany
    Search for more papers by this author

Abstract

Summary

In general, leaves are 13C-depleted compared with all other organs (e.g. roots, stem/trunk and fruits). Different hypotheses are formulated in the literature to explain this difference. One of these states that CO2 respired by leaves in the dark is 13C-enriched compared with leaf organic matter, while it is 13C-depleted in the case of root respiration. The opposite respiratory fractionation between leaves and roots was invoked as an explanation for the widespread between-organ isotopic differences. After summarizing the basics of photosynthetic and post-photosynthetic discrimination, we mainly review the recent findings on the isotopic composition of CO2 respired by leaves (autotrophic organs) and roots (heterotrophic organs) compared with respective plant material (i.e. apparent respiratory fractionation) as well as its metabolic origin. The potential impact of such fractionation on the isotopic signal of organic matter (OM) is discussed. Some perspectives for future studies are also proposed .

I. Introduction

Until recently, changes in the 13C signal of ecosystem-respired CO2 have been attributed to the changes in photosynthetic discrimination caused by changes in environmental conditions (Fung et al., 1997). However, the generally accepted hypothesis that no discrimination occurs downstream of photosynthetic CO2 fixation is now in question. As leaves are generally 13C-depleted compared with all other organs (e.g. roots, stem/trunk and fruits) in C3 plants, it is suggested that post-photosynthetic discriminations do probably occur, leading to the observed isotopic difference between autotrophic and heterotrophic tissues/organs. Different hypotheses are formulated in the literature to explain this difference. Several studies have shown that CO2 respired by leaves in the dark is 13C-enriched compared with leaf organic matter (see Supporting Information, Notes S1, for references used for Fig. 1(a); and for a review see Ghashghaie et al., 2003; Werner & Gessler, 2011), while it is 13C-depleted in the case of root respiration (Badeck et al., 2005; Klumpp et al., 2005; Schnyder & Lattanzi, 2005; Bathellier et al., 2008; Gessler et al., 2009; Wegener et al., 2010; Kodama et al., 2011; Zhu & Cheng, 2011). The opposite respiratory fractionation between leaves and roots discussed in the following could partly explain the widespread between-organ isotopic differences (reviewed by Badeck et al., 2005; Cernusak et al., 2009). Although the 13C enrichment in leaf-respired CO2 has now been confirmed by different research groups for many C3 species (reviewed by Ghashghaie et al., 2003; Werner & Gessler, 2011), data concerning root respiratory fractionation are scarce (see review by Werth & Kuzyakov, 2010).

Figure 1.

(a, b) Distribution of different classes of apparent respiratory fractionation values (ΔR) of leaves (a) and roots (b) of C3 herbs (dark gray bars), C3 woody plants (light gray bars), and C4 herbs (black bars) under varying growth conditions and measured in the dark using different methods. (c) A few data on stem respiration of five woody C3 species are also presented. References used for leaf data from 23 herbaceous C3, 20 woody C3 (including six coniferous) and four herbaceous C4 species are listed in Supporting Information Notes S1. For roots, data from 12 herbaceous C3 species, eight woody C3 species (including two coniferous) and five herbaceous C4 species listed in Table 1 are used. ΔR is calculated as the difference between the carbon isotope composition of leaf or root material (δ13CS) available in the literature (bulk organic matter, water-soluble fraction or soluble sugars, as respiratory substrates) and that of leaf- or root-respired CO213CR) as a product of respiration. Negative ΔR values correspond to 13C enrichment and positive ΔR values to 13C depletion in respired CO2 compared with the substrate. Vertical dashed lines indicate no respiratory fractionation (i.e. ΔR = 0).

In this paper, after summarizing the basics of photosynthetic and post-photosynthetic discrimination, we mainly review the recent findings on the isotopic composition of CO2 respired (δ13CR) by leaves (autotrophic organs) and roots (heterotrophic organs) compared with respective plant material (i.e. apparent respiratory fractionation) as well as its metabolic origin. We also briefly describe apparent fractionation during stem respiration. The potential impact of such fractionation on the isotopic signal of organic matter (OM) is also discussed, and the effect of difference in the presumable source of respired organic material is assessed. Some perspectives for future studies are also proposed.

II. Photosynthetic carbon isotope discrimination

Discrimination against the heavy carbon isotope 13C occurs during photosynthetic CO2 uptake, leading to 13C depletion in plant OM compared with atmospheric CO2. Photosynthetic discrimination is, on average, c. 20‰ in C3 and c. 4‰ in C4 leaves. It has been extensively studied and robust models have been developed for both plant types (Farquhar et al., 1982; Farquhar, 1983). The simple version of the C3 model is based on the two main discriminating steps: discrimination during CO2 diffusion from the ambient air into the leaves through stomata (4.4‰) and during carboxylation by ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco, 29‰). Although the complete version of the model includes leaf internal resistance to CO2 diffusion and discrimination during day respiration and photorespiration, these aspects are less well investigated (for a review, see Brugnoli & Farquhar, 2000). In some C3 species, leaf mesophyll resistance has been shown to be high enough to lead to a relevant decrease in the CO2 mole fraction in the chloroplasts (relative to the intercellular air spaces), thus decreasing net photosynthetic discrimination compared with the values expected from the simple model (Evans et al., 1986, 2009; von Caemmerer & Evans, 1991). Photorespiratory fractionation has been shown to be up to 12‰ against 13C, thus blurring the on-line photosynthetic discrimination measurements (Lanigan et al., 2008). The impact of day respiratory fractionation on the net photosynthetic discrimination is difficult to determine, because the mitochondrial respiration, mainly the tricarboxylic acid cycle (TCA or Krebs cycle), is strongly inhibited in the light (Tcherkez et al., 2005). Recently, day-respired CO2 has been shown to be 13C-depleted by 0–10‰ compared with C3 leaf OM (Tcherkez et al., 2011). However, the simple version of Farquhar's model has been validated for many C3 species and is often used rather than the complete version, owing to its complexity, to determine internal resistances and photorespiratory discrimination in parallel to Ci : Ca.

According to the simple version of Farquhar's model, the overall discrimination during photosynthetic assimilation of CO2 in C3 leaves is a linear function of the intercellular to ambient CO2 mole fraction ratio (i.e. Ci : Ca). Changes in environmental conditions (e.g. air humidity, light intensity, CO2 concentration) affecting stomatal conductance, and thus Ci : Ca, lead to changes in photosynthetic discrimination in C3 leaves (C4 discrimination is much less dependent on Ci : Ca; for a recent review, see Cernusak et al., 2013). Changes in internal leaf conductance and photosynthetic capacities also affect Ci : Ca through modification of the sink strength for CO2. Thus, carbon isotope composition of plant OM is, in general, considered to properly reflect the carbon isotope discrimination of a given photosynthesizing system during photosynthetic CO2 assimilation. However, the carbon isotope signature of plant OM not only integrates discrimination processes during net CO2 assimilation (i.e. photosynthesis, photorespiration and day respiration, already included in the complete version of discrimination model), but also post-photosynthetic processes, which could potentially discriminate, for example, night respiration and photoassimilate export from source leaves to sink tissues (for a recent review, see Cernusak et al., 2009). Post-photosynthetic discrimination could explain the observed widespread interorgan isotopic differences (Badeck et al., 2005) as well as the changes in leaf OM isotopic composition compared with recent photoassimilates. For instance, night respiration of C3 leaves was shown to release, in general, 13C-enriched CO2 compared with photosynthetic assimilates (reviewed by Ghashghaie et al., 2003; Werner & Gessler, 2011). Such fractionation is expected to change the carbon isotope signature of the remaining leaf material at the end of the night period compared with the photosynthetic products fixed during the daytime (Ghashghaie et al., 2003).

