Effects of intraleaf variations in carbonic anhydrase activity and gas exchange on leaf C18OO isoflux in Zea mays

Authors

  • Hagit P. Affek,

    1. Department of Environmental Sciences and Energy Research, Weizmann Institute of Science, Rehovot 76100, Israel. Current address: Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA.
    Search for more papers by this author
  • Maria J. Krisch,

    Corresponding author
    1. Current address: James Franck Institute and Department of Chemistry, The University of Chicago, Chicago, IL 60637, USA
      Author for correspondence: Dan Yakir Tel: +972 8 9342549 Fax: +972 8 9344124 Email: dan.yakir@weizmann.ac.il
    Search for more papers by this author
  • Dan Yakir

    1. Department of Environmental Sciences and Energy Research, Weizmann Institute of Science, Rehovot 76100, Israel. Current address: Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA.
    Search for more papers by this author

Author for correspondence: Dan Yakir Tel: +972 8 9342549 Fax: +972 8 9344124 Email: dan.yakir@weizmann.ac.il

Summary

  • • Variation in the C18OO content of atmospheric CO218Oa) can be used to distinguish photosynthesis from soil respiration, which is based on carbonic anhydrase (CA)-catalyzed 18O exchange between CO2 and 18O-enriched leaf water (δ18Ow).
  • • Here we tested the hypothesis that mean leaf δ18Ow and assimilation rates can be used to estimate whole-leaf C18OO flux (isoflux), ignoring intraleaf variations in CA activity and gas exchange parameters.
  • • We observed variations in CA activity along the leaf (> 30% decline from the leaf center toward the leaf ends), which were only partially correlated to those in δ18Ow (7 to 21‰), δ18O and δ13C of leaf organic matter (25 to 30‰ and −12.8 to −13.2‰, respectively), and substomatal CO2 concentrations (intercellular CO2 concentrations, ci, at the leaf center were ∼40% of those at the leaf tip).
  • • The combined effect of these variations produced a leaf-integrated isoflux that was different from that predicted based on bulk leaf values. However, because of canceling effects among the influencing parameters, isoflux overestimations were only ∼10%. Conversely, use of measured parameters from a leaf segment could produce large errors in predicting leaf-integrated C18OO fluxes.

Introduction

Natural variations in the 18O content (δ18O) of CO2 provide a useful tracer for photosynthetic activity as a consequence of a sequence of events: first, δ18O of chloroplast water is high because of evaporative effects; secondly, in the chloroplasts, exchange of oxygen between CO2 and H2O is catalyzed by carbonic anhydrase (CA); thirdly, a large fraction, proportional to the assimilation flux, of the 18O-labeled CO2 diffuses from the chloroplast back to the atmosphere (see Yakir, 1998 for a review). At the leaf scale, this ‘retroflux’ of 18O-enriched CO2 from the leaf back to the atmosphere is observed as discrimination against C18OO associated with leaf assimilation, 18Δ (Farquhar & Lloyd, 1993). The effect of 18Δ on atmospheric CO2 above plant canopies was used to partition net CO2 fluxes into ecosystem photosynthesis and respiration, taking advantage of the enrichment in evaporation of leaf water as compared with the less enriched soil water (Yakir & Wang, 1996; Bowling et al., 2003). This effect is also observed at the global scale as latitudinal and seasonal changes in the δ18O of atmospheric CO2, reflecting global-scale plant productivity (Francey & Tans, 1987; Farquhar et al., 1993; Ciais et al., 1997). The quantitative use of the 18O-CO2 signal, however, still critically depends on better understanding of processes influencing 18Δ (Gillon & Yakir, 2000a,b, 2001). This must include considerations of the large heterogeneity in the isotopic composition of leaf water, which has been repeatedly observed (Yakir et al., 1989; Luo & Sternberg, 1992; Luo & Sternberg, 1992; Wang & Yakir, 1995; Helliker & Ehleringer, 2000; Gan et al., 2002, 2003, with Luo & Sternberg (1992) referring to variations in δD), and possible heterogeneity in other factors that can influence the δ18O of CO2, including CA activity as well as leaf internal CO2 concentration, which has not yet been considered in this context.

