- Top of page
- Materials and Methods
The 13C : 12C isotopic ratio of leaf-respired CO2 (δ13Cr) is a powerful tool for elucidating key physiological, ecological and atmospheric processes (see Ghashghaie et al., 2003 for a recent review). However, current difficulties in measuring and predicting δ13Cr, and uncertainties associated with the factors influencing variation in it, still hinder the quantitative utilization of this isotopic signal to trace carbon fluxes in plants and ecosystems.
In contrast to early assumptions, recent studies have established that leaf-respired CO2 can be 13C-enriched compared with key respiratory substrates (Duranceau et al., 1999; Ghashghaie et al., 2001; Tcherkez et al., 2003; Xu et al., 2004). The theoretical basis for this enrichment has recently been revisited and provides a mechanistic base for fractionation during dark respiration, and its dependence on environmental conditions and species (Ghashghaie et al., 2003). A nonstatistical distribution of 13C and 12C within glucose molecules provides the foundation for both variations in δ13C of key metabolites (Schmidt & Gleixner, 1998; Hobbie & Werner, 2004) and observed ‘apparent’ fractionation in respiration (i.e. the difference between δ13Cr and the δ13C of the expected substrate pool) (Rossmann et al., 1991; Gleixner & Schmidt, 1996; Tcherkez et al., 2004). During the decarboxylation of pyruvate formed during glycolysis, by pyruvate dehydrogenase (PDH), carbon atoms from the C3 and C4 position in the original glucose molecule are evolved as CO2. Because these positions are 13C-enriched relative to the glucose molecule, this reaction liberates 13C-enriched CO2, while acetyl coenzyme A (acetyl-CoA) relatively depleted in 13C is formed (DeNiro & Epstein, 1977; Melzer & Schmidt, 1987). Tcherkez et al. (2003) highlighted how, under these circumstances, the degree to which CO2 released during leaf dark respiration is 13C-enriched could be controlled by the balance between PDH activity and subsequent acetyl-CoA oxidation in the Krebs cycle. Although laboratory manipulations provide good means to identify biochemical and physiological responses, the stimuli are usually isolated and enhanced and it is often complicated to predict the applicability, or the relevance, of lab results to field conditions. Fluctuating environmental and physiological conditions in the field are often short-term and mild and involve complex feedbacks, making specific field experiments a critical step in applying new discoveries in the lab to the ecosystem scale.
Despite recent advances in understanding ‘apparent’13C fractionation during leaf dark respiration, there has been no attempt to extend these process-based studies to the field, and specifically no attempt to assess variability in δ13Cr under field conditions. Recent controlled environment studies have provided a middle ground between the lab and the field. Recently, Xu et al. (2004) demonstrated the enrichment of δ13Cr relative to possible substrates under greenhouse conditions and in Biosphere 2, and emphasised the difficulties in predicting δ13Cr from substrate analysis, whereas Klumpp et al. (2005) performed a novel growth-chamber study to show how δ13Cr of different plant components may integrate into whole-canopy δ13Cr. In contrast, the scope of field studies has been more limited, typically highlighting the large range of δ13C of respired CO2 of specific ecosystem components, and reconciling them with the isotopic composition of CO2 respired from the ecosystem as a whole or the soil component, rather than looking for short-term variability in δ13C of respired CO2 (McDowell et al., 2004; Tu & Dawson, 2005; for reviews, see Yakir & Sternberg, 2000; Pataki et al., 2003). Understanding variability in δ13C of CO2 respired by land ecosystems is important, as these isotopic labels are used in global scale models as a tracer to constrain CO2 fluxes between the land biosphere and the atmosphere, and to partition the effects of land and ocean effects on the atmospheric CO2 budget (Yakir, 2003; Ciais et al., 2005). Critically, these existing models generally assume that respired CO2 has the same isotopic composition as bulk organic material, and is invariable in time, an assumption that may be unjustified.
As we have highlighted, there is a gap between laboratory studies, which indicate possible environmental effects on δ13Cr via changes in the PDH : acetyl-CoA oxidation ratios, and field studies that must account for such effects. In an effort to address this issue, here we tested the hypothesis that leaf δ13Cr in forest canopies would display significant variation over a 24-h period that was independent of variation in the δ13C of possible respiratory substrates.