Another consequence of photosynthetic discrimination is that the CO2 left in the atmosphere becomes 13C-enriched. This effect is used for retrieving the distribution of sources and sinks from the spatiotemporal distribution of carbon isotopes in the atmosphere (Fung et al., 1997), as well as for estimating the isotopic signature of ecosystem-respired CO2 using Keeling plots and disentangling the relative contribution of photosynthetic and respiratory fluxes to the net ecosystem exchanges (Pataki et al., 2003; Mortazavi et al., 2006). For these studies, the carbon isotope signal of ecosystem-respired CO2 was assumed to reflect the carbon isotope signature of photosynthetic products and changes in photosynthetic discrimination resulting from changing environmental conditions. However, the hypothesis that no discrimination occurs after net photosynthetic CO2 fixation in the leaves has been refuted in recent years (Ghashghaie et al., 2003; Ogée et al., 2003; Badeck et al., 2005; Bowling et al., 2008). There is now growing interest in post-photosynthetic discrimination as well as its impact on the isotope composition of both plant OM and CO2 respired by different plant organs and ecosystem components (for recent reviews, see Bowling et al., 2008; Brüggemann et al., 2011; Werner et al., 2012), because of the potential impacts on predicted isofluxes, additional constraints on spatiotemporal variation in the component fluxes determining the integrated isotopic composition of the ecosystem-respired CO2 and the attribution of respiratory CO2 to its carbon sources.

III. Post-photosynthetic discrimination

Carbon isotope composition of leaf bulk OM is often used as a reference for photosynthetic discrimination in both plant- and ecosystem-level studies. Badeck et al. (2005) compiled data from the literature on c. 80 plant species, including C3 and C4, herbaceous and woody, as well as annual and perennial plants. They clearly showed that the 13C signal in OM varies between organs in C3 plants, the leaves in most cases being 13C-depleted compared with all other organs: only slightly compared with green stems, c. 1‰ in average compared with roots and c. 2‰ compared with tree trunks. A difference in carbon isotope composition is also observed between different tissues of a given organ; for example, leaf lamina were shown to be 13C-depleted compared with leaf ribs in different herbs and woody species (Badeck et al., 2009 and references therein). Heterotrophic leaves (before photosynthetic onset) are 13C-enriched compared with green autotrophic leaves as well (Bathellier et al., 2008; Lamade et al., 2009). It is concluded that fractionation mechanisms do probably occur after net CO2 fixation in the leaves, leading to the observed 13C differences between autotrophic and heterotrophic tissues/organs. Indeed, in the event that no fractionation occurs downstream of photosynthetic CO2 fixation in the leaves, no between-organ isotopic difference can be expected, because of the conservation law. Various hypotheses about post-photosynthetic discrimination have already been advanced to explain the isotopic differences between autotrophic and heterotrophic organs/tissues – in the main these are as follows:

  1. opposite respiratory fractionation between leaves and heterotrophic organs (Badeck et al., 2005; Klumpp et al., 2005; Bathellier et al., 2008);
  2. export of isotopically heavier assimilates from leaves to sink organs as a result of fractionation during phloem loading/transport (Hobbie & Werner, 2004; Gessler et al., 2007) and/or developmental variation in photosynthetic discrimination with lower Ci : Ca in mature leaves thus exporting 13C-enriched assimilates to sink organs (Francey et al., 1985; Cernusak et al., 2001);
  3. higher rate of CO2 fixation by phosphoenolpyruvate carboxylase (PEPc) in heterotrophic tissues/organs compared with C3 leaves (Terwilliger & Huang, 1996);
  4. differential use of 13C-depleted day vs 13C-enriched night sucrose between leaves and sink tissues (Tcherkez et al., 2004; Gessler et al., 2008);
  5. emission of volatile organic compounds (VOCs) and ablation of surface waxes, both of which are, in general, 13C-depleted compared with photosynthetic products;
  6. discrimination during root exudation (Badeck et al., 2005);
  7. use of assimilates produced under contrasting environmental conditions and thus contrasting in δ13C for asynchronous growth of different organs (Pate & Arthur, 1998; Cernusak et al., 2002, 2005).

Cernusak et al. (2009) recently reviewed the hypotheses explaining why leaves are generally 13C-depleted compared with all other organs.

The basis of these processes is fractionation that occurs at metabolic branching points mainly during C–C bond making or cleavage, resulting in a nonuniform intramolecular 13C distribution as well as between-metabolite isotopic differences. As a consequence, when a given molecule with heterogeneous 13C distribution is cleaved during a given enzymatic reaction, a so-called ‘fragmentation fractionation’ (Tcherkez et al., 2004) occurs, resulting in new molecules carrying the positional 13C signal of the fraction of the educt molecule that makes up the individual product. For instance, heterogeneous 13C distribution within molecules could (in addition to enzymatic isotope effects) result in 13C-enriched or 13C-depleted CO2 evolved during decarboxylation of such molecules, releasing heavy or light carbon atom positions and leaving behind 13C-depleted or 13C-enriched molecules, respectively.

Already within the Calvin cycle in the chloroplasts (daytime), aldolase, which condenses two trioses to form fructose 1-6 bisphosphate, discriminates in favour of 13C, enriching C-3 and C-4 positions of hexose molecules (Rossmann et al., 1991; Gleixner & Schmidt, 1997; Gilbert et al., 2011, 2012), thus enriching transitory starch in 13C, while 13C-depleted trioses left behind are transported to the cytosol, forming sucrose. Consequently, day sucrose is expected to be 13C-depleted relative to the average of all assimilates, while night sucrose coming from degradation of transitory starch will be 13C-enriched. Thus, a differential use of 13C-depleted day sucrose vs 13C-enriched night sucrose is expected to impact on the observed between-organ isotopic differences (see Hypothesis 4 earlier and the model of Tcherkez et al. (2004), as well as experimental evidence in Gessler et al. (2008)). As the fractionation by aldolase already occurs in the chloroplasts, the term ‘post-carboxylation’ fractionation was suggested by Gessler et al. (2008) for fractionation after carboxylation by Rubisco in the chloroplasts. The term post-photosynthetic discrimination is generally used for processes after net CO2 fixation and sugar synthesis in the leaves.

Amongst post-photosynthetic processes that are potentially discriminating, it is primarily respiration that has been investigated during the past decade. By contrast with photosynthetic discrimination, respiratory discrimination is complex, not only because different pathways involving different enzymatic decarboxylations contribute to respiration (i.e. various substrates can be used), but also because the isotope effects of the enzymes involved, and consequently the 13C signature of CO2 evolved, can change depending on the relative commitment of the substrates at metabolic branching points to decarboxylations and to other reactions. Moreover, if more than one substrate with different isotopic signature is involved in feeding respiration, the source carbon used and its isotopic composition cannot be accurately determined. For all these reasons, the term ‘apparent’ respiratory fractionation (denoted ΔR) is generally used (Ghashghaie et al., 2003) to describe the isotopic difference between bulk OM taken as source carbon (or putative substrates, mainly sugars) and the product (overall released CO2).

As phloem sugars are often 13C-enriched compared with leaf sugars, fractionation during phloem loading and transport of assimilates (Hypothesis 2) was suggested (Gessler et al., 2007) and theoretical models were proposed (Hobbie & Werner, 2004; Werner & Gessler, 2011). However, the gradients in sugar isotope composition along the tree trunks reported in the literature are contradictory, that is, there is either 13C enrichment or 13C depletion from the top to the bottom (Keitel et al., 2003; Gessler et al., 2004, 2007; Scartazza et al., 2004; Werner et al., 2012). The effects of other post-photosynthetic fractionation processes, such as, for example, ablation of leaf waxes, emission of VOCs and anaplerotic reactions, on the isotopic composition of OM have not yet been experimentally quantified, except for the pioneering work by Nalborczyk (1978) discussed later.