The primary control on the δ18O of CO2 is the δ18O of the liquid water with which it was last in contact. CO2 equilibrates isotopically with water according to the following reaction:

image( Eqn 1)

(l, liquid; g, gas; aq, aqueous) which involves a temperature-dependent equilibrium fractionation between the oxygen in the CO2 and in water (Brenninkmeijer et al., 1983). In the presence of CA, which is ubiquitous in leaves, equilibrium can be reached nearly instantaneously, with a turnover rate of up to 106 s−1 (Silverman, 1982), and typical rates of 100–1400 µmol CO2 m−2 s−1 on a leaf area basis (Gillon & Yakir, 2001). The quantity of water usually involved in the CO2–water interaction is many orders of magnitude greater than the quantity of CO2 present, so that isotopically equilibrated CO2 takes on the oxygen isotopic ratio of the water in which it is dissolved, plus the temperature-dependent equilibrium fractionation, regardless of its initial δ18O value. However, full isotopic equilibrium is not always attained, depending on CA activity and internal CO2 concentrations (Gillon & Yakir, 2000a,b). Furthermore, leaf water is not well mixed and, as noted above, large heterogeneity in the isotopic composition of bulk leaf water (δ18Ow) has been repeatedly demonstrated. Farquhar & Gan (2003) recently provided a mathematical basis for describing the progressive enrichment in leaves, based on the string of lakes approach of Gat & Bowser (1991) and an internal Péclet effect (Farquhar & Lloyd, 1993; Barbour et al., 2004).

The overall discrimination against 18O during leaf CO2 assimilation, 18Δ, can be described as in Farquhar & Lloyd (1993), modified in Gillon & Yakir (2001):

image( Eqn 2)

where Ra and Rl, the oxygen isotope ratios of CO2 in the air and CO2 in equilibrium with water at the site of exchange; ā, the weighted average fractionation during diffusion of CO2 from the atmosphere to the chloroplast (8.8‰ in stagnant air and 0.8‰ in solution, with the weighted average taken as 7.4‰; Gillon & Yakir, 2001); ξ = ccs/(ca − ccs), with ca and ccs the CO2 concentrations in the atmosphere and at the site of oxygen exchange between CO2 and water in the leaves (assumed to be near the ‘chloroplast surface’), respectively (ξ represents the retro-diffusion flux back to the atmosphere after 18O exchange with leaf water); δa and δl, the δ18O values of CO2 in the atmosphere and in equilibrium with leaf water (strictly, the estimated δ18O of water in the chloroplast), respectively; δ = (Rsample/Rstandard − 1)103‰ (the standards are V-PDB, Vienna peedee belemnite, for CO2 and V-SMOW, Vienna standard mean ocean water, for water samples); θeq, the extent of isotopic equilibrium between CO2 and water, where θeq = 1 at full equilibrium. (Note that, unlike the primary effect on 13C, the secondary effect on 18O during carboxylation by the photosynthetic enzyme RUBISCO is expected to cause negligible discrimination.)

To assess the effects of leaf discrimination on CO2 in the air above the leaf, both at physiological and at larger scales, an ‘isoflux’ term is often used. The 18O assimilation isoflux in this case refers to the net assimilation flux of C18OO:

image(Eqn 3)

where A, assimilation flux; RA, the molar ratio 18O/16O in the assimilated CO2. 18ISOFLUXA has units of µmol m−2 s−1 ‰ and use of the δ notation (a measured quantity) makes it conceptually equivalent, but not identical, to the actual C18OO flux (cf. Bowling et al., 2003; note that we use the convention in atmospheric studies of negative fluxes out of the atmosphere, resulting in positive isofluxes).

The objective of this study was to examine the effect of the large variations in δ18Ow observed in leaves on the integrated leaf 18ISOFLUXA, and to test the hypothesis that mean leaf δ18Ow can be used to estimate whole-leaf 18ISOFLUXA. This hypothesis must be based on the assumption that other influencing parameters, primarily CA activity and ccs, are either constant or would vary in concert with δ18Ow and would therefore not influence estimates of 18ISOFLUXA. Accepting or rejecting this hypothesis can have important consequences for the use of 18O in CO2 in physiological, ecological, and large-scale studies using the 18O approach, and to interpretations of variations in the δ18O of atmospheric CO2.

Materials and Methods

Gas exchange

All parameters were measured in 10-cm segments (segments were numbered so that segment 1 was the base of the leaf) along at least three corn (Zea mays L.) leaves in field-grown plants in the southern coastal plains of Israel at around midday (11:00–13:00 h). The incident light intensity (I) was measured with a LiCor PAR sensor (LiCor, Lincoln, NE, USA) along the leaves before and after each gas exchange measurement and averaged. Gas exchange parameters (net assimilation rate, A, stomatal conductance, gs, and intercellular CO2 concentration ci) were measured in attached leaves using a portable gas exchange system (Li-6400; LiCor) with light supplied by the instrument LED array and adjusted to match the sunlight intensity measured just before measurement (this provided near-ambient but more stable conditions during measurements). In some cases, light response curves were obtained in attached leaves under field conditions by stepwise changes of the light intensity of the instrument (first increasing I from ambient values to saturation, and then decreasing back to low values). The observed gas exchange parameters were also used to estimate the extent of 18O exchange between CO2 and water (θeq; see the Results section).

Isotopic analysis

Leaf water was extracted by vacuum distillation at 80°C. δ18O values were determined by equilibration of 0.5 ml of water with CO2 for 24 h at 29°C followed by cryogenic purification of a CO2 aliquot. δ18O values of the CO2 were measured by dual inlet isotope ratio mass spectrometer (IRMS; MAT250; Finigan, Bremen, Germany). Values were calibrated on the V-SMOW scale by simultaneously measuring an internal water standard (having a δ18O value of −4.5‰ periodically calibrated to the international V-SMOW standard obtained from the International Atomic Energy Agency, Vienna, Austria).

Ground whole dry leaf segments were analyzed for δ13C by a conventional online combustion elemental analyzer (EA1109 CHN-O; Carlo Erba Instuments, Milan, Italy) connected to IRMS (Optima; Micromass, Manchester, UK). Leaf organic δ18O measurements were obtained by pyrolysis on graphite, using the same elemental analyzer, followed by IRMS measurement of the CO produced (Saurer et al., 1998).

Carbonic anhydrase activity, θeq and ccs

Carbonic anhydrase (CA) was extracted, within 2 h of collecting leaves, from leaf discs (1.7 cm2) taken at 10-cm intervals along the leaves. Extraction and activity measurement assay were performed using the method described by Gillon & Yakir (2000a,b). CA was extracted by grinding leaf discs in extraction buffer (50 mm HEPES-NaOH, pH 8.3, 0.5 mm EDTA, 10 mm dithiothreitol, 10% glycerol, and 1% triton X-100) at 4°C. The in vitro CA activity assay was performed at 2°C by adding saturated aqueous CO2 into an assay buffer (20 mm Na-barbitol, pH 8.3) containing the enzyme extract, and the rate of pH decrease yielded CAassay. The in vivo CA activity at leaf temperature and saturating light (CAleaf) was then estimated as CAassay converted to leaf conditions: CAleaf = F CAassay [(17.5 + Km)/17.5] [ccs/(ccs + Km)], where F is a temperature correction factor, F = Q10(Tleaf–Tassay)/10, assuming Q10 = 2 (Burnell & Hatch, 1988), and Km is the CO2 concentration at half maximal activity, taken as 2.8 mm (Hatch & Burnell, 1990). The CO2 concentration at the chloroplast surface (ccs) was estimated from A = gw(ci − ccs), using light-saturated A and ci values and where gw, the internal wall conductance, is assumed to be 1 mmol m−2 s−1 (Gillon & Yakir, 2000b). The extent of isotopic equilibrium between CO2 and water (θeq) was determined as θeq = 1 − ekτ/3, where kτ =CAleaf/Fin and Fin = A[ccs/(ca − ccs) + 1] (Gillon & Yakir, 2000a).

Results

The light intensity (I) incident along the leaf in its natural orientation under field conditions was variable. I values were maximal, ∼700 µmol m−2 s−1, around the mid-leaf sections and declined toward both the tip and the base to ∼400 µmol m−2 s−1 (Fig. 1a). Gas-exchange parameters, measured in 10-cm segments along the leaf, were generally consistent with leaf orientation toward the sun and for most parameters reflected the two main parts of the leaf: the upward section from the base to approximately mid-leaf, and the downward section from mid-leaf to the tip (Fig. 1).

Figure 1.

(a) Light intensity (I), (b) net assimilation rate (A), (c) intercellular CO2 concentration (ci) and (d) stomatal conductance (gs), measured at 10-cm intervals along corn (Zea mays) leaves. Leaf segments (10 cm each) are labeled from the base (lowest number) to the tip of the leaf. A, gs, ci were measured under artificial light equivalent to the natural light intensity incident on the leaf, as measured shortly before the gas exchange measurements. To account for leaf-to-leaf variations in absolute values, gas exchange parameters in each leaf were normalized to the maximal value of each parameter before averaging. The horizontal line indicates the normalized mean value along the leaf, with the actual bulk leaf value given at the top of each panel. The light response curve in (e) was measured in the middle of one leaf.

A light response curve measured at mid-leaf (Fig. 1e) indicated that A was light saturated at light intensities of c. 1100 µmol m−2 s−1. Hence, A increased, together with I, from 15 ± 0.8 µmol m−2 s−1 (mean ± standard error; n = 3) at the leaf base to 20 ± 3 µmol m−2 s−1 at mid-leaf (Fig. 1b). A was relatively constant, however, from mid-leaf to the tip (19 ± 0.8 µmol m−2 s−1; n = 12), in spite of the decrease in I. Stomatal conductance (gs) increased along the leaf from 0.11 ± 0.02 mol m−2 s−1 (n = 3) at the leaf base to 0.21 ± 0.01 mol m−2 s−1 (n = 3) at the tip, with a small decrease at mid-leaf (Fig. 1d). Although modest, variations in gs were consistent with changes in A from the leaf base to mid-leaf, resulting in relatively constant values of ci (81 ± 8 µl l−1; n = 12). This was followed by an increase in ci values from mid-leaf to 184 ± 9 µl l−1 (n = 3) at the tip (Fig. 1c), which was correlated with the increase in gs but relatively constant A. As the ambient CO2 concentration (ca, recorded before each gas exchange measurement) did not vary much, the ratio ci/ca varied in a similar manner to the variations in ci from the leaf base to mid-leaf (0.23 ± 0.04; n = 12) and to the tip (0.53 ± 0.03; n = 3).

The ratio ci/ca also influences the δ13C of leaf organic matter (δ13Corg) which on average decreased from −12.8 ± 0.1‰ (n = 4) at the base to −13.2 ± 0.1‰ (n = 4) at the tip of the leaves (Fig. 2b). However, in the first half of the leaf, from base to mid-leaf, this 13C depletion was mainly a consequence of the first segment being relatively enriched (−12.8 ± 0.1‰; n = 4), whereas segments 2–4 did not greatly change (−12.9 ± 0.06‰; n = 12). Overall, therefore, high ci values were correlated with more depleted δ13Corg in leaf organic matter. Note, however, that the correlation between ci and δ13Corg may be fortuitous as these parameters represent different time scales and likely also spatial scales, as most organic matter is produced at the leaf base during leaf development.

Figure 2.

Isotopic composition along corn (Zea mays) leaves: (a) the oxygen isotope composition of leaf water (δ18Ow); (b) the oxygen (δ18Oorg; closed circles) and carbon (δ13Corg; open circles) isotope composition of leaf organic matter. The equation denotes the curve fit for δ18O only. Values are mean ± standard error for four leaves. Leaf segments (10 cm each) are labeled from the base (lowest number) to the tip of the leaf.

The oxygen isotopic composition of leaf water (δ18Ow) showed pronounced enrichment along the leaf, from 7.2 ± 0.7‰ (n = 4) at the base to 21.2 ± 0.6‰ (n = 4) at the tip (Fig. 2a). Stem water in these plants was −1.9 ± 0.1‰. A similar trend of enrichment along the leaf, although less pronounced, was observed in the oxygen isotopic composition of leaf organic matter (δ18Oorg; Fig. 2b). δ18Oorg values increased from 25.7 ± 0.7‰ (n = 3) at the leaf base to 30.1 ± 0.4‰ at the tip. This resulted in a good correlation between δ18Ow and δ18Oorg, with a best-fit line of δ18Oorg = 23.96 + 0.27δ18Ow (R2 = 0.92), indicating an enrichment trend along the leaf that was only c. 30% of that in δ18Ow. Note that, as for 13C discussed above, the isotopic signals in water and organic matter represent different temporal and possibly also spatial scales. During cellulose synthesis in the leaf base, extensive exchange with stem water can occur (Farquhar et al., 1998; Yakir, 1998; Roden et al., 2000).

The activity of the enzyme carbonic anhydrase (CAassay) measured in 10 leaves showed a similar pattern to that in incident light intensity along the leaf (although large leaf-to-leaf variations were observed). On average, CAassay increased from 136 ± 17 µmol m−2 s−1 (n = 10) at the leaf base to 175 ± 22 µmol m−2 s−1 at mid-leaf and decreased to 122 ± 38 µmol m−2 s−1 at the tip (Fig. 3a). The mean bulk leaf CAassay was 160 ± 18 µmol m−2 s−1.

Figure 3.

Rate of activity of carbonic anhydrase (CA) along corn (Zea mays) leaves: (a) measured activity of leaf extracts in assay conditions (CAassay) at 10-cm intervals along the leaf. To account for leaf-to-leaf variations in absolute values, CAassay values were normalized to maximal activity in each leaf. The horizontal line indicates the normalized mean value along the leaf, with the actual bulk leaf value indicated at the bottom of the panel. Values are mean ± standard error for 10 leaves, with leaf segments (10 cm each) labeled from the base (lowest number) to the tip of the leaf. (b) CA activity under in vivo conditions (CAleaf) estimated as in Gillon & Yakir (2000b) using leaf segment CAassay values (mean of 10 leaves), measured leaf temperature and intercellular CO2 concentration (ci) at saturating light intensity. (c) Extent of isotopic equilibrium between CO2 and water (θeq) calculated using CAleaf and the light-saturated assimilation rate (full equilibrium at θeq = 1).

A similar pattern was observed when the potential in vivo activity of carbonic anhydrase (CAleaf, the measured assay activity converted to leaf conditions) was estimated using the gas exchange parameters (A and ci) at saturating light intensity (Fig. 1e). CAleaf increased from 173 µmol m−2 s−1 at the leaf base to 228 µmol m−2 s−1 at mid-leaf and decreased to 158 µmol m−2 s−1 at the tip of the leaf (Fig. 3b). CAleaf was used to calculate the extent of isotopic equilibrium between CO2 and water (θeq) using measured light-saturated assimilation rates as a constant basis. Complete isotopic equilibrium (θeq = 1) was never reached in the corn leaves (Fig. 3c) and θeq generally varied according to variation in CA activity, but with a smaller range, from 0.86 at the leaf base to 0.92 at mid-leaf and 0.83 at the tip.

The combined effect of variations along the leaf in δ18Ow, gas exchange parameters and θeq on the leaf 18ISOFLUXA was estimated using calculated values of leaf discrimination against 18O, 18Δ, according to Eqns 2 and 3. δa was taken as −0.2‰ (the average value measured at the month in which leaves were sampled in our Negev station of the National Oceanic and Atmospheric Administration's Climate Monitoring and Diagnostic Laboratory, NOAA-CMDL, network; http://www.cmdl.noaa.gov/ccgg/index.htm). 18Δ increased slightly from 8.9‰ at the leaf base to 9.4‰ at mid-leaf and then sharply to 24.2‰ at the tip. In the lower half of the leaf, the small change in 18Δ likely reflected the slight decrease in ci and calculated ccs values, balanced by an increase in δ18Ow. The sharp increase in 18Δ in the upper half of the leaf likely reflected the combined increase in ccs and δ18Ow. Average 18Δ for the leaf, calculated using mean values for the entire leaf in Eqns 2 and 3, was 11.6‰. As a result of the increase in A from the leaf base to mid-leaf (Fig. 1b), 18ISOFLUXA increased from 137 to 194 µmol m−2 s−1 ‰. A sharper increase to 480 µmol m−2 s−1‰ was observed from mid-leaf to the tip, reflecting the increase in 18Δ, in spite of the constant A (Fig. 4). 18ISOFLUXA calculated with mean parameters for the entire leaf was 222 µmol m−2 s−1 ‰.

Figure 4.

The 18O assimilation flux calculated as 18ISOFLUXA= –Aa −18Δ) for (a) C4 plants and (b) simulated C3 plants assuming ca − ci drawdown in a C3 grass to be 0.53 of that measured in corn (Zea mays), where ci is the intercellular CO2 concentration and ca is the CO2 concentration in the atmosphere. δa denotes the δ18O value of atmospheric CO2. 18Δ, the overall discrimination against 18O during leaf CO2 assimilation, was estimated using the measured extent of isotopic equilibrium (θeq) along corn leaves (closed circles in panel a), or using a range of possible θeq values typical of C4 grasses (θeq = 0.4, triangles), C3 grasses (θeq = 0.8, diamonds) or dicots (θeq = 1, squares). The measured assimilation rate (A) along the leaf (mean of three leaves) was used in 18ISOFLUXA estimates. Closed and open symbols denote C4 and C3 values, respectively. The horizontal lines indicate whole-leaf mean 18ISOFLUXA values. Leaf segments (10 cm each) are labeled from the base (lowest number) to the tip of the leaf.

Discussion

The primary objective of this study was to evaluate the effect of the large variations in δ18Ow observed in leaves (Yakir et al., 1989; Luo & Sternberg, 1992; Wang & Yakir, 1995; Helliker & Ehleringer, 2000; Gan et al., 2002, 2003) on the integrated leaf 18ISOFLUXA (Eqn 3). A reasonable hypothesis is that 18ISOFLUXA varies along leaves in concert with δ18Ow, and consequently the mean δ18Ow value of bulk leaf water can serve as a good predictor of the whole-leaf 18ISOFLUXA. Rejection of this hypothesis complicates estimation of 18ISOFLUXA, and can result in a very different leaf 18ISOFLUXA from that estimated based on the relatively easy to predicted mean leaf δ18Ow value. For example, a case in which δ18Ow increases along the leaf but CAleaf and ccs, the primary controls of leaf 18ISOFLUXA (cf. Eqns 2,3), are constant along the leaf would produce an 18ISOFLUXA that simply reflected the bulk, or mean, δ18Ow value; however, this could be very different in a case were the contribution to the 18ISOFLUXA of the very high δ18Ow observed in approaching the leaf tip was diminished as a result, say, of low ccs and/or low CA activity in this part of the leaf.

The results show the expected large progressive enrichment in δ18Ow along corn leaves, but also variable patterns of other parameters influencing the transfer of 18O from leaf water to CO2. Some parameters, such as I, showed bimodal patterns along the leaf, consistent with the light regime under natural field conditions. Other parameters, such as CAleaf, A and ci (which was used to estimate ccs), showed secondary response patterns producing a nonlinear progression along the leaves. As a result, the leaf 18ISOFLUXA did not have a similar pattern to δ18Ow along the leaf and reflected instead the balance of the various patterns. We therefore had to reject our hypothesis. However, as discussed below (see The leaf 18ISOFLUXA), contrasting effects, such as high CAleaf associated with low ccs, reduced the effects of the complex gradients along the leaf on the total leaf isoflux, thus reducing the potential error introduced by the conventional use of leaf bulk δ18Ow to predict 18ISOFLUXA.

The water component

The observed large progressive enrichment in δ18Ow (measured as the difference between the δ18O of segment water and that of the source, stem, water was linearly related to the distance from the base of the leaf and was similar to that observed previously (Helliker & Ehleringer, 2000, 2002a). This progressive enrichment process was described mathematically for a chain of lakes by Gat & Bowser (1991) and was recently adapted to leaves by making it continuous and incorporating a Péclet effect, to account for the competing effects of advection and diffusion to and from the site of evaporation (Farquhar & Gan, 2003). While these models describe the increasing enrichment along leaves, the integrated δ18O value of the entire system (i.e. the mean bulk leaf water) predicted by the models approaches the value predicted by the original evaporation model of Craig & Gordon (1965; cf. Flanagan, 1993), which was developed for one well-mixed water pool. This simpler bulk leaf model cannot, however, be used to accurately estimate the leaf 18ISOFLUXA if intraleaf variations in other relevant parameters are not constant (or vary in concert with δ18Ow along the leaf).

It is well established that the δ18O of leaf organic matter, δ18Oorg, is linked to δ18Ow and can provide a time-integrated record of δ18Ow (e.g. Yakir, 1992). Indeed, there was a high linear correlation between δ18Oorg and δ18Ow along the corn leaves, as observed in other plants (e.g. Helliker & Ehleringer, 2002a). However, the slope of ∼0.3 indicated a more moderate enrichment along the leaf for δ18Oorg than for δ18Ow. This is expected because of the exchange of organically bound oxygen with water pools other than the segment water represented by our δ18Ow values (Yakir, 1992; Saurer et al., 1997; Farquhar et al., 1998). Calculating the exchange parameters recently proposed by Barbour & Farquhar (2000), we obtained a mean PexPx value (representing, respectively, the proportion of exchangeable oxygen atoms during synthesis of cellulose, and the proportion of xylem water present in the cells where cellulose, the dominant organic species in the leaf material, is synthesized) of 0.73 with an ɛ0 value (representing the 18O discrimination factor between carbonyl oxygen and water) of 25.4‰. This PexPx estimate is much higher than the values of 0.25–0.38 observed in short grasses (Helliker & Ehleringer, 2002a; Barbour et al., 2004). The structural organic material, which comprises most of the material in our samples, is formed primarily in the leaf-base meristem. This suggests that, while δ18Oorg values are initially determined in the leaf region with high rates of assimilation, they are consequently influenced by exchange with the water at the leaf-base site during leaf elongation (Helliker & Ehleringer, 2002a,b; Barbour et al., 2004). Therefore, a leaf-segment δ18Oorg is likely to be influenced by the temporal progression in the δ18O of water during leaf development (including variations in longitudinal Péclet effect in the leaf as it increases in length). Such complicating factors must be considered before δ18Oorg can be used as a direct record of δ18Ow along leaves.

From water to CO2

The transfer of the 18O signal from leaf water to CO2 depends on the exchange of oxygen between the leaf water and CO2 (Eqn 1). Because of the short residence time of CO2 inside the leaves (normally less than 1 s) and the relatively slow noncatalyzed CO2 hydration rate, this exchange critically depends on the concentration of CO2 at the site of CO2–H2O exchange (ccs) and on the rate of CA activity. The value of ccs should be intermediate between that at the substomatal cavities, ci, and that at the site of the photosynthetic enzyme RUBISCO (Gillon & Yakir, 2000a; see Materials and Methods). Leaf-scale physiology therefore strongly influences the evolution of the leaf C18OO flux.

Clearly, the physiological parameters measured here varied along the corn leaf independent of the linear enrichment in δ18Ow (Figs 1, 2). The dominant factors influencing intraleaf physiology were probably incident light (Fig. 1) and developmental stage (young at the base and senescing at the tip; not examined here). In response to variations in these dominant parameters, gs, ccs and CAleaf seem to have varied to produce relatively stable A and θeq across most of the leaf (Figs 1, 3). Most prominently, the pattern in incident light, which was maximal at mid-leaf, was largely paralleled by the patterns in ccs and CAleaf and all three parameters changed by 30–50% along the leaves. In contrast, both θeq and A varied by less than 10% along the leaf (excluding one data point for A). Optimization of A and CA activity (which determines θeq) probably helps to maximize leaf productivity, while it is θeq and ccs that dominate 18Δ and the leaf 18ISOFLUXA (Eqns 2,3).

The activity of carbonic anhydrase in the present study was significantly higher than the mean value estimated previously for C4 plants (Gillon & Yakir, 2000b, 2001), although, in these studies too, cultivated corn had the highest CA activity among the C4 plants. Therefore, while C4 plants were reported to have a mean θeq value of ∼0.4, here we obtained for cultivated corn plants θeq values of ∼0.9 (Fig. 3c). In estimating leaf isoflux we therefore also simulated the effects of the observed intraleaf variations when mean θeq is 0.4 (see The leaf 18ISOFLUXA).

As for 18O, the δ13C of leaf organic matter, δ13Corg, can potentially provide a long-term integrated record of ci (Farquhar et al., 1982, 1998) and therefore of ccs, and possibly of their variations along the leaf. Indeed, the organic 13C record along the leaf showed a similar pattern to that of ci (excluding one data point at the base of the leaf; cf. Fig. 1c vs Fig. 2b). The general trend of 13C depletion along the leaf in corn and sugar cane (Saccharum spp.) was observed previously and could also be linked to ci (Sasakawa et al., 1989; Meinzer & Saliendra, 1997). However, different patterns of incident light along the leaf led to different trends in ci, with decreasing values along the sugar cane leaves (Meinzer & Saliendra, 1997) and increasing values in the corn leaves. These contrasting patterns are consistent with the differences in estimated bundle sheath leakiness. Leakiness was estimated at ∼0.40 in corn, while a value of 0.32 was calculated in sugar cane (Meinzer & Saliendra, 1997). These differences in leakiness are significant; the theoretical model for C4 discrimination (Farquhar, 1983) predicts that, at a leakiness of ∼0.35, the pattern of 13C vs ci/ca is inverted, with an inverse correlation below this value and a direct correlation above this value. However, as noted above for δ18Oorg, leaf organic matter (mostly cellulose) is formed in the leaf-base meristem and not at the particular segment sampled. As discussed for δ18Oorg, δ13Corg too is likely to be influenced by the temporal progression in ci during leaf development. For example, the tip of the leaves likely contained cellulose formed in the base meristem of the young leaves when vapor pressure deficit and temperature were lower than during the sampling time at peak season, consistent with the more depleted δ13Corg observed. Note that, while such temporal effects would also apply to δ18Oorg, they would be overshadowed by the oxygen exchange with tissue water discussed above, which should not influence δ13Oorg. However, for both δ13Oorg and δ18Oorg, the complicating factors involved must be considered before these parameters can be used as a direct record of ci or δ18Ow along leaves.

The leaf 18ISOFLUXA

The effects of δ18Ow and leaf physiology were integrated by estimating 18Δ (Eqn 2) and 18ISOFLUXA (Eqn 3). Note that 18ISOFLUXA was dominated by 18Δ as A and δa did not vary much along the leaf. In the bottom half of the leaf, a relatively small change in 18Δ and 18ISOFLUXA was observed, reflecting the slight decrease in ci and the calculated ccs values, balanced by an increase in δ18Ow (Fig. 4a). In the top half of the leaf, the sharp increase in 18Δ and greater 18ISOFLUXA reflected the combined increase in ccs and δ18Ow (Fig. 4a).

The sensitivity of the patterns in 18ISOFLUXA to θeq values was examined by using θeq = 0.4, which is typical for C4 plants (Gillon & Yakir, 2000b). This resulted in only slight changes in the bottom part of the leaf but considerably greater 18ISOFLUXA toward the leaf tip (480 µmol m−2 s−1 ‰ using the observed θeq and 270 µmol m−2 s−1 ‰ with θeq = 0.4). Using θeq = 1, 18ISOFLUXA varied along the leaf, from 146 µmol m−2 s−1 ‰ at the leaf base to 201 µmol m−2 s−1 ‰ at mid-leaf to 563 µmol m−2 s−1 ‰ at the tip. This variable sensitivity to θeq values demonstrated the importance of ci (and, when available, ccs), which controls the gross, one-way retro-diffusion flux from the leaf back to the atmosphere. Therefore, high ci (and ccs) results in high 18O-labeled retro-diffusion flux and consequently high 18Δ. Thus, a combined effect of θeq and ci (or ccs) strongly impacts the leaf 18ISOFLUXA. Such an effect can be expected in C3 leaves when θeq is near 1 and ci values are generally much higher than in C4 leaves.

For such a comparison, we simulated C3 grass leaves (e.g. wheat, Triticumaestivum) by maintaining the same patterns observed along corn leaves, but assuming the stomatal drawdown in CO2 concentration (from ca to ci) in a C3 leaf to be 0.53 of observed values in corn, yielding a mean ci value of 219 µl l−1, compatible with a typical ratio of 2.1 for ci values between C4 and C3 leaves (Lloyd & Farquhar, 1994). This allowed a comparison, as a first approximation, of the expected 18ISOFLUXA along leaves between C3 plants and C4 monocots. As expected, with both θeq and ci values high in the ‘C3 leaf’, 18ISOFLUXA was much greater than in typical C4 leaves (Fig. 4b). High ci also makes the leaves more sensitive to variations in θeq values. For θeq = 1, which is typical for C3 dicots, or θeq = 0.8, which is typical for C3 grasses (Gillon & Yakir, 2001), 18ISOFLUXA at the tip of the leaf was 1178 and 940 µmol m−2 s−1 ‰, respectively (Fig. 4b).

Finally, we compared the weighted average leaf 18ISOFLUXA, obtained by following the variations along the leaves, with that obtained by using a single bulk leaf value for δ18Ow and a mean value for each of the physiological parameters (Table 1). The errors introduced in this case ranged between 7% (for C4 assuming θeq = 0.4) and 11% (for C4 assuming θeq = 1) and were c. 7–8% for simulated C3 grass leaves. The error would of course be much greater if an arbitrary segment of a leaf was used. In this case, the range would be between 35 and 56% underestimation when using segments from the leaf base (for C4, θeq = 0.4 and for C3, θeq = 1, respectively) and between 57 and 111% overestimation when using segments near the leaf tip (for C4, θeq = 0.4 or 1, respectively). Although the patterns of ci and CA activity along the leaf did not co-vary with δ18Ow, leading us to reject our initial hypothesis of a direct correlation between 18ISOFLUXA and δ18Ow, the results showed that the overestimation of 18ISOFLUXA caused by ignoring these variations and using instead the bulk leaf values is relatively small. Therefore, under current levels of uncertainty in modeling leaf and ecosystem isofluxes, it is probably still reasonable to use bulk leaf values to estimate the total leaf 18ISOFLUXA, but not values from any specific part of the leaf.

Table 1. 18ISOFLUXA estimates using different values for the extent of isotopic equilibrium between CO2 and water (full equilibrium at θeq = 1)
 C418ISOFLUXA (µmol m−2 s−1 ‰)C318ISOFLUXA (µmol m−2 s−1 ‰)
θeq = 1θeq = 0.8θeq = 0.4θeq = 1θeq = 0.8θeq = 0.4
  1. For C4, 18ISOFLUXA is calculated from measured gas exchange parameters and the δ18O of leaf water (δ18Ow) in corn (Zea mays). For C3, gas exchange parameters are simulated assuming ca − ci drawdown in a C3 grass to be 0.53 of that measured in corn. Values are given for the base of the leaf, the mid-leaf and the tip based on measurements in 10-cm leaf segments. ‘Average’ refers to the weighted average obtained from leaf isofluxes estimated for each leaf segment. ‘Bulk’ refers to the whole-leaf 18ISOFLUXA calculated using the leaf mean values for δ18Ow and gas exchange parameters. Errors were estimated as error = 100(1 − bulk/average) and refer to estimating leaf 18ISOFLUXA without the weighting of isotopic and physiological parameters involved in calculating the 18ISOFLUXA of leaf segments (see Eqn 3).

Base 146135112 257215130
Mid 201187159 450373221
Tip 5634642701178940480
Bulk 238212161 538440247
Average 267235172 588479266
Error (%)10.9 9.8 6.6  8.4 8.1 6.9

Acknowledgements

We are grateful to D. Hemming and J. S. Gillon for helpful comments and to E. Negreanu, R. Ben-Meir and T. Lin for technical support. We would also like to particularly thank A. Schwartz for his help in different aspects of this project. This research was supported by grants from BSF, GLOWA-JR and Minerva Foundation to DY.

Ancillary

Advertisement