Materials and Methods
- Top of page
- Materials and Methods
In early August 2003, samples of leaf-respired CO2 and leaf material were taken simultaneously with leaf gas-exchange measurements, on six occasions during a 24-h period in two forest canopies in central Italy. These measurements were made on two different days 1 wk apart, first in an c. 50-yr-old Quercus ilex (L.) forest at Castelporiziano (41°45′ N, 12°22′ E) and second in an 11-yr-old Quercus cerris (L.) plantation at Roccarespampani (42°23′ N, 11°51′ E). Each sampling took place at three heights: the top (c. 15 m), middle (c. 10 m) and bottom (c. 5 m) of each canopy, with each level being accessible from a canopy access tower. With respect to the photosynthetic photon flux densities (PPFDs) incident on the top of the canopy at midday, only 30% and 15% reached the middle, and 7% and 5% reached the bottom of the Q. ilex and Q. cerris canopies, respectively. The two measurement campaigns were performed midway through a dry summer on days that experienced weather conditions (temperature, humidity and cloudiness) that were representative of the season, based on our long-term experience at the sites.
Collection and analysis of leaf-respired CO2
At each sampling, leaf area was removed from three branches on each of three trees at each of three canopy heights. In total, at least 0.05 m2 was removed from each tree as groups of leaves attached to a small amount of branch. The leaf area collected from each tree was immediately placed inside one of three dark, custom designed 0.5 l Perspex® cylinders (Fig. 1). Leaf sampling and CO2 collections were made at three heights within the canopy. However, sampling and collection was completed for each canopy height before leaves from the next canopy height were sampled. Screw caps containing O rings sealed the cylinder before the two-step procedure of collecting leaf-respired CO2 into a 200-ml glass flask began. First, both the cylinder and the glass flask, into which the respired CO2 was ultimately collected, were scrubbed of CO2. This was achieved by drawing ambient air through a soda lime column placed in line before both the cylinder and then the glass flask. An infrared gas analyzer (IRGA) (LICOR 7000; LICOR Inc., Lincoln, NB, USA) placed after the glass flask was used to pull the air through the system at 1 l min−1 and to monitor the CO2 mole fraction of the air leaving the glass flask. As CO2 free air was drawn through both the cylinder and glass flask, the CO2 mole fraction declined and ultimately reached a steady-state value, which reflected the addition of leaf-respired CO2 into the CO2 free cylinder, typically 0–5 µmol m−2 s−1. This point usually took less than 5 min to reach. At this point, the pump was switched off and the leaves were left to respire in the cylinder. The second part of the process was the collection of the respired CO2. After c. 10 min of incubation, CO2-free air was pumped through the collection system and glass flask, but bypassing the leaf chambers, to ensure the lines and flask were free of CO2 (monitored on the IRGA). The flow of CO2-free air was then diverted through the cylinder, and the respired CO2 collected in the cylinder was carried through a DRIERITE® column (WA Hammond Drierite Co., Xenia, OH, USA) into the 200-ml glass flask. The positioning of the IRGA after the flask was used to follow the accumulation of CO2 in the flask. As the air was passed through the leaf chambers the CO2 mole fraction, measured by the IRGA, increased steadily to a maximum of at least 400 µmol mol−1. At this point, the pump was switched off and the glass flask was sealed. This protocol ensured that respired CO2 had been collected a maximum of 30 min after the removal of the leaves from the trees.
Figure 1. Schematic of leaf-respired CO2 collection system. The CO2 collection system enabled incubation of leaves from three trees at one of three canopy heights at a time. The system enabled incubation of leaf-respired CO2 in CO2-free air, owing to the positioning of a CO2 scrubber before the incubation cylinders. After an incubation period, an infrared gas analyzer (IRGA) drew air through the system at 1 l min−1 and was used to monitor the build up of dry leaf-respired CO2 in a 200 ml collection flask. Each asterisk denotes a junction in the tubing that can be opened or closed to direct air flow through specific incubation chambers, or to bypass the chambers and flush the system and flask with CO2 free air prior to collection. Thick arrows indicate common air flow lines; thin arrows indicate individually selected flow lines depending on the configuration. See the Materials and Methods section for more detail.