IV. Apparent dark respiratory fractionation in leaves

The carbon isotopic composition of plant-respired CO2 has long been considered to be similar to plant OM. However, pioneering investigations in the early 1970s showed some variability in carbon isotope composition of dark-respired CO2 by leaves compared with leaf material and thus in ‘apparent’ respiratory isotope fractionation (reviewed by Ghashghaie et al., 2003). Because of this variability and difficulties in identifying substrates used for respiration, this topic was not investigated further for nearly 25 yr. Carbon isotope composition of plant-respired CO2 has long been considered to be equal to that of photosynthetic products, thereby reflecting photosynthetic fractionation only. However, experimental data published during the past decade have clearly demonstrated not only that a nonnegligible ‘apparent’ dark respiratory fractionation occurs in C3 leaves, but also that it is highly variable, confirming the pioneering data. Although CO2 respired in the dark by C3 leaves is shown to be generally 13C-enriched compared with leaf OM (or leaf sugars) under typical conditions, it substantially changes (13C enrichment up to 15‰ or 13C depletion in some cases) among species and with environmental conditions (see reviews by Ghashghaie et al., 2003; Badeck et al., 2005; Werner et al., 2012), showing marked diel dynamics depending on functional groups as well (see the recent review by Werner & Gessler, 2011, and references therein). Published data clearly show this large variability between species and plant types (Fig. 1). Nevertheless, respiratory CO2 of C3 herbs and C3 woody plants is, on average, more 13C-enriched than that of C4 plants (= 0.0023 and = 0.0095, respectively, Wilcoxon test). However, the 13C enrichment in respired CO2 is not significantly different between C3 herbs and woody plants (= 0.0625). Leaf-respired CO2 of C3 species (both herbaceous and woody plants) is generally 13C-enriched. A few exceptions showing 13C-depleted respired CO2 correspond, in the case of herbaceous species, to respiration of young seedlings of sunflower and peanut (Smith, 1971), and thus 13C depletion in respired CO2 may result from the use of lipids as substrate; and, in the case of woody species, to very young Quercus ilex (Werner et al., 2009), probably because the respiration was measured at the end of the night period (and so carbohydrate reserves were exhausted and lipids were used as substrates), and to some coniferous trees (Troughton et al., 1974; Mortazavi et al., 2005). The nature of the organic material used as the presumable source (OM vs water-soluble or a carbohydrate fraction) for the determination of apparent discrimination does not impact on these main findings. In none of the 41 cases of C3 herbaceous leaves for which the isotopic signature of OM as well as water-soluble organic matter (WSOM) or a carbohydrate fraction was reported did the sign of the apparent discrimination differ between calculation on an OM basis or a WSOM/carbohydrate basis. For these cases, OM was, on average, 2.25‰ more negative than WSOM or the carbohydrates. Thus, apparent fractionation calculated on a WSOM or carbohydrate basis qualitatively matches the results on an OM basis, but leads to a lower estimate of discrimination. Among 36 cases of C3 leaves of woody species for which the isotopic signature of OM as well as WSOM or a carbohydrate fraction was reported, only three differed in the sign of the apparent discrimination between calculation on an OM basis and calculation on a WSOM/carbohydrate basis. OM was on average 1.04‰ less negative (i.e. 13C-enriched) than WSOM or the carbohydrates. The three exceptions with changing sign of the respiratory discrimination stem from measurements on Prosopis velutina (see Fig. 2).

Figure 2.

Apparent respiratory fractionation (ΔR) calculated with measurements of carbohydrates (CH; sucrose or starch) or the water-soluble fraction (WSOM) as the putative substrate vs apparent respiratory fractionation calculated with organic matter (OM) as the reference for C3 herbs (left panel) and woody C3 plants (right panel). Open symbols, leaf respiration; closed symbols, root respiration. Solid lines correspond to a 1 : 1 relationship and dashed lines indicate no respiratory fractionation (i.e. ΔR = 0). Positive values (indicated by white arrows) correspond to 13C depletion and negative values (indicated by black arrows) to 13C enrichment in respired CO2 compared with plant material used as putative substrate. Literature data used are those from Fig. 1 for which both OM and CH or WSOM were available.

The observed 13C enrichment in dark-respired CO2 by C3 leaves was initially proposed (Ghashghaie et al., 2001) to be primarily the result of the heterogeneous 13C distribution in hexose molecules mentioned earlier, combined with a higher contribution of pyruvate dehydrogenase reaction, PDH (releasing 13C-enriched CO2 (i.e. C-1 of pyruvate coming from C-3 and C-4 positions of glucose)), relative to TCA cycle (releasing 13C-depleted CO2 (i.e. C-2 and C-3 of pyruvate coming from C-1, C-2, C-5 and C-6 of glucose)), to the overall respiration. Tcherkez et al. (2003) obtained a positive linear relationship (Fig. 3) between leaf δ13CR and respiratory quotient (RQ = CO2 evolved/O2 consumed), demonstrating that when glycolysis could not supply carbon skeletons needed for the TCA cycle (because of the increase in respiration rate with increasing temperature, or because of the decrease in carbohydrate pool size under prolonged darkness), reserves such as fatty acids (known to be 13C-depleted) are oxidized to replete the TCA cycle, thereby releasing 13C-depleted CO2. Under typical dark conditions, the leaf RQ is c. 1, indicating the use of carbohydrates as the main substrate for respiration and that the respired CO2 is 13C-enriched, while under high temperatures and after a long period of darkness, the RQ is much lower, indicating the switch to the use of less oxygenated substrates, such as fatty acids, and that the respired CO2 is 13C-depleted (c. −30‰), approaching the isotopic signature of lipids (Fig. 3, open circles; Fig. 4, right-side panel). According to the intramolecular 13C-distribution values for hexose molecules given by Rossmann et al. (1991), leading to a fragmentation fractionation, leaf δ13CR is expected to vary between c. −21‰ (if only PDH contributes to respiration) and c. −27‰ (if the TCA cycle is the sole contributor), suggesting that the variability in metabolic pathways could lead to the observed variability in leaf δ13CR (Tcherkez et al., 2003). The variability in the isotopic signature of leaf-respired CO2 could thus potentially be used as an indicator of metabolic pathways and substrates used for respiration (see reviews by Ghashghaie et al., 2003; Werner & Gessler, 2011; Ghashghaie & Tcherkez, 2013).

Figure 3.

Variations of carbon isotope composition of CO2 δ13CR respired by attached leaves (open symbols) and washed attached roots (closed symbols) of Phaseolus vulgaris as a function of respiratory quotient (RQ = CO2 evolved/O2 consumed). Regression lines are also presented. Data points correspond to individual measurements on different plants. Leaf data are from Tcherkez et al. (2003) and root data are from Bathellier et al. (2009). Additional data extracted from the literature for several other plant species (tomato, castor bean, peanut, pea and radish) are indicated by stars (James, 1953; Park & Epstein, 1961; Smith, 1971). Redrawn from Ghashghaie & Tcherkez. 2013. Advances in botanical research, chap. 8, vol. 67, © 2013, with permission from Elsevier.

Figure 4.

(a–f) Variations of respiration rate, R (a, d), respiratory quotient, RQ (b, c) and carbon isotope composition of CO2 respired (c, f) by washed attached roots (a–c) and attached leaves (d–f) of Phaseolus vulgaris during continuous darkness. (c, f) Dashed horizontal lines indicate the mean values of carbon isotope composition of carbohydrates (soluble sugars), soluble proteins and total lipids extracted from roots (c) and leaves (f). Closed circles, data published in Bathellier (2008); open circles, data from Tcherkez et al. (2003). Error bars correspond to ± SE (= 3). The leaf mass area index of the first adult leaves at this developmental stage was c. 28 ± 2.25 g DW m−2 (= 3). The leaf respiration rate on mass basis was thus between c. 7 and 31 nmol g−1 DW s−1.

However, the flux balance between PDH and the TCA cycle alone cannot explain the measured variations in leaf δ13CR, ranging between −15‰ and −33‰, observed for different species under varying conditions (Tcherkez et al., 2003; Barbour et al., 2007; Priault et al., 2009; Werner et al., 2009; Wegener et al., 2010), because they exceed the estimated range mainly for the 13C-enriched values. Isotope effects operating by PDH and decarboxylating enzymes involved in the TCA cycle could also influence the overall δ13CR. Moreover, as pyruvate is at a metabolic branching point, the fractionation by PDH will depend on the relative commitment of pyruvate to decarboxylation or other reactions. Therefore, the leaf δ13CR and its variability reflect the concerted influence of metabolic pathways and fluxes, isotope effects of the enzymes involved, as well as the intermolecular 13C distribution in hexoses and the resulting fragmentation fractionation.

In addition, new techniques, mainly tunable diode laser spectroscopy (TDLS; Bowling et al., 2003), allowing high time-resolution measurements during light-to-dark transition, demonstrated strong 13C enrichment in leaf-respired CO2 up to 11‰ compared with phloem sap sugars in Ricinus (i.e. δ13CR was nearly −15‰; Barbour et al., 2007). This result was confirmed using rapid in-tube sampling of respired CO2 on other C3 species (Priault et al., 2009; Werner et al., 2009; Wegener et al., 2010). Such strong 13C enrichment in δ13CR of illuminated leaves immediately after transfer to the dark cannot be explained by decarboxylation of the 13C-enriched carbon atom position of pyruvate during the PDH reaction (expected to be c. −21‰ only) alone, but rather is attributed to a rapid consumption of 13C-enriched malate accumulated during the light period (Barbour et al., 2007; Gessler et al., 2009; Werner et al., 2011). Because of the inhibition of mitochondrial respiration in the light (Tcherkez et al., 2005), organic acids such as malate (or oxaloacetate) accumulate in the cytosol during the light period and are subsequently decarboxylated by malic enzyme in the mitochondria upon transfer to darkness, resulting in a peak in the respiration rate lasting 5–20 min, the so-called light-enhanced dark respiration (LEDR). In C3 plants, malate and oxaloacetate are indeed 13C-enriched metabolites compared with sugars, because they originate from anaplerotic fixation of HCO3 by PEPc. Because hydration of CO2 by carbonic anhydrase discriminates in favour of 13C, HCO3 is enriched in 13C by c. 9‰ compared with CO2, and PEPc discriminates only by c. 2.2‰ against 13C; thus malate issued from the anaplerotic pathway, and thereby CO2 evolved by its decarboxylation upon transfer to darkness (LEDR), are 13C-enriched.

To our knowledge, the impact of respiratory discrimination on the 13C signature of plant OM has not yet been investigated in a systematic manner. A 13C depletion by c. 1‰ in the leaf OM remaining at the end of the night period compared with photosynthetic products was estimated using a simple mass balance calculation and assuming an average value of −4‰ dark respiratory fractionation (Ghashghaie et al., 2003). Thus, the 13C-enriched leaf-respired CO2 during the night could partly explain the generally observed 13C depletion in leaf organic material compared with other organs (for a recent review, see Cernusak et al., 2009).

Reports in the literature on the carbon isotope fractionation during respiration by C4 leaves and by organs other than leaves (e.g. tree trunks/twigs, roots) are still scarce. A few pioneering data on C4 plants (Smith, 1971; Troughton et al., 1974) and preliminary data on maize leaves (Ghashghaie et al., 2003) show a small difference (< 2‰) between respired CO2 and bulk OM or leaf sugars, the difference being either negative or positive (see Fig. 1a). Higher (re)fixation of respired CO2 by PEPc in C4 leaves compared with C3 leaves could be at the origin of the low ‘apparent’ respiratory fractionation observed in C4 leaves, but investigations at the metabolic level are needed to confirm this hypothesis. One exception is the high 13C enrichment (10.5‰) in respired CO2 compared with leaf OM observed for Paspalum dilatatum cultured in sand at 15°C (Schnyder & Lattanzi, 2005).

V. Root-respired CO2 is generally 13C-depleted compared with root OM, except in C3 woody plants

In contrast to leaves, root respiratory fractionation is poorly investigated, mainly because of difficulties in disentangling the relative contribution of roots and soil microorganisms (Grossiord et al., 2012). Understanding of root respiratory fractionation is necessary, in particular because root respiration is a major contributor to soil CO2 efflux and thus is an important component of ecosystem respiration. Root δ13CR can be determined under controlled conditions either on excavated and cleaned/washed roots (or root samples), that is, roots without soil and without microorganisms (except for roots including mycorrhizas), or on a simplified culture system (e.g. hydroponic medium or sand/vermiculite). Otherwise, in the case of in situ respiration measurements, the soil-respired CO2 carries the isotopic signature of both roots and microorganisms of the rhizosphere and thus labelling experiments are needed to distinguish the contribution of each component.

The isotope composition of root-respired CO2 of herbaceous C3 plants (alfalfa, perennial ryegrass and sunflower) cultured in pots filled with sand showed that, in contrast to shoots, root-respired CO2 was 13C-depleted (up to 5.6‰ for ryegrass) compared with root OM but varied with species and with growth conditions (Klumpp et al., 2005; Schnyder & Lattanzi, 2005). Similarly, a 13C depletion by c. 1–3‰ in root-respired CO2 compared with root organic material (or root sucrose) was observed on washed roots (without soil) of bean plants, both attached and detached (Badeck et al., 2005; Bathellier et al., 2008). Whatever the culture support used (peat soil, sand or vermiculite), the respired CO2 of washed roots had the same isotope signature, c. −28‰ for bean plants (Bathellier, 2008). So far, published data show that, similar to leaves, root δ13CR is also variable within species and conditions (Table 1). On average, apparent root respiratory fractionation did not differ between C3 and C4 herbs (> 0.1, Wilcoxon test), while it was significantly different for roots of C3 woody plants compared with C3 as well as C4 herbs (< 0.001). Root-respired CO2 is 13C-depleted compared with root material in herbaceous C3 species, with two exceptions, Melissa officinalis (Wegener et al., 2010) and Triticum aestivum (Kodama et al., 2011), surprisingly showing 13C-enriched root-respired CO2 at the end of the night period, perhaps because the carbohydrate reserves were exhausted and lipids were used as substrates. In contrast to C3 herbs, deciduous C3 trees (Eucalyptus, Acer and Acacia) and shrubs (Halimium halimifolium and Rosmarinus officinalis) have 13C-enriched root-respired CO2. For instance, 13C enrichment in root δ13CR is > 9‰ for Acer negundo (Moyes et al., 2010) and for a Mediterranean semideciduous shrub, H. halimifolium (Dubbert et al., 2012), except when the latter is cultured in hydroponic medium (Wegener et al., 2010). This contrast in root δ13CR values between herbaceous and woody species is not observed for leaves, which show 13C enrichment in respired CO2 for both herbaceous and woody C3 species (see Fig. 1b), with few exceptions in both plant types. Since in trees (and shrubs) the root respiration includes the respiration of associated mycorrhizas (even in cleaned or washed roots), this could explain the opposite root δ13CR values between trees and herbaceous plants. However, δ13CR values should be measured on mycorrhizas alone to confirm this hypothesis. Further experiments in the field and in the pots of the same species should also be conducted to examine a potential effect of field vs potted conditions. Another hypothesis can be built on the assumption that, in lignified roots of woody species, C4-like assimilation of CO2 may occur (J. Bloemen, pers. comm.), that is, high activity of C4 enzymes and subsequent decarboxylation of 13C-enriched C4 metabolites such as malate, similar to the high C4 activity observed on C3 stems and petioles (Hibberd & Quick, 2002) and tree twigs (Berveiller & Damesin, 2008). However, as we do not expect diurnal changes in fixation and release of C4 compounds in roots, this hypothesis requires additional assumptions on variation in time of the rates of assimilation and release; alternatively, it may be an artefact of the release of C4 compounds subsequent to extraction of roots from their natural environment. Structural and metabolic studies on roots of woody species should be undertaken in order to examine this hypothesis.

Table 1. Carbon isotope composition of root material analysed (δ13COM) and root-respired CO213CR) as well as root apparent respiratory fractionation (ΔR) calculated as ΔR = δ13COM – δ13CR
Species and conditionsδ13CR (‰)δ13COM (‰)ΔR (‰)Root materialCulture conditionsMethods for δ13CRSources
  1. DAS, d after sowing; WSOM, water-soluble fraction extracted from root material.

  2. Positive ΔR values correspond to 13C depletion and negative ΔR values to 13C enrichment in respired CO2 compared with root material analysed. Growth conditions and methods for respired CO2 collection for isotope analysis are indicated. C3 woody species (trees and shrubs) are presented separately from the C3 herbs. The coniferous species are indicated by grey shading.

C3 plants (herbs)
Medicago sativa L. (alfalfa)
Low light pretreatment + 2.95OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
High light pretreatment + 1.52OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
High light/high nitrogen + 3.73OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
High light/low nitrogen + 2.84OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
Low light/high nitrogen + 2.39OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
Low light/low nitrogen + 2.68OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
Lolium perenne L. (perennial ryegrass) + 5.39OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
25°C/23°C−24.9−21.7+ 3.2OMSand + nutrientsOn-line δ13CR on attached roots in sandSchnyder & Lattanzi (2005)
15°C/14°C−28.3−22.8+ 5.5OMSand + nutrientsOn-line δ13CR on attached roots in sandSchnyder & Lattanzi (2005)
Helianthus annuus L. (sunflower)
Low density−23.26−20.65+ 2.61WSOMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
Low density−23.26−22.79+ 0.46OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
High density−24.81−21.44+ 3.36WSOMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
High density−24.81−22.81+ 2.0OMSand + nutrientsOn-line δ13CR on attached roots in sandKlumpp et al. (2005)
DAS 40–42−28.97−28.24+ 0.73OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−29.77−28.46+ 1.31OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Phaseolus vulgaris L. (bean)
DAS 22−29.2−27.6+ 1.6SucrosePeat soil + nutrientsOn-line δ13CR on attached washed rootsBadeck et al. (2005)
DAS 22−29.2−28.2+ 1.0OMPeat soil + nutrientsOn-line δ13CR on attached washed rootsBadeck et al. (2005)
DAS 22−29.4−26.2+ 3.2SucroseVermiculite + nutrientsOn-line δ13CR on attached washed rootsBathellier et al. (2008)
DAS 22−29.4−27.8+ 1.6OMVermiculite + nutrientsOn-line δ13CR on attached washed rootsBathellier et al. (2008)
Mean of a 6-d dark period−27.45−26.5+ 0.95SucroseVermiculite + nutrientsOn-line δ13CR on detached washed rootsBathellier et al. (2009)
Ricinus communis L. (castor bean)
Daily mean values−31.5−29.3+ 2.2WSOMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Daily mean values−31.5−28.7+ 2.8OMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Morning−32.5−30.4+ 2.1WSOMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Morning−32.5−29.3+ 3.2OMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Evening−30.5−28.2+ 2.3WSOMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Evening−30.5−28.1+ 2.4OMSand + nutrientsKeeling plots on excised cleaned rootsGessler et al. (2009)
Glycine max L. (soybean)
DAS 40–42−29.66−27.87+ 1.79OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−29.35−27.73+ 1.62OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Triticum aestivum L. (wheat)−27.3−27.1+ 0.20OMSandCO2 trapping by NaOH, attached rootsCheng (1996)
DAS 40–42−27.28−26.48+ 0.80OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−28.85−27.91+ 0.94OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Daily mean values−25.05−24.18+ 0.87WSOMField (sandy-loamy silt soil)Keeling plots on excised cleaned rootsKodama et al. (2011)
End of the night−22.7−24.2−1.50WSOMField (sandy-loamy silt soil)Keeling plots on excised cleaned rootsKodama et al. (2011)
End of the day−28.1−24.7+ 3.40WSOMField (sandy-loamy silt soil)Keeling plots on excised cleaned rootsKodama et al. (2011)
Melissa officinalis L.
End of the night−24.6−27.31−2.71WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
End of the day−28.0−27.13+ 0.87WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
Mean of all root parts (18:00)−28.7−28.1+ 0.6WSOMHydrocultureIn-tube δ13CR on excised rootsWegener et al. (2010)
Mean of all root parts (18:00)−28.7−28.5+ 0.2OMHydrocultureIn-tube δ13CR on excised rootsWegener et al. (2010)
Salvia officinalis L.
End of the night−25.6−24.87+ 0.73WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
End of the day−27.5−24.94+ 2.56WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
Oxalis triangularis A.St.-Hil.
End of the night−28.6−28.78+ 0.5WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
End of the day−29.3−26.82+ 1.2WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
Arachis hypgaea (peanut)
Stage 14 (BBCH)−30.83−29.75+ 1.08WSOMVermiculite + nutrientsIn-tube δ13CR on excised rootsJ. Ghashghaie et al. (unpublished)
Stage 14 (BBCH)−30.83−31.35−0.52OMVermiculite + nutrientsIn-tube δ13CR on excised rootsJ. Ghashghaie et al. (unpublished)
Solanum tuberosum L. (potato)
Tuber (day 23)−27.88−27.09+ 0.79OMGermination in the dark of tubers in air (without soil)On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
Sprout (day 23)−27.42−25.43+ 1.99OM On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
Tuber (day 23)−27.88−26.06+ 1.82Starch On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
Sprout (day 23)−27.42−23.80+ 3.62Starch On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
Tuber (day 23)−27.88−26.33+ 1.55Sucrose On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
Sprout (day 23)−27.42−24.61+ 2.81Sucrose On-line δ13CR on excised sproutsMaunoury-Danger et al. (2009)
C3 woody species (trees & shrubs)
Acer negundo −17.9−27.2−9.3OMMature treesδ13CR of soil & roots – soil without rootMoyes et al. (2010)
Eucalyptus delegatensis L.
March 2005−24.86−26.78−1.92WSOMMature treesKeeling plots on excised cleaned rootsGessler et al. (2007)
March 2005−24.86−28.03−3.17OMMature treesKeeling plots on excised cleaned rootsGessler et al. (2007)
Prosopis velutina
Pre-monsoon−22.2−26.4−4.2OMField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Pre-monsoon−22.2−24.3−2.1SucroseField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Monsoon−23−25.7−2.7OMField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Monsoon−23−23.9−0.9SucroseField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Acacia longifolia L.
May−23.9−25.7−1.8WSOMOpen site treesIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−23.9−27.5−3.6OMOpen site treesIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−19.0−26.8−7.8WSOMOpen site treesIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
August−19.0−26.4−7.4OMOpen site treesIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
May−23.6−26.2−2.6WSOMForest site treesIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−23.6−27.3−3.7OMForest site treesIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−22.5−28.1−5.6WSOMForest site treesIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
August−22.5−27.3−4.8OMForest site treesIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
Rosmarinus officinalis L.
May−22.8−25.9−3.1WSOMOpen siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−22.8−27.8−5.0OMOpen siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−21.4−25.7−4.3WSOMOpen siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−24.3−26.6−2.3WSOMForest siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−24.3−26.9−2.6OMForest siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−20.4−28.0−7.6WSOMForest siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
August−20.4−27.7−7.3OMForest siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
Halimium halimifolium L.
May−23.4−26.3−2.9WSOMOpen siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−23.4−27.0−3.6OMOpen siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−17.4−25.4−8.0WSOMOpen siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−24.1−27.8−3.7WSOMForest siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
May−24.1−27.7−3.6OMForest siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
August−18.5−27.9−9.4WSOMForest siteIn-tube δ13CR on excised cleaned rootsDubbert et al. (2012)
August−18.5−27.5−9.0OMForest siteIn-tube δ13CR on excised cleaned rootsM. Dubbert et al. (unpublished)
Halimium halimifolium L.
End of the night−26.1−23.8+ 2.3WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
End of the day−27.3−24.8+ 2.5WSOMHydrocultureIn-tube δ13CR on excised root tipsWegener et al. (2010)
Mean of all root parts (18:00)−32.2−30.1+ 2.1WSOMHydrocultureIn-tube δ13CR on excised rootsWegener et al. (2010)
Mean of all root parts (18:00)−32.2−29.8+ 2.4OMHydrocultureIn-tube δ13CR on excised rootsWegener et al. (2010)
  Pinus ponderosa −0.4 to + 1.5OMFieldPhotoassimilate labellingDijkstra & Cheng (2007)
  Pinus pinaster−27.31−26.62+ 0.69OMField (sandy podsol)In-tube δ13CR on excised washed rootsEpron et al. (2011)
C4 plants
Paspalum dilatatum
25°C/23°C + 0.6OMSand + nutrientsOn-line δ13CR on attached roots in sandSchnyder & Lattanzi (2005)
15°C/14°C + 5.9OMSand + nutrientsOn-line δ13CR on attached roots in sandSchnyder & Lattanzi (2005)
Zea mays L. (maize)
Full nutrient solution−15.8−15.1+ 0.7OMHydrocultureCO2 trapping by NaOH, attached rootsWerth & Kuzyakov (2005)
0.1 × Full nutrient solution−14.6−14.8–0.2OMHydrocultureCO2 trapping by NaOH, attached rootsWerth & Kuzyakov (2005)
Deionised water−14.2−14.5–0.3OMHydrocultureCO2 trapping by NaOH, attached rootsWerth & Kuzyakov (2005)
DAS 40–42−16.90−13.93+ 2.97OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−18.49−13.98+ 4.51OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Amaranthus tricolor L.
DAS 40–42−20.71−13.64+ 7.07OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−20.91−13.90+ 7.01OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Sorghum bicolor L.
DAS 40–42−20.14−13.41+ 6.73OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
DAS 60–62−19.60−13.41+ 6.19OMSand : perlite : soil (50 : 9 : 1)CO2 trapping by NaOH, attached rootsZhu & Cheng (2011)
Sporobolus wrightii
Pre-monsoon−14.7−13.9+ 0.8OMField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Pre-monsoon−14.7−13.8+ 0.9SucroseField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Monsoon−17.3−13.5+ 3.8OMField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)
Monsoon−17.3−16.5+ 0.8SucroseField samplingIn-tube δ13CR on excised cleaned rootsSun et al. (2012)

Surprisingly, in the case of coniferous trees, root δ13CR is similar to root OM or even slightly 13C-depleted (Dijkstra & Cheng, 2007; Epron et al., 2011). However, the literature on coniferous tree roots is scarce and more data are needed to confirm this result. It is worth emphasising the point that CO2 respired by needles of some coniferous trees also showed an opposite respiratory fractionation compared with leaf δ13CR of deciduous trees. For instance, δ13CR of Pinus radiata needles (plus stem) was 13C-depleted by 3.7‰ compared with OM (Troughton et al., 1974), but this was not the case for other coniferous needles analyzed, probably because the stem was included together with needles in the case of P. radiata. Devaux et al. (2009) extracted high amounts of pinitol from the phloem sap of Pinus pinaster. As pinitol is a metabolite 13C-depleted compared to sugars, one could suggest that its use for respiration explains the 13C depletion in CO2 respired by leaves and roots of conifers compared with other trees. However, more data on conifers are needed and labelled pinitol should also be used to test this hypothesis.

Intriguingly, during germination in darkness, potato tubers and sprouts (which are heterotrophic organs, as are roots) also show a 13C depletion in respired CO2 compared with starch (Maunoury-Danger et al., 2009). For C4 plants, almost no fractionation in maize roots was observed when they were cultured in nutrient solution (Werth & Kuzyakov, 2005), while root-respired CO2 was 13C-depleted up to 7‰ in different C4 species (including maize) cultured in a mix of sand, perlite and soil (Schnyder & Lattanzi, 2005; Zhu & Cheng, 2011).

Similar to leaves, the nature of the organic material used as a presumable source (OM vs WSOM or a carbohydrate fraction) for the determination of apparent discrimination in roots does not impact on these main findings (Fig. 2). In 10 instances of roots of C3 herbaceous plants for which the isotopic signature of OM as well as WSOM or a carbohydrate fraction was reported, only one showed a difference in the sign of the apparent discrimination between calculation on an OM basis and calculation on a WSOM/carbohydrate basis. For these cases, OM was, on average, −0.71‰ more negative than WSOM or the carbohydrates. Thus, apparent fractionation calculated on a WSOM or carbohydrate basis qualitatively matches the results on an OM basis but leads to a lower estimate of discrimination. An exception to this was Arachis hypogea, with apparent discrimination on a WSOM basis leading to depleted (1.08‰) respiratory CO2 in line with all other measurements, while on an OM basis a slight enrichment (−0.52‰) was observed. In 14 cases of roots of C3 woody species for which the isotopic signature of OM as well as WSOM or a carbohydrate fraction was reported, only one differed in the sign of the apparent discrimination between calculation on an OM basis and calculation on a WSOM/carbohydrate basis. OM was, on average, 0.62‰ more negative than WSOM or the carbohydrates (see Fig. 2).

Based on these initial studies, the following patterns appear to emerge: although both leaf and root respiratory apparent fractionations are variable among species, they have opposite signs (with few exceptions), that is, root respiratory fractionation values are positive in C3 plants (13C depletion in root δ13CR compared with root OM), except for deciduous trees and shrubs, while leaf respiratory fractionation is negative (13C enrichment in leaf δ13CR compared with leaf OM) except for some pine trees. The difference in mean apparent respiratory fractionation between leaves and roots is significant at < 0.001 for C3 as well as for C4 herbs (Wilcoxon test), while leaf and root apparent respiratory fractionations are not significantly different for C3 woody species (> 0.1). Given this difference, the question arises as to when the leaf–root δ13CR divergence appears during plant ontogeny.

VI. δ13CR of leaves and roots diverges at leaf autotrophy onset

Interestingly, δ13CR values of both leaves and roots were shown (Fig. 5) to be close to the isotopic composition of seed OM (c. −26‰) in bean seedlings, but diverged upon leaf autotrophy onset: leaf-respired CO2 became progressively 13C-enriched, reaching values of c. −20‰ when leaves were fully expanded, while that of roots became 13C-depleted, reaching values of c. –29‰, corresponding to the values already reported for adult plants (Bathellier et al., 2008). Similar results were also obtained on peanut seedlings, with leaves and roots having δ13CR values similar to those of bean plants despite more negative δ13C values for peanut seeds (c. −29‰) because of high lipid content (J. Ghashghaie et al., unpublished). By contrast, maize leaves show an opposite trend, that is, 13C depletion in respired CO2 during ontogeny (J. Ghashghaie et al., unpublished).

Figure 5.

Changes in apparent respiratory fractionation (ΔR) during early ontogeny calculated as the isotopic difference between the putative substrate (δ13CS) and respired CO213CR) for bean (squares) and peanut (circles) roots (closed symbols) and leaves (open symbols). The horizontal solid line indicates no respiratory fractionation (i.e. ΔR = 0). Sucrose and water-soluble fraction were taken as source carbon for respiration for bean and peanut, respectively. During ontogeny, the δ13C values of soluble fraction varied from c. −27 to −29‰ and from c. −26 to −29.7‰ in peanut leaves and roots, respectively. In bean plants, the δ13C values of sucrose changed from c. −25.5 to −27‰ and from c. −24 to −27‰ in leaves and roots, respectively. As the growth rates were different between the two species, the advance in phenological phases is expressed on the BBCH scale (growth stage of mono-and dicotyledonous plants. Developmental stages; BBCH monograph, 2001, Federal Biological Research Centre for Agriculture and Forestry). The seedlings are heterotrophic until stage 10, which corresponds to the beginning of leaf autotrophy. Values for bean plants are from Bathellier et al. (2008) and for peanut from J. Ghashghaie et al. (unpublished).

The observed divergence between root and leaf δ13CR cannot result from the difference in isotope composition of the respiratory substrate between the two organs because sucrose becomes 13C-depleted in both organs during leaf autotrophy acquisition; that is, the isotopic difference between respired CO2 and sucrose (putative substrate) increases in leaves, while it remains low and constant in roots (Bathellier et al., 2008). Parallel changes in the isotopic composition of bulk OM and sucrose in both leaves and roots suggest that photosynthetic products transported from source leaves to roots progressively change the isotopic signature of both leaf and root OM but with a time lag as a result of the transport.

Changes in respiratory fractionation (ΔR) of roots and leaves during ontogeny of peanut plants match quite well those observed for bean plants (Fig. 5). Obviously, respiratory fractionation, being negligible for both organs at the beginning of germination, changes in opposite directions upon leaf autotrophy onset; that is, ΔR becomes more negative in leaves, reaching −8‰ in both species and positive in roots up to + 3.5‰, presumably because of differences in respiratory metabolic pathways between autotrophic and heterotrophic tissues.

VII. Metabolic pathways potentially implied in 13C depletion in root-respired CO2

Recently, the metabolic origin of root δ13CR compared with leaf δ13CR was investigated in bean plants (Bathellier et al., 2009). Surprisingly, root δ13CR does not follow the leaf δ13CR pattern, remaining low and almost stable whatever the carbohydrate pool size during a prolonged dark period (Figs 3, 4c,f). Indeed, despite a decrease in respiration rate (Fig. 4a) indicating a decrease in carbohydrate pool size, and despite substantial changes in RQ (Figs 3, 4b) suggesting a substrate switch from carbohydrates to less oxygenated metabolites (e.g. lipids), root δ13CR did not change (Figs 3, 4c). Similar results were also observed when roots were detached to prevent assimilate supply from leaves (Bathellier, 2008). This cannot originate from variation in the δ13C value of root metabolites, which might compensate for the switch of respiratory substrate, because all the major root metabolites had invariant δ13C (see Fig. 1 in Bathellier et al., 2009). Clearly, leaves and roots do behave differently, presumably because of differences in respiratory metabolic pathways between autotrophic and heterotrophic tissues. Obviously, root respiration under starvation involves metabolic changes that nevertheless result in respired CO2 with δ13C similar to that under typical nonstarving conditions.

13C depletion in root δ13CR could be partly explained by ‘fragmentation fractionation’ during decarboxylation reactions involved in respiratory pathways; that is, the TCA cycle releases light carbon atom positions of glucose (C-1, C-2, C-5 and C-6) and the pentose phosphate pathway (PPP) releases C-1 of glucose. However, recent results on differences in isotopomer frequencies of C-1 in glucose extracted from different organs (Gilbert et al., 2012) point to the need for further studies on the variability of δ13C released from C-1 glucose during PPP activity. In addition, isotope effects of the enzymes involved in decarboxylations could operate, releasing 13C-depleted CO2. In the PPP, 6-phosphogluconate dehydrogenase fractionates against 13C by c. 9.6‰ during decarboxylation of C-1 of glucose (Rendina et al., 1984). Because it originates from C-3 and C-4 positions of glucose, the C-1 of pyruvate decarboxylated by PDH is expected to be 13C-enriched. However, when pyruvate is only partly engaged in the PDH reaction, the isotope effect of this enzyme will operate at its maximum level, leading to a 13C depletion in CO2 evolved up to 23.8‰ compared with pyruvate (the in vitro value reported by Melzer & Schmidt, 1987). Also, several enzymes associated with the TCA cycle could fractionate against 13C (i.e. citrate synthase fractionates by c. 20‰ (Tcherkez & Farquhar, 2005), NADP-dependent isocitrate dehydrogenase fractionates by up to 5.7‰ (Lin et al., 2008) and 2-oxoglutarate dehydrogenase may fractionate by c. 20‰ (Tcherkez & Farquhar, 2005)), so that TCA cycle-derived CO2 is clearly 13C-depleted. In addition, the anaplerotic CO2 fixation by PEPc may influence the isotope composition of TCA cycle intermediates, unless all the PEPc-derived carbon atoms within oxaloacetate molecules are subsequently decarboxylated (Edwards et al., 1998). Taking all these effects together, root-respired CO2 is expected to be generally 13C-depleted compared with substrates. The question is how the changes in relative activities of these decarboxylating pathways might affect overall root δ13CR.

Positional labelling experiments on attached roots of bean plants immersed in glucose or pyruvate solutions 13C-labelled in C-1, C-2 or C-3 were conducted to estimate the relative contributions of PDH, the TCA cycle and PPP to the overall root respiration (see Fig. 6, redrawn from Bathellier et al., 2009). Labelling experiments under typical dark conditions revealed an important PPP activity (c. 22% of the overall respiration, the same rate as previously reported for maize root tips (24%) by Dieuaide-Noubhani et al., 1995). Similar experiments after a few days of continuous darkness showed that the prolonged dark treatment mainly affected the TCA cycle, which seemed to become notably reduced and fuelled mainly by the lipid/protein recycling, and the ongoing synthesis of glutamate was sustained by the anaplerotic action of PEPc, with no effect on overall root δ13CR (Bathellier et al., 2009). Fig. 7 summarises the relative activities of the metabolic pathways involved in respiration and the prevalence of decarboxylations leading to 13C enrichment or 13C depletion in overall respired CO2 from roots (left-hand panels) and leaves (right-hand panels) under typical dark conditions (beginning of the night) or after a prolonged period of darkness. Root δ13CR was estimated to range between c. −27‰ and c. −30‰ whatever the dark conditions, because the main contributors to the respiration (PPP, TCA cycle and lipid degradation) result in evolution of 13C-depleted CO2 as a result of the ‘fragmentation fractionation’ and the isotope effects of the enzymes involved. This is in agreement with the values observed by Bathellier et al. (2009), shown on Figs 3 and 4(c). These authors suggested that the observed invariance in the isotope composition of root-respired CO2 under continuous darkness could be driven by compensations between both the different fractionating steps and the composition of the respiratory substrate mix (for details of the labelling experiments and calculations, see Bathellier et al., 2009).

Figure 6.

Positional labelling of root respiration using glucose or pyruvate labelled at C-1, C-2 or C-3 positions (adapted from Bathellier et al., 2009). Labelled carbon atom positions are presented as grey circles and corresponding labelled CO2 evolved as grey ellipses. Biosynthetic pathways corresponding to the synthesis of lipids and secondary metabolites from acetyl-CoA, and biosynthesis of amino acids (e.g. aspartate and glutamate) from tricarboxylic acid (TCA) cycle intermediates, as well as the pentose phosphate pathway (PPP) are indicated by dashed lines. The anaplerotic pathway for (re)fixation of CO2 (after its hydration to HCO3) by phosphoenolpyruvate (PEPc) to feed the TCA cycle with oxaloacetic acid (OAA) is also presented.

Figure 7.

Metabolic pathways involved in root (left panels) and leaf (right panels) respiration. The relative importance of the pathways reported by Bathellier et al. (2009) for roots and by Tcherkez et al. (2003) for leaves of Phaseolus vulgaris under typical dark conditions (upper panels) and after prolonged darkness (lower panels) is shown by the thickness of the arrows. The grey areas indicate the prevalence of decarboxylating pathways for each case, with pyruvate dehydrogenase (PDH) and malic enzyme (ME) releasing 13C-enriched CO2 while the tricarboxylic acid cycle (TCA) and the pentose phosphate pathway (PPP) release 13C-depleted CO2. Degradation of lipids feeds the TCA cycle with 13C-depleted molecules as well. In the case of leaf respiration under typical conditions, decarboxylation of malate by ME is high at the light-to-dark transition but decreases rapidly in the dark. Biosynthesis of amino acids (e.g. aspartate and glutamate) from TCA cycle intermediates and the resulting anaplerotic feeding of the TCA cycle with oxaloacetic acid (OAA) via phosphoenolpyruvate carboxylase (PEPc) activity are also shown.

Based on the data discussed earlier, two pathways importantly involved in root-respiratory processes could be at the origin of 13C depletion in root-respired CO2 (i.e. PPP and the anaplerotic pathway). The role of a higher PPP activity in roots (compared with leaves) on the 13C depletion in overall root-respired CO2 has already been studied (as discussed in the three preceeding paragraphs). However, the potential impact of the anaplerotic pathway on root δ13CR is still unknown. Indeed, CO2 (re)fixation by PEPc discriminates in favour of 13C (malate formed is thus 13C-enriched compared with other organic acids of the TCA cycle coming from acetyl-CoA), leaving behind 13C-depleted CO2, which should lead to 13C depletion in root net respired CO2. This should also lead to 13C enrichment in root compared with leaf OM (for a mass balance estimation, see Badeck et al., 2005). The impact of CO2 fixation by PEPc will, however, depend on the rate of reassimilation of respired CO2 by this enzyme.

Furthermore, respiratory metabolism is affected by nitrogen nutrition (NO3 or NH4+) through acid–base regulation, cellular pH stat and the related activities of carboxylases as well as the fate of carbon introduced by these carboxylases (Raven & Farquhar, 1990). Cramer et al. (1993) demonstrated that incorporation of 14C in maize roots is higher when plants are cultured with NH4+. To maintain the cellular pH and supply carbon for biosynthesis of amino and organic acids, CO2 fixation by PEPc via the anaplerotic pathway is activated. Schweizer & Erismann (1985) observed opposite effects of N nutrition on PEPc activity in leaves vs roots of nonnodulated bean plants; PEPc activity was low under NH4+ and high under NO3 nutrition in primary leaves but the reverese was seen in roots. With a proteomic study on starved maize plants, which were subsequently subjected to high NO3 concentration in the culture medium, Prinsi et al. (2009) suggested that the nutritional status of the plant may affect two different post-translational modifications of PEPc, consisting of monoubiquitination (thus reduction of its affinity for PEP) and phosphorylation (activation) in roots and leaves, respectively. They also showed increased amounts of the enzymes involved in PPP (i.e. glucose 6-phosphate dehydrogenase and 6-gluconate dehydrogenase) in roots and proteins involved in the regulation of photosynthesis and also lipid metabolism in leaves. Further investigations on the impact of the anaplerotic pathway on the isotopic signature of both root OM and root-respired CO2 under varying nitrogen nutrition conditions are needed to understand its potential contributions to the between-organ isotopic differences in different species.

In addition, as mentioned earlier, C4 leaves discriminate only slightly during respiration (in contrast to C3 leaves). It would be interesting to examine whether this is related to the high PEPc activity in C4 leaves (analogous to roots having higher PEPc activity and lower respiratory fractionation).

The measurement of root respiration, and specifically of the isotopic signature of root-respired CO2, raises a series of methodological issues because it requires separation of the root-derived signal from that resulting from soil microbial activity. Two main approaches to tackle this issue have been employed so far in the studies cited here. Some researchers used quasi-sterile growth media (sterilized sand or nutrient solutions) while others extracted the roots from the soil and cleaned them before measuring respiration (see Table 1). The advantage of the first method is that in situ measurements on the soil/root system can be done, avoiding potential artefacts as a result of extraction of roots and cleaning. However, this comes at the price of a restricted range of soil types that can be studied. When applying the second method, it cannot be ruled out that the separation of the roots from their in situ environment can lead to alterations in root respiratory metabolism. In particular, if the commitment of different catabolic and anaplerotic pathways depends on ion exchange with the soil medium, exposing the roots to air can be expected to lead to adjustments in the respiratory metabolism. In addition, if roots fix CO2 through PEPc, the fraction of assimilated carbon stemming from respired organic material and thus carrying relatively negative signatures relative to the fraction stemming from ambient air that is less depleted in 13C is most likely going to change after extraction of roots from the soil. In consequence, the isotopic signature of root-respired CO2 will change. Currently little is known about these potential artefacts and the timescales on which they establish.

VIII. A glance at stem respiration

Plant stems connect leaves and roots, providing transport of water and minerals from roots to leaves, transport of assimilates from leaves to roots, and mechanical stability to the above-ground plant organs anchored within the soil. Very often plant stems also contain photosynthetically active tissues and can store reserves. As such, stems are characterized by a complex mix of vertical and radial gas transport pathways as well as often relatively high distances of living cells from the stem surface. In consequence, CO2 exchange fluxes across the stem surface are determined by the stem internal CO2 partial pressure gradients, which, in turn, depend on respiration of diverse stem tissues, CO2 exchange with the fluids transported within xylem and phloem, and stem photosynthesis. In a similar manner, the putative substrates of stem respiration also vary with local stem photosynthesis, import from phloem sap and export to phloem sap. An in-depth discussion of the current state of knowledge about fractionation during stem respiration goes beyond the scope of current review. Here, we only report empirical results on apparent fractionation without addressing the issue of identification of the source carbon. Some aspects of the mentioned issues related to CO2 exchange fluxes across the stem surfaces are dealt with in recent reviews and papers by Teskey et al. (2008) on transport of CO2 within tree trunks, by Cernusak et al. (2001) on stem photosynthesis, and by Damesin et al. (2005) on different methods for CO2 sampling.

Stem-respired CO2 of C3 species is, in most cases, 13C-enriched (Fig. 1c, see Notes S2 for references used). Thus, apparent respiratory discrimination in stems resembles the phenomenon described for leaves with a lower average ΔR of −1.65‰, as compared with a mean of −3.8‰ in leaves. The set of measurements currently available is dominated by experiments on tree stems of mainly European tree species.

Leaf and stem respiration during ontogeny of current-year shoots of Fagus sylvatica from sleeping buds to 3 months after bud-burst showed negative ΔR for both organs throughout the observation period (Eglin et al., 2009). Heterotrophic buds had higher δ13CR because 13C-enriched reserves were used, while respired CO2 of photosynthesizing leaves and stems progressively became more 13C-depleted after bud-burst.

IX. Impact on carbon isotope composition of plant OM

Post-photosynthetic discrimination, discussed earlier, could lead to 13C depletion in the bulk OM of autotrophic organs (i.e. leaves) relative to the average isotopic composition of photosynthetic assimilates, mainly because of generally 13C-enriched CO2 evolved during the night. An opposite trend will result from exhalation of 13C-depleted volatile compounds evolved from leaves or ablation of leaf waxes in some species. Heterotrophic organs could become 13C-enriched mainly because of 13C-depleted CO2 evolved via PPP during respiration, CO2 (re)fixation by PEPc (Badeck et al., 2005) and probably because of the transport of 13C-enriched assimilates from autotrophic source to sink heterotrophic organs (Gessler et al., 2007). Such changes in the isotopic signature of organic material, as a result of post-photosynthetic discrimination compared with that of photosynthetic primary products carrying the isotopic imprint of photosynthetic discrimination, could introduce bias in the use of such signatures as references in both ecosystem partitioning studies and in the estimation of water-use efficiency.

However, the impact of post-photosynthetic discrimination on the isotopic signature of OM has not been experimentally demonstrated as yet, except by Nalborczyk (1978), who experimentally showed for different plant species that the 13C : 12C ratio (natural abundance) in leaf OM increases with increasing rates of 14C-labelled CO2 fixation by PEPc in the dark (Fig. 8). Further investigations on the potential impact of discrimination during different post-photosynthetic processes on isotope composition of plant OM are needed to improve our interpretation of the isotopic signature of plant OM as well as between-organ differences.

Figure 8.

Carbon isotope composition (δ13C) of leaf bulk organic matter as a function of carboxylation rate in the dark, measured using 14C-labelled CO2 on sunflower, tomato, rape, barley, rye, cucumber, wheat and lupine. δ13C was determined in ambient air before labelling on plants grown under low light (drawn using data from Nalborczyk, 1978).

X. Conclusions

Despite high variability in both leaf δ13CR and root δ13CR, and despite few exceptions for both organs, the present review has collected evidence of opposite ‘apparent’ carbon isotope discrimination during respiration in roots as compared with leaves in herbaceous species. In contrast to leaves, which in general release 13C-enriched respired CO2 compared with leaf OM, the CO2 evolved by roots is 13C-depleted compared with root material. This could partly explain the between-organ isotopic difference already reported in the literature.

Interestingly, significant differences between functional groups (C3 herbs vs C3 woody species, and C3 vs C4 herbs) are seen mainly for roots. While leaf δ13CR is 13C-enriched in both C3 herbs and C3 woody species (and also in some C4 herbs), roots of C3 herbs show opposite respiratory fractionation compared with roots of C3 woody plants – root δ13CR being 13C-depleted in C3 herbs and 13C-enriched in C3 woody species compared with root material. The respiration of mycorrhizas associated with tree roots could explain the opposite respiratory fractionation observed between roots of C3 herbs and C3 woody species.

The route to further progress in exploring the causes of these patterns will involve scrutinising potential artefacts during measurements of root respiration; assessing the role of PEPc activity; and studying differences in root fractionation processes in soils/substrates differing in pH and ammonium vs nitrate availability.

The divergence in 13C between leaves and roots in C3 herbs is shown to establish upon the heterotrophy–autotrophy transition, suggesting that heterotrophic tissues/organs behave differently from autotrophic ones with respect to respiratory metabolism. The relative activities of metabolic pathways releasing 13C-enriched or 13C-depleted CO2 (PDH and malic enzyme vs TCA cycle and PPP) are shown to be at the origin of the isotopic signature of leaf- vs root-respired CO2. Metabolic studies similar to those discussed should be conducted during ontogeny to elucidate the metabolic origin of the observed divergence at autotrophy onset.

The analyses of apparent respiratory discrimination based on OM as reference material or, alternatively, based on carbohydrates or WSOM as putative substrate for experiments that provided both the isotopic signature of OM and carbohydrates or WSOM did not provide any evidence of contradictory results on apparent respiratory discrimination between the use of the two groups of references. The sign of the apparent respiratory discrimination essentially did not change, while the absolute magnitudes differed as a result of systematic differences between the isotopic signatures of OM and carbohydrates.

The results presented in the current paper help to constrain the analysis of carbon isotope exchange fluxes in ecosystems with multiple sources of respiratory CO2 (Ogée et al., 2003; Werner et al., 2007; Wingate et al., 2010) and allow a better understanding of the origin of soil CO2 isotopic signatures (Brüggemann et al., 2011). Furthermore, they indicate that there are good prospects for the development of on-line carbon isotope exchange measurements (Barbour et al., 2007), which can be applied to in vivo diagnosis of active metabolic pathways.

Acknowledgements

The authors are grateful for financial support through the SIBAE network (COST Action: COST ES0806) coordinated by Nina Buchmann, which facilitated fruitful discussions with the SIBAE members. We also thank these colleagues for their encouragement to prepare a review on this topic. Many thanks also to the colleagues who kindly shared their data (even unpublished in some cases).

Ancillary