Measurement of Gross Photosynthesis, Respiration in the Light, and Mesophyll Conductance Using H₂¹⁸O Labeling^,

H₂¹⁸O labeling and IRMS can be used to measure gross photosynthesis, mesophyll conductance, and respiratory uptake in the light in leaflets exposed to 21% and 2% O₂.

Abstract

A fundamental challenge in plant physiology is independently determining the rates of gross O₂ production by photosynthesis and O₂ consumption by respiration, photorespiration, and other processes. Previous studies on isolated chloroplasts or leaves have separately constrained net and gross O₂ production (NOP and GOP, respectively) by labeling ambient O₂ with ¹⁸O while leaf water was unlabeled. Here, we describe a method to accurately measure GOP and NOP of whole detached leaves in a cuvette as a routine gas-exchange measurement. The petiole is immersed in water enriched to a δ¹⁸O of ∼9,000‰, and leaf water is labeled through the transpiration stream. Photosynthesis transfers ¹⁸O from H₂O to O₂. GOP is calculated from the increase in δ¹⁸O of O₂ as air passes through the cuvette. NOP is determined from the increase in O₂/N₂. Both terms are measured by isotope ratio mass spectrometry. CO₂ assimilation and other standard gas-exchange parameters also were measured. Reproducible measurements are made on a single leaf for more than 15 h. We used this method to measure the light response curve of NOP and GOP in French bean (Phaseolus vulgaris) at 21% and 2% O₂. We then used these data to examine the O₂/CO₂ ratio of net photosynthesis, the light response curve of mesophyll conductance, and the apparent inhibition of respiration in the light (Kok effect) at both oxygen levels. The results are discussed in the context of evaluating the technique as a tool to study and understand leaf physiological traits.

Gross O₂ production (GOP; the rate of water splitting at PSII), mitochondrial respiration in the light (R_L), the oxygenation of Rubisco, and the photosynthetic carbon oxidation cycle (i.e. photorespiration; Hagemann et al., 2013) are fundamental rate properties of plants. When assessed as net O₂ fluxes, the rates of these processes are confounded. Since each of the component processes represents a distinct biochemical pathway with important cellular functions, quantifying the contribution of these processes to the net assimilatory flux is of particular interest. The utilization of O₂ as an electron acceptor around PSI via the Mehler reaction also can contribute to O₂ uptake (Badger et al., 2000). Precise measurements of O₂ fluxes from leaves would provide a better understanding of these important processes and shed light on diverse metabolic responses to environmental variation (Furbank et al., 1982; Badger et al., 2000; Hillier, 2008; André, 2011).

A variety of methods have been developed in leaves and phytoplankton to measure net O₂ production (NOP; Kok, 1948; Cornic et al., 1989; Laisk et al., 1992; Laisk and Oja, 1998). GOP has been measured directly using an ¹⁸O tracer and mass spectrometry (Mehler and Brown 1952; Hoch et al., 1963; Ozbun et al., 1964; Radmer and Ollinger, 1980; Ruuska et al., 2000). Two general approaches have been used to measure GOP: (1) measuring the increase of ¹⁶O₂ from unlabeled water in a chamber containing dioxygen only as ¹⁸O₂; and (2) measuring ¹⁸O¹⁶O of photosynthetic O₂, produced from splitting ¹⁸O-labeled water. All published GOP data for leaves use the first approach to determine GOP and, given NOP measurements, O₂ uptake (Berry et al., 1978; Gerbaud and André, 1979; Canvin et al., 1980; Peltier and Thibault, 1985; Biehler and Fock, 1995; Haupt-Herting and Fock, 2002; André, 2011). Typically, this method relies on leaf fractions or small leaves that are placed in a gas-tight gas-exchange chamber connected to a membrane inlet mass spectrometer (Berry et al., 1978; Biehler and Fock, 1995). The chamber air is then replaced with O₂ present as ¹⁸O₂, and the evolution of ¹⁶O₂ and ¹⁸O₂ is monitored. The decrease of ¹⁸O₂ concentration in the system constrains O₂ uptake, and the increase of ¹⁶O₂ gives GOP. GOP, determined in this way, is a reliable measure of the electron transport rate (J; André, 2013), which then can be used to estimate mesophyll conductance (g_m; Flexas et al., 2012) and chloroplastic [CO₂] (C_c; Renou et al., 1990; André, 2011). Nevertheless, among current methods used to measure g_m (for review, see Flexas et al., 2012), the ¹⁸O method was limited by a restriction associated with obtaining ¹⁸O₂ and the need for specific equipment (e.g. membrane inlet mass spectrometry). As a consequence, to date, measurements of NOP and GOP have received little attention.