Download figure to PowerPoint
The custom designed cylinders and the collection method were designed to protect against leaks and to avoid fractionation. Tests showed that, over a 15-min sampling period, the maximum observed leak rate of ambient CO2 was < 2 µmol CO2 mol−1, orders of magnitude smaller than respiratory flux over the same time. Extensive testing with known δ13CO2 showed no fractionation, and good agreement between this incubation approach, with the much more labour-intensive Keeling-plot approach (Pataki et al., 2003). In addition, numerical modelling of this incubation approach using simple diffusion models indicated diffusion fractionation effects of < 0.5, which was not significant for this study (K. Maseyk, unpublished data).
Subsequent measurement of δ13Cr was made on a subsample of c. 2.0 ml of air extracted from each 200-ml flask. The CO2 was cryogenically trapped from this sample, separated from N2O using a Carbosieve G packed column (SIS Inc., Ringoes, NJ, USA) at 70°C, and the isotopic ratio determined using a continuous-flow isotope ratio mass spectrometer (IRMS) (20–20; Europa, Northwich, UK). Five flasks filled with calibrated standard air were measured with every 10 sample flasks.
Extraction and analysis of leaf carbohydrates
Extractions of leaf soluble sugars, starch, lipids and cellulose were performed on subsamples of the leaf material removed from three trees at each canopy height for the CO2 collections (described earlier in this section). Each subsample contained leaves taken from at least three branches (up to 0.015 m2 leaf area) and was immediately killed in liquid N2 before undergoing analysis of key leaf extracts. Extraction was performed using the procedures and protocols described in Wanek et al. (2001). Briefly, leaf soluble sugars and lipids were extracted from 100 g of leaf material into a 1-ml methanol, chloroform, water solution (MCW) (12 : 5 : 3, v/v/v). Starch extractions were prepared using enzymatic hydrolysis with α-amylase, while the cellulose fraction was purified from the residue left after soluble sugars, lipid and starch extraction with a solvolytic method using acidified diglyme.
Measurement of the δ13C of leaf soluble sugars (δ13Css), starch (δ13Cst), lipids (δ13Cl) and cellulose (δ13Cc) was determined by combustion of smooth tin capsules (Elemental Microanalysis Ltd, Okehampton, UK) containing the extracts in an elemental analyser (Carlo Erba 1108; Carlbo Erba, Milano, Italy). The CO2 produced during combustion was separated with a PoropaQ GC column (80 mesh, at 60°C) and carried on-line to a continuous-flow IRMS (Optima; Micromass Ltd, Manchester, UK). Calibration of each run to was done by measuring four samples of the Acetanilide (Elemental Microanalysis Ltd #B20CC) international standard at the start of each run and two samples of a cellulose laboratory working standard for every 12 extract samples, with a correction applied to account for the blank cup effect.
The δ13C of leaf-respired CO2 and leaf extracts (Rs) was expressed relative to Vienna Pee Dee Belemnite (VPDB) (RVPDB) as
- δ13C = [(Rs/RVPDB) − 1] × 1000
with a precision of ±0.1.
Leaf gas exchange, photosynthetic discrimination and cumulative CO2 uptake
Simultaneously with leaf sampling, leaf net CO2 uptake (A) and H2O efflux (E) were measured on three attached leaves, on each of three trees, at each of the three canopy levels in each study site. These leaves were from the same location within the canopy as the leaves sampled as described above. Measurements were made using a portable, open IRGA (LI 6400, LICOR Inc.) calibrated for CO2 and H2O against a known CO2 standard and using a dew point generator (LI 610, LICOR Inc.), respectively. All measurements were made under ambient PPFDs, air temperature (Tair) and H2O partial pressure on leaves held in a manner so as to approximate their normal orientation to the sun. Measurements were made on leaves that had no, or minimal, signs of damage. Discrimination against 13C during photosynthesis (Δ) results in 13C-depleted photosynthate and was predicted from gas exchange data using Farquhar et al. (1989) as both
- ((Eqn 1) )
- ((Eqn 2) )
[a, fractionation occurring during diffusion through air (4.4); b, net fractionation during carboxylations by Rubisco and PEPC (28)]. The partial pressures of CO2 in the substomatal cavity and the atmosphere are denoted by pi and pa, respectively. Equation 2 incorporates fractionation during dark respiration, where e is the fractionation occurring during dark respiration, here taken as δ13Cr–δ13Css, Rd is the rate of leaf dark respiration (µmol CO2 m−2 s−1) measured at night and k is the carboxylation efficiency, determined from the initial slope of response curves of A against pi for upper-canopy leaves only (k = 0.1; data not shown). Note that Δ is the leaf discrimination against 13C with respect to atmospheric δ13C (δ13Ca) and determination of δ13C of plant photosynthate, δ13Cp = δ13Ca − Δ.
Cumulative CO2 uptake through the sampling period (mol CO2 m−2) was determined by integrating beneath the curve of A over the sampling period, using a two-dimensional smoothing function in a graphical software package (sigma plot 8.02; SPSS Inc, Chicago, IL, USA).
For the purposes of statistical analysis, our replicates were three trees. Two-factor anova was used to test for: (i) an effect of time of day, position in the canopy and their interaction on δ13Cr, δ13Css, δ13Cst, δ13Cl and δ13Cc; and (ii) differences between δ13Cr and potential respiratory substrates δ13Css, δ13Cs and δ13Cl and their interaction with time of day. Whenever the interaction between the two factors was statistically significant, a post hoc Tukey test was performed to identify differences between individual means. Results are described as significant where P < 0.05. Statistical analysis was performed using a statistical software package (Systat 7.0, SPSS Inc, Chicago, IL, USA).
- Top of page
- Materials and Methods
We provide the first experimental evidence for a large diel cycle in δ13Cr under field conditions throughout the profile of two forest canopies. This cycle appeared to be the result of a progressive 13C-enrichment of δ13Cr of leaf ‘dark respiration’ during the photoperiod (of up to 7 for individual leaves), and a progressive decrease during the night. For leaves from the top and middle of both canopies, a robust linear relationship between δ13Cr and cumulative CO2 assimilation was found, providing a potential basis for predicting δ13Cr values from measurements or estimates of photosynthesis. These results have important implications for using δ13C to understand carbon flows through forested ecosystems. First, these results provide insights to metabolic changes that occur over the daily cycle in the field (see later in this section). Second, these results should influence sampling strategies of canopy air for estimation of ecosystem-scale δ13Cr signals by the Keeling-plot approach, which relies on constant δ13Cr (Pataki et al., 2003). Finally, variations in land ecosystem δ13Cr values of the magnitude observed here could cause annual net atmosphere-biosphere flux estimates from inverse models to vary significantly (Randerson et al., 2002; Scholze et al., 2003).
Although this study shows variation in δ13Cr and a potential means to estimate it, the data are not sufficient to explain the underlying mechanism for the variation. However, recent studies provide three possible mechanisms. Variation in δ13Cr could be driven by changes in: (1) primary photosynthetic discrimination and resulting changes in δ13C of respiratory substrates; (2) shifts among respiratory substrates, which are known to have distinct δ13C values; or (3) a shift in the ratio of CO2 produced during initial decarboxylation of pyruvate by PDH, to the oxidation of acetyl-CoA in the Krebs cycle. Only the third option can explain the magnitude of the diel cycle in δ13Cr reported here.
The first possibility was tested by calculating Δ from gas exchange data (Farquhar et al., 1989). We found that, for the upper-canopy Q. ilex leaves and the entire Q. cerris canopy, the enrichment pattern differed completely from those of δ13Cr (Table 1). Only in the middle and bottom of the Q. ilex canopy was a decrease in Δ (indicating 13C-enrichment in photosynthate) observed during the photoperiod. Even in these cases, the change in Δ was insufficient to explain the variation in δ13Cr, and for such change to be fully expressed in δ13Cr, we have to assume a direct link between photosynthesis and respiration via a rapid-turnover carbohydrate pool without interactions with the larger carbohydrate pools (e.g. soluble sugars, starch, cellulose and lipids, which did not show a daily pattern at all). Although recent studies indicated the existence of rapid turnover pools, its link to δ13Cr was partial (Affek & Yakir, 2002; Nogués et al., 2004) or required several days (Lotscher et al., 2004; Barbour et al., 2005). It was notable that the maximum ‘apparent’ fractionation during dark respiration observed in this study (c. 9, δ13Cr – δ13Css) had a small effect on Δ of 0.25 (Table 1). This effect was insignificant within the specific context of this study in which we attempted, but failed, to reconcile the variation in δ13Cr with Δ. Of course, the study of fractionation during dark respiration is, in itself, an active area of study, and fractionation of this magnitude is significant, with important implications for high precision attempts to predict Δ.
The second possibility, that simple switching between respiratory substrates was driving the variation in δ13Cr, could also be ruled out as a primary explanation because, for the upper-canopy Q. ilex leaves, the difference in δ13C between substrates was only up to 5. This is not sufficient to explain the observed daily change in δ13Cr up to 7 (Fig. 2a), even in the unlikely event of a complete shift from the most enriched to the most depleted substrate. Although there is no reason to assume different mechanisms in different leaves, it is worth noting that, for the other leaves, the full daily range in δ13Cr was smaller than the differences in the δ13C of the substrates. In theory, it follows that a shift from the oxidation of pure lipid predawn to soluble sugars or starch during the day could have produced the cycle observed (Tcherkez et al., 2003). However, this would still require c. 5 fractionation to reconcile the absolute substrate 13C with δ13Cr. The results are consistent with conclusions of recent studies that describe how the δ13C of specific carbohydrate extractions cannot be used as reliable surrogates for δ13Cr (Xu et al., 2004).
The third possibility was recently summarized in a quantitative manner by Tcherkez et al. (2003), who showed that shifts in the proportion of respired CO2 derived from the decarboxylation of pyruvate by PDH (relatively 13C-enriched CO2) to that derived from Krebs cycle decarboxylation reactions (relatively 13C-depleted CO2) could yield a range of up to 10 in δ13Cr. Such a range is consistent with this study and it is likely that here, too, changes in the rate of PDH activity relative to Krebs cycle activity were an important factor in the observed dynamic changes in δ13Cr. Certainly, during the dry Mediterranean summer months when growth has ceased there is a low metabolic demand for respiratory products of the Krebs cycle (Rambal et al., 2004). We speculate that under such conditions the synthesis of secondary compounds derived from acetyl-CoA depleted in 13C, particularly lipids, may be favoured relative to the oxidation of these compounds. This reasoning is consistent with the fact that the greatest 13C-enrichment of δ13Cr was observed at the top of the canopy, where the highest carbon accumulation took place against a background respiration rate (measured at night) that was no different between canopy layers. Further support is provided by the observation of a linear relationship between cumulative CO2 uptake and the daylight 13C-enrichment of δ13Cr above predawn values for sun leaves from two different canopies (Fig. 3).
Irrespective of the basis of the novel linear relationship between cumulative carbon uptake by the leaves and δ13Cr, these relationships provide empirical means to estimate δ13Cr from assessments of leaf carbon gain. It is also possible that these relationships hold at the canopy scale. Clearly, this approach, its basis and the scaling of the phenomenon should be further studied, because the empirical relationships observed do not reveal the underlying mechanism. In particular, these relationships should be tested for robustness, given the different relationship, albeit still linear, for middle-canopy leaves in the Q. cerris canopy and no correlation for leaves from the bottom of both the Q. ilex and Q. cerris canopies in this study. We suggest that limitations of intermittent gas-exchange measurements to capture fully sun-spot photosynthesis under the light environment within the lower canopy likely contributed to a large underestimation of leaf CO2 uptake over the photoperiod for lower-canopy leaves. It is also possible that in the Q. cerris canopy, where light penetration to the middle canopy was lower than in the Q. ilex canopy, we also underestimated leaf CO2 uptake.
In conclusion, there is a short-term diel cycle in 13Cr throughout forest canopies, which is reproducible in different oak species and at different layers in the canopy. The dynamics of this cycle cannot be predicted from analysis of key respiratory substrates but are consistent with recently proposed processes shown to influence the 13C-enrichment of δ13Cr in laboratory studies and may be predicted from simple assessments of leaf carbon accumulation.