The second approach, relying on ¹⁸O-labeled water, has been used to measure NOP and GOP in phytoplankton (Grande et al., 1989; Goldman et al., 2015). With this method, water in the cells is labeled with H₂¹⁸O. If implemented with a whole leaf in a flow-through gas-exchange system, the leaf would be labeled with H₂¹⁸O via the transpiration stream. In the cuvette, NOP raises the O₂ concentration, and GOP also raises the δ¹⁸O of O₂. The rates of NOP and GOP are measured from the difference in O₂/N₂ (Grande et al., 1989) and δ¹⁸O of O₂ in air entering and leaving the cuvette. O₂ uptake is the difference between these two rates (GOP – NOP). The measurement of the change in O₂ concentration is a standard practice for high-precision measurements: changes in O₂ concentration are routinely measured from changes in the ratio of O₂/N₂ rather than from the O₂ concentration or partial pressure, which also depend on temperature and humidity.

The advantage of this method is that gas-exchange properties are determined with a steady-state gas-exchange protocol rather than in a closed gas-exchange system. The challenge of this technique is that there are only small differences in the O₂/N₂ ratio and δ¹⁸O of O₂ between outgoing and incoming air. These small changes can be measured by high-precision isotope ratio mass spectrometry. For example, a rate of NOP resulting in a CO₂ drawdown or O₂ rise of 100 µL L⁻¹ would cause O₂/N₂ in air exiting the chamber to be increased by 0.47‰. If water molecules were labeled to +10,000‰, GOP of 100 μmol mol⁻¹ would cause δ¹⁸O of O₂ to increase by 5‰. For comparison, δO₂/N₂ can be measured by isotope ratio mass spectrometry to a precision of ±0.005‰ and δ¹⁸O of O₂ can be measured to a precision of ±0.03‰ or better (Bender et al., 1994). Thus, one can measure [O₂] and δ¹⁸O changes associated with NOP and GOP to a precision of about ±1 µL L⁻¹ out of 210,000 µL L⁻¹ (Bender et al., 1998). Our H₂¹⁸O labeling approach, as described below, allows us to simultaneously measure a full set of fundamental gas-exchange properties, including NOP and GOP. We also can measure light and CO₂ response curves, thus characterizing many important physiological properties of plants simultaneously. As will be shown, we can regulate flow rate, temperature, humidity, and other properties of air in the chamber.

As a demonstration of the utility of this technique, we measured GOP and NOP of French bean (Phaseolus vulgaris) as a function of irradiance at 21% and 2% O₂. We also measured CO₂ assimilation. We then used these results to examine ∆O₂/∆CO₂ during exposure to light, the light response of g_m, and the Kok effect (apparent inhibition of respiration in the light; Kok, 1948, 1956; Hoch et al., 1963) under normal and nonphotorespiratory conditions.

RESULTS

Measurement of Net and Gross Photosynthesis

Gas-exchange equations predict that net assimilation (A_net) increases linearly at low light intensity, where light is limiting, and plateaus at high light levels, where CO₂ assimilation is limited by Rubisco activity. In Figure 1, the linear portion of the curve is restricted to light intensities lower than 160 μmol m⁻² s⁻¹. The asymptote of the photosynthesis light response curve represents the maximum rate of A_net, NOP, or GOP. The asymptote is determined using a nonrectangular quadratic function. At 21% O₂, GOP_max (the subscript max refers to the asymptotic value) is 34.2 μmol m⁻² s⁻¹ and NOP_max is 13.5 μmol m⁻² s⁻¹. At 2% O₂, GOP_max is 19 μmol m⁻² s⁻¹ and NOP_max is 15.8 μmol m⁻² s⁻¹. At any given [O₂], NOP_max and A_max are indistinguishable, as also suggested by Figure 2. The difference between GOP and NOP at 21% O₂ is greatest at maximum light intensity. This point corresponds to a reduction in C_i values (Supplemental Fig. S1). In this discussion, we acknowledge that ambient CO₂ (C_a) changed in the leaf chamber (Supplemental Fig. S1). During the experiments at 21% O₂, C_i increases from 245 ± 5 μmol mol⁻¹ at maximum light intensity to 458 ± 8 μmol mol⁻¹ in darkness. The variation in C_i with irradiance is a consequence of changes in net photosynthesis and respiration as well as variable stomatal closure and C_a in the cuvette (from 379 ± 4 to 426 ± 3 μmol mol⁻¹). As a consequence, C_a is slightly higher in the dark, where respiration adds CO₂ to ambient air, than in the light, where photosynthetic assimilation removes CO₂. The ambient CO₂ in the cuvette depends on the difference between net CO₂ assimilation by the leaf and the flux of CO₂ into the cuvette. At 2% O₂, C_i increases with decreasing light intensity. The range of increase is enhanced by the higher gross assimilation rate at high irradiances (from 129 ± 19 to 437 ± 13 μmol mol⁻¹). For irradiances above 200 μmol photosynthetically active radiation (PAR) m⁻² s⁻¹ at 2% O₂, C_i was lower than 200 μmol mol⁻¹ and considered limiting for photosynthesis. The discussion of the results focuses on the lower irradiance range (0–200 μmol PAR m⁻² s⁻¹).

Figure 1.

Light response of net CO₂ assimilation and NOP for French bean leaves exposed to 21% O₂ (black symbols) or 2% O₂ (white symbols). A and B, Light responses of net CO₂ assimilation. C and D, Light responses of NOP. Nonrectangular quadratic curves are represented by solid lines for net fluxes and by dashed lines for GOP. Errors bars represent se for n = 3.

Figure 2.

Correlation between NOP and net CO₂ assimilation for leaves exposed to 21% O₂ (black symbols) or 2% O₂ (white symbols). The dashed line represents a 1:1 relation.

All measured responses of GOP to increasing light intensities follow a similar nonrectangular quadratic function (Fig. 1, C and D). As suggested by the higher sd of fluxes in the replicate experiments (Fig. 1), GOP is most variable from leaf to leaf at high light intensities. However, this variability is less pronounced for measurements at 2% O₂ (Fig. 1). Nevertheless, the precision of the measurements is better at 2% than at 21% O₂ (for description about d¹⁸O accuracy and precision, see Supplemental Material).

Figure 2 shows a positive correlation of nearly 1:1 between NOP and A_net. At 21% O₂, the linear regression between NOP and A_net (photosynthetic quotient) is 1.066 ± 0.013 (r² = 99.6%, P < 0.01) O₂ per CO₂. At 2% O₂, it is 0.978 ± 0.013 (r² = 99.5%, P < 0.01) O₂ per CO₂. The ratio NOP/A_net for values obtained in the linear portion of the curve (light intensities lower than 160 μmol m⁻² s⁻¹) is 0.982 ± 0.097 (r² = 90.2%, P < 0.01) at 21% O₂ and 0.956 ± 0.013 (r² = 99.5%, P < 0.01) at 2% O₂. The correlation between NOP and A_net in the Rubisco-limited part of the curve (light intensity above 160 μmol m⁻² s⁻¹) is 1.026 ± 0.027 (r² = 99.2%, P < 0.01) O₂ per CO₂ at 21% O₂ and 0.999 ± 0.067 (r² = 97.3%, P < 0.01) O₂ per CO₂ at 2% O₂. The correlation between NOP and A_net for all the data presented in Figure 2 is 1.006 ± 0.012 (r² = 99.1%, P < 0.01). This correlation validates the measurement of net photosynthesis using either O₂ or CO₂ fluxes.

Stomatal Conductance and g_m

Stomatal conductance (g_s) increases with light intensity to reach a maximum around 0.1 mol m⁻² s⁻¹ (Fig. 3A), but g_s is higher at 2% O₂ than at 21% O₂. Light saturation of g_s occurs at about 160 μmol PAR m⁻² s⁻¹ at 2% O₂ and 500 μmol PAR m⁻² s⁻¹ at 21% O₂.

Figure 3.

Light response of g_s, C_c, and g_m for French bean leaves exposed to 21% O₂. A, g_s. B, C_c. C, g_m. Dotted lines represent asymptotic values for each parameter. Errors bars represent se for n = 3.

At 21% O₂, g_m increases with irradiance to an asymptotic value. As irradiance rises above about 110 μmol m⁻² s⁻¹, C_c decreases slightly from 161 ± 16 to 109 ± 5 μmol mol⁻¹ (Fig. 3, B and C). Nevertheless, C_c tends to be conserved, and variation in C_i seems to be compensated for by an increase of g_m with increasing light intensities. In fact, g_m reaches an asymptote by 400 to 450 μmol PAR m⁻² s⁻¹. At 21% O₂, this asymptote is around 0.1 mol CO₂ m⁻² s⁻¹. At 2% O₂, because O₂ uptake is similar to R_L, g_m and C_c cannot be calculated (see Eq. 18 in “Materials and Methods”).

The Kok Effect under Photorespiratory and Nonphotorespiratory Conditions

Considering the light-limited part of the experiment, Figure 4 summarizes the data for the linear portion of the light response curve. Linear regression of GOP and NOP in the low-light-intensity part of the photosynthesis-irradiance curve (0–160 μmol PAR m⁻² s⁻¹) is used for leaves exposed to either 21% or 2% O_2. At 21% O₂, the linear portion of the curve occurs between 0 and 90 μmol PAR m⁻² s⁻¹. At 21% O₂, NOP and GOP diverge as light intensities rise, reflecting increasing photorespiration at higher rates of A_net or GOP. The slope of GOP versus irradiance for values between 0 and 110 μmol m⁻² s⁻¹ is 0.07 ± 0.004 μmol O₂ produced per quanta (r² = 98.2, P < 0.01) at 21% O₂. The slope of NOP is 0.054 ± 0.002 O₂ molecules produced per quanta (r² = 99.7%, P < 0.01). The linear portion of GOP and NOP versus irradiance curves at 2% O₂ lies below 110 μmol m⁻² s⁻¹. At 2% O₂, the slopes of GOP and NOP with increasing light intensities were similar, 0.066 ± 0.002 O₂ molecules produced per quanta (r² = 99.7, P < 0.01).

Figure 4.

Low-light response of NOP (squares) and GOP (triangles) for individual leaves of French bean exposed to 21% O₂ (black symbols) or 2% O₂ (white symbols). A through F show individual low-light responses, and G and H show averaged values. NOP values are corrected for C_i variation using Kirschbaum and Farquhar (1987). Dashed-dotted lines represent the linear regression of the values obtained for light intensities between 40 and 100 μmol m⁻² s⁻¹. Error bars represent se for n = 3.

Similarly, using linear regressions of the NOP and A_net values obtained for light intensities in the linear portion of the NOP and A_net light response curves above 35 μmol PAR m⁻² s⁻¹ (as defined above) for individual leaves, a rate of respiration in the light can be obtained following the Kok method. On average, R_o-apparent is 0.54 ± 0.16 μmol m⁻² s⁻¹. In comparison, the averaged value of dark respiration (R_dark) is 1.06 ± 0.07 μmol m⁻² s⁻¹. As a consequence, R_o-apparent for 21% O₂ is inhibited by 46% at light intensities above the break point in the light response curve. At 2% O₂, linear regressions of GOP and NOP against irradiance are parallel (Fig. 4B), and the linear regression of NOP intercepts the y axis at the average value of dark respiration (R_dark = 0.79 ± 0.12 μmol m⁻² s⁻¹). Therefore, at 2% O₂, there is no discernible Kok effect and no apparent light inhibition of mitochondrial respiration.

Dark respiration values (1.06 ± 0.07 μmol m⁻² s⁻¹ at 21% O₂ and 0.79 ± 0.12 μmol m⁻² s⁻¹ at 2% O₂) are compared with R_o-corrected or R_c-corrected values calculated from either O₂ or CO₂, respectively (Fig. 5). At 21% O₂, the C_i correction increases R_o slightly from 0.59 ± 0.05 to 0.68 ± 0.04 μmol m⁻² s⁻¹. Thus, after correction, respiration is 36% lower in the light than in the dark. For R_c, the apparent value is 0.74 ± 0.14 μmol m⁻² s⁻¹ and the corrected value is 0.64 ± 0.04 μmol m⁻² s⁻¹. Although the impact of the C_i correction on the true values of R_o and R_c are different, at 21% O₂, these differences are not statistically significant using a Student’s t test comparison with P > 0.05. At 2% O₂, the correction for C_i variations decreases R_c from 1.15 ± 0.39 to 0.64 ± 0.04 μmol m⁻² s⁻¹ and decreases R_o from 0.86 ± 0.28 to 0.72 ± 0.04 μmol m⁻² s⁻¹. The respiratory quotient for apparent respiration in the light is then 1.06 ± 0.14 at 21% O₂ and 1.12 ± 0.05 at 2% O₂.

Figure 5.

Comparison of respiration rates measured in the dark (black bars) and apparent respiration rate in the light using the Kok method on A_net (gray bars) or using the Kok method on NOP (white bars) for French bean leaves exposed to 21% O₂ or 2% O₂. Values of R_Light were corrected for variations in C_i. Results of Student’s t test are represented by double asterisks when the difference from the other group is very significant (P ≤ 0.01) and by single asterisks at P < 0.05. Each group is represented by a different letter, a or b. Error bars represent se for n = 3.

DISCUSSION

Assessment of the Gas-Exchange Method

The technique developed here is a robust gas-exchange method combining cavity ring-down spectrometers (CRDS) with a high-precision isotope ratio mass spectrometer (IRMS) to more fully characterize leaf carbon, oxygen, and water vapor exchanges under near-ambient conditions. This technique takes advantage of the ability of the IRMS to achieve a precision of ±0.03‰ in the measurement of δ¹⁸O and ±0.005‰ in δO₂/N₂. This technique does not enable the measurement of previously unmeasured properties. For example, GOP has been measured before by ¹⁸O₂ labeling (Berry et al., 1978; Canvin et al., 1980; Peltier and Thibault, 1985; Biehler et al., 1997; Badger et al., 2000; Haupt-Herting and Fock, 2000; Ruuska et al., 2000). However, our technique has two main virtues. First, the use of a steady-state gas-exchange system allows the measurement of leaf properties such as R_L and g_m that could not be easily measured in the closed-system chambers used previously to measure GOP. Second, this instrument allows us to constrain and vary major environmental properties such as O₂ concentrations, irradiance, and temperature in a systematic way and, thus, to study a variety of leaf physiological processes.

Compared with previously used closed gas-exchanges systems (Laisk et al., 1992; Badger et al., 2000; Haupt-Herting and Fock, 2002), the technique developed here presents significant improvements that overcome several previous limitations. However, special care needs to be taken when preparing detached leaves. Detaching leaves from their main branch can disrupt the continuity of the water stream and lead to xylem embolism and leaf wilting (Savvides et al., 2012; Tombesi et al., 2014). As in Gauthier et al. (2010), where detached leaves were maintained alive for up to 16 h, the protocol described here (see above) routinely allowed measurements for more than 15 h. In past experiments using membrane inlet mass spectrometry, the cost of ¹⁸O₂ limited the length of experiments to only a few hours (Atkins and Canvin, 1971; Harris et al., 1983), also limiting the range of measurements and their application. The technique presented overcomes this limitation by using ambient air and a limited amount of H₂¹⁸O. Second, the utilization of entire leaves has two advantages: it limits leaf desiccation (Flexas et al., 2006) and reduces stress reactions associated with leaf cutting. Furthermore, it increases the difference between entering and exiting air and increases repeatability and precision (Supplemental Figs. S4 and S5). Finally, compared with the techniques used on attached leaves, enclosing a leaf, petiole, and water supply reduces potential system leaks (Yakir, 1992; Haupt-Herting and Fock, 2000).

The limitations associated with working on detached leaves also need to be recognized. There is evidence that g_s is regulated by water potential and plant water status (Buckley, 2017), yet in our experiments here, leaf water potential was not monitored. Thus, the low g_s and resulting C_i observed in our experiments may be associated with detaching the leaves. Alternatively, the low g_s also could have been generated by stomatal patchiness (Buckley et al., 1997; Mott and Buckley, 2000). Finally, we note the limitation associated with variable C_a inside the chamber, which led to limiting C_i values (less than 200 µL L⁻¹) under nonphotorespiratory conditions. This limiting C_i could lead to a limitation of carbamylation and the apparent inactivation of Rubisco (Sage et al., 2002). Future experiments should continue to quantify and minimize the overall stomatal response while attempting to study leaf physiological parameters under near-ambient conditions and constant C_a.

This system enables the measurement of GOP and NOP concurrently with many other commonly measured properties, such as A_net, g_s, and carbon isotope discrimination. There are a number of ways in which this instrument could be applied to important topics in plant physiology. For example, the measurement of GOP and electron flow through the photosynthetic electron transport chain also can be used to quantify the quantum yield of photosynthesis if light absorption is measured (Singsaas et al., 2001). Similarly, precise measurements of GOP coupled with NOP can constrain O₂ uptake and provide some insight into the role of the Mehler reaction (Badger et al., 2000; Haupt-Herting and Fock, 2002). Finally, GOP can be compared directly with fluorescence, which is its proxy for electron flow (Genty et al., 1989; von Caemmerer, 2000; André, 2011). Calibrating fluorescence against measured values of GOP would allow accurate estimates of this property in a wide range of studies.

NOP data, together with measurements of A_net, allow us to determine the stoichiometry of photosynthesis and respiration, expressed as ∆O₂/∆CO₂. Doing so adds the capacity to assess processes such as nitrate assimilation (Bloom et al., 1989), anaplerotic carbon fixation (Oja et al., 2007; Angert et al., 2012), or the production of lipids (Aubert et al., 1996; Devaux et al., 2003; Tcherkez et al., 2003), in addition to potential changes in the respiratory substrates and cellular energetics (Tcherkez et al., 2003; Nogués et al., 2004). In our experiments, the photosynthetic quotient was nearly constant and very close to 1, as seen by the strong 1:1 relationship between NOP and A_net (Fig. 2). This is to be expected if both the photosynthates created and the respiratory substrates used during the measurement were Glc, starch, and other basic carbohydrates. However, considering the low rate of R_L, the influence of the respiratory substrates on the net assimilation quotient is relatively small. In addition, the ratio NOP/A_net at 21% O₂ was slightly higher than 1 by a few percent. This small imbalance in O₂ production compared with CO₂ assimilation could emerge from O₂ produced during nitrogen assimilation in leaves (Bloom et al., 1989). In addition, we are mindful that, in low light, calculations of ∆O₂/∆CO₂ due to respiration in the light from CO₂ and O₂ exchange are inherently uncertain. This is because they rely on extrapolating NOP and A_net to zero irradiance. Although photosynthesis and respiratory quotients of leaves are rarely measured in gas-exchange experiments, they have the potential to add valuable information. As discussed here, with our system, it is possible to measure these properties under a variety of ambient conditions.

Response of g_m to Light Intensity

Measuring GOP allows for the calculation of both C_c and g_m (Renou et al., 1990; André 2011; Flexas et al., 2012). We found, in French bean, that g_m doubled when light increased from 200 to 1,100 μmol PAR m⁻² s⁻¹. Similar responses have been reported previously from C isotope discrimination (Δ¹³C; Hassiotou et al., 2009; Douthe et al., 2011, 2012) or fluorescence measurements (Flexas et al., 2007; Xiong et al., 2015). Douthe et al. (2011) found that, when light increased from 200 to 1,100 μmol PAR m⁻² s⁻¹, g_m increased by 60% for three Eucalyptus spp. In a similar light intensity range, Flexas et al. (2007) found a 40% increase of g_m with irradiance for Olea and Vitis spp. Finally, Hassiotou et al. (2009) reported an increase with irradiance of 22% of g_m in six species of Banksia. The response of g_m to irradiance is likely to be species dependent and may be a strategy for plants to maintain a positive carbon balance in leaves (Flexas et al., 2006). However, no consensus exists regarding the mechanism for the rapidity of the response (Douthe et al., 2011). Our results demonstrate the capacity of our system to determine g_m, adding to the short list of studies using alternative approaches to the most commonly used techniques (fluorescence or ¹³C) to study environmental responses of g_m.

Major differences exist between previously reported g_m methods and the current O₂ method. The O₂ method has some distinct advantages for studying g_m responses to environmental factors. Carbon isotope approaches use a combination of isotope effects associated with carbon metabolism (Tazoe et al., 2011) to estimate g_m. The relevant fractionation processes that need to be considered include the diffusion of CO₂ through the stomata (a = 4.4‰), the fractionation for dissolution and diffusion through water (a₁ = 1.8‰), fractionation caused by Rubisco (b = 30‰), and fractionation associated with respiration and photorespiration (e = 5.1‰ and f = 11.6‰, respectively; Gong et al., 2015). g_m is then calculated according to Tazoe et al. (2011):

(21)

where A_net is net CO₂ assimilation, p_i and p_a are CO₂ partial pressure of the ambient and intercellular airspaces, Γ* is the compensation point in the absence of R_L, and Δ_obs is the observed CO₂ discrimination calculated from the difference in C isotopes entering and leaving an open gas-exchange system (Evans et al., 1986). Photorespiratory and respiratory fluxes are determined using Γ* and R_L. By making these measurements at 2% O₂, the impact of R_L and Γ* on the estimate of g_m remains small and negligible (Tazoe et al., 2009). Comparing Equations 18 and 21, we can see that both equations constraining g_m invoke four common terms: R_L, A_net, Γ*, and C_i. The O₂ method also requires determining U − R_L, while the ¹³C method requires estimating five isotope effects and the apparent discrimination (Δ). Furthermore, Gong et al. (2015) have shown that differences in the δ¹³C between the tank gas used in gas-exchange experiments and the ambient air the plants were grown in can lead to errors in calculated values of photosynthetic discrimination. This problem is avoided when using the O₂ method, where the only unique term needed to calculate g_m is U (Eq. 16), determined from measurements of GOP and NOP. The main assumption of the O₂ method is that oxygen consumption is negligible for processes other than photorespiration and respiration (such as photoreduction by chlororespiration or the Mehler reaction). Finally, Equation 21 has been shown to overestimate g_m (Farquhar and Cernusak, 2012), and the ternary effect of transpiration rate on A_net needs to be considered for the estimates of g_m using the ¹³C method.

Mitochondrial Respiration in the Light

We found that, at 21% O₂, respiration was light inhibited (based on the nonlinearity in the light response curve of A_net), as has been commonly found in other experiments (Atkin et al., 2005; Tcherkez et al., 2008; Heskel et al., 2013). However, at 2% O₂, light did not inhibit mitochondrial respiration (i.e. there was no Kok effect). The lack of a Kok effect under ambient [CO₂] has been found previously in studies of maize (Zea mays; Cornic and Jarvis 1972) and rice (Oryza sativa; Ishii and Murata, 1978). In the case of maize, a C4 plant where photorespiration is insignificant, Cornic and Jarvis (1972) measured the light response curve at low irradiances using 0%, 21%, and 100% O₂ and did not observe a Kok effect. They concluded that the lack of a Kok effect could be explained by the lack of photorespiration under natural conditions. For rice, Ishii and Murata (1978) did not observe a Kok effect at 2% O₂ but observed a larger Kok effect with 50% O₂. This led them to conclude that photorespiration was the proximate cause of the Kok effect.

However, contrary to the two studies on maize and rice, Sharp et al. (1984) found the occurrence of a Kok effect in sunflower (Helianthus annuus) leaves under elevated [CO₂]. Since photorespiration is inhibited under high CO₂, the explanation of a direct influence of photorespiration on the Kok effect may be too simple. Sharp et al. (1984) also reported a reduction of dark respiration under low O₂. However, they suggested that it originated from the limitation of O₂ uptake by the mitochondrial electron transport chain when O₂ was less than 2% (their low-O₂ partial pressures were 1%). Our data also suggested a lower dark respiration rate at 2% O₂, but the difference was not statistically significant (Student’s t test, P = 0.056). Furthermore, Griffin and Turnbull (2013) and Yin et al. (2011) reported significant light inhibition of respiration in both C3 and C4 plants. Tcherkez et al. (2008) found that the inhibition of respiration in the light is independent of the CO₂/O₂ ratio but that, under low-O₂ conditions, the degree of inhibition may diminish. More recently, Farquhar and Busch (2017) suggested that most of the Kok effect could be explained by a variation in C_c under limiting irradiances. However, in Vicia faba, this explanation was not confirmed. Instead, Buckley et al. (2017) found that (1) the break point (Kok effect) was unaffected by [O₂], confirming the findings of Tcherkez et al. (2008), and (2) the activity of the oxygen components of mitochondrial respiration decline progressively with irradiance (Buckley et al., 2017). In fact, the degree of inhibition of respiration in the light may reflect a strong plasticity of the respiratory system in C3 plants. Comparison of this flexibility among C3 plants is needed, and the gas-exchange system presented here can contribute physiological information for future investigations.

CONCLUSION

We have presented a method that allows the standard gas-exchange techniques to be extended to the measurement of NOP and GOP. With this method, a broad suite of gas-exchange properties can be measured while regulating light, CO₂, O₂, temperature, and humidity. The light response curves presented here were used to demonstrate the method and to determine several key parameters that could help improve our understanding of fundamental leaf processes. Two significant insights are highlighted. First, mitochondrial respiration was not inhibited in the light when leaves of French bean were exposed to nonphotorespiratory conditions. This result reinforces some earlier conclusions that photorespiration could indirectly play a major role in regulating the intensity of the Kok effect in some plants, but it remains to be further investigated. Second, g_m increases with an increase in light intensity. g_m inferred from O₂ fluxes represents an alternative and complementary approach that may clarify uncertainties in its measurement. For example, Farquhar and Busch (2017) demonstrated that the uncertainty associated with the measurement of respiration in the light using the Kok effect is of the same order of magnitude as the uncertainty in C_c using conventional CO₂ gas exchanges. Comparable calculations using the O₂ method are currently missing, and a better understanding of the limitations of O₂ diffusion toward the chloroplast is needed. Major uncertainties remain around the nature of the Kok effect and g_m. Measuring O₂ fluxes could add useful information to compute more accurate C_c values and thus examine the Kok effect with sufficient precision (Tcherkez et al., 2017a, 2017b). Finally, this technique creates opportunities to study the long-term response of gross photosynthesis to environmental changes and to investigate the balance between photochemical energy production by the electron transport chain and the utilization of this energy for carbon assimilation and other cell processes.

MATERIALS AND METHODS

Plant Material

Seeds of French bean (Phaseolus vulgaris) were germinated on wet filter paper in petri dishes for 5 to 6 d in the laboratory. Once true leaves appeared, plantlets were transferred to a greenhouse and placed in trays containing potting mix for 1 week before being transferred to 10-L pots. Photosynthetic photon flux density during a 16-h photoperiod was maintained above 500 μmol m⁻² s⁻¹ using supplemental high-pressure sodium lighting. Ambient temperature was maintained at approximately 23°C/14°C day/night. All pots were watered for 5 to 10 min, two to three times per day depending on plant size, by drip irrigation. Nutrient solution (Miracle-Gro enriched with iron-EDTA; Scotts) was applied twice per week. Mature green leaves (approximately 3 to 4 weeks old) were detached from the plants under water, and the apical leaflet was used for the experiment.

Gas-Exchange System

Gas-exchange measurements were made on fully expanded leaves placed in a glass and aluminum chamber similar to that described by von Caemmerer and Hubick (1989). The chamber’s internal dimensions were 150 × 190 × 15 mm for a volume of 427.5 cm³. The air within the chamber was well mixed with a 15-cm-long cross-flow fan driven by an external DC-powered, magnetically coupled drive (MagneDrive; Autoclave Engineers). The upper surface of the cuvette was sealed with a 10-mm-thick tempered glass plate measuring 170 × 260 mm. An air-tight seal was achieved by pressing the glass lid against a Viton O-ring with eight clamps to apply homogenous pressure. Air temperature inside the chamber was controlled by a thermoelectric (Peltier) temperature controller (TC-36-25; TE Technology) in thermal contact with the bottom of the cuvette. Two type K thermocouples coupled to amplifier breakout boards (MAX31850K; Adafruit Industries), interfaced to an Arduino Uno R3 (http://arduino.cc), were used to record air and leaf temperatures. The cuvette was uniformly illuminated by three controllable customized SPYDRx LED systems (Fluence Bioengineering) placed above the chamber. A small quantum sensor (Versatile Mini; Heinz Walz) inside the cuvette was used to measure PAR reaching the adaxial side of the leaf.

During each experiment, a constant flow of compressed air entered the chamber after passing through two Drierite columns (W.A. Hammond Drierite) and a glass cryogenic trap (at −50°C) to remove moisture. Some air exiting the chamber also was dried. A schematic of the gas-exchange system is presented in Figure 6. Dried compressed air was allowed to flow through a glass capillary into the reference side of the changeover valve of the mass spectrometer (Finnigan Delta plus XP; Fig. 6). A mass-flow meter (Aalborg) was used to measure the flow rate of air exiting the chamber. A second small fraction of exiting air was sent to an isotopic H₂O CRDS (L2130-I; Picarro) for analysis of the H₂O mixing ratio in air (Supplemental Figs. S6 and S7C) and δ¹⁸O of water vapor (δ¹⁸O_v; Supplemental Figs. S6 and S7D). Transpiration rate (E) and g_s were calculated according to Buck (1981) and von Caemmerer and Farquhar (1981). Exiting air that was not sent to the water analyzer was dried by passage through a Drierite column and a glass water trap. A fraction of this air was sent to a CRDS CO₂ analyzer, which measured CO₂ (Supplemental Figs. S6 and S7A) and δ¹³C in CO₂ (Supplemental Figs. S6 and S7A). Note that the δ¹³C data were not used in any of the calculations of this article. A separate fraction of dried exiting air flowed through a glass capillary tube to the sample side of the mass spectrometer to measure δO₂/N₂ and δ¹⁸O of O₂ (Supplemental Figs. S6 and S7, B and E, respectively).

Figure 6.

Schematic of the gas-exchange system used for the experiments. MFC, Mass flow controller; WT, water trap; CRDS, cavity ring-down spectrometer; IRMS, isotope ratio mass spectrometer; MFM, mass flow meter. Subscript v is water vapor.

Petiole Box

For each experiment, the leaf’s petiole was placed in a 40- × 30- × 20-mm sealed aluminum box containing 22 cm³ of ¹⁸O-labeled water. The entire leaf and petiole box was placed inside the leaf chamber. In this way, any potential leak caused by the petiole or leaf blade crossing the side wall of the chamber was avoided (as described previously by von Caemmerer and Hubick [1989] and Yakir et al. [1994]). The aluminum lid of the petiole box had a 10- × 10-mm aperture that allowed for the passage of the petiole. Underneath the lid, a 38- × 28- × 1-mm Viton sheet with smaller aperture was custom fit to the variable diameter of each leaf’s petiole. The aluminum plate and Viton sheet suppressed evaporation from the petiole box to the point where it did not have a discernible influence on the concentration or δ¹⁸O of water vapor exiting the cuvette. Thus, transpiration was the only significant source of water vapor (v) leaving the cuvette.

Leaf Water Labeling

An entire fully expanded leaf was detached from a plant, and the petiole was immediately immersed in tap water. The petiole was then recut under water to remove any potential embolism. After 30 min in darkness, the side leaflets were removed and the apical leaflet was used in the experiment. The leaf was transferred to the petiole box filled with labeled water (as described above) at 8,806‰ and 8,735‰ δ¹⁸O for 21% O₂ and 2% O₂ experiments, respectively. The petiole was again recut under water inside the box, and the entire system (leaf + petiole box) was sealed inside the chamber. The course of leaf water labeling was followed using the CRDS (Fig. 6) to measure δ¹⁸O_v of transpired water vapor exiting the chamber (Supplemental Figs. S6 and S7D).

Calculating GOP, NOP, and O₂ Uptake

GOP is defined as the rate of O₂ production by water splitting and NOP as the sum of all O₂ fluxes between air and the leaf. The δ¹⁸O of leaf water (δ¹⁸O_L) is calculated first. Leaf water is the substrate for photosynthesis, and the δ¹⁸O of photosynthetic O₂ approximates δ¹⁸O_L (Guy et al., 1987). Mass balance equations constrain GOP and NOP, given the observed production rate of the two major isotopologues of O₂, ¹⁶O₂ and ¹⁸O¹⁶O. O₂ uptake is then calculated by difference. Other isotopologues of O₂ have very low abundances, make a trivial contribution to O₂ fluxes, and are neglected. Errors introduced in GOP by this neglect are 0.04% for ¹⁷O¹⁶O and less for the less abundant ¹⁷O₂, ¹⁸O¹⁷O, and ¹⁸O₂.

Air is run through the cuvette until δ¹⁸O of water vapor exiting the chamber reaches a constant value, at which point the water balance is at steady state. Then, the isotopic composition of water entering the leaf through the petiole should be similar to that of water exiting the leaf through the transpiration stream. Thus, the flux of H₂¹⁸O derived from transpiration (i.e. F_18O-v), is calculated as:

(1)

However, the H₂¹⁸O/H₂¹⁶O ratio of water vapor (v) is less than its liquid source due to kinetic and equilibrium isotope fractionation during evaporation (Harwood et al., 1998). Applying the standard Craig-Gordon model of evaporation (Farquhar et al., 1989):

(2)

where R_e is the isotopic ratio at the site of evaporation, R_s is the isotopic ratio of the source of water (here, xylem), R_v is the isotopic ratio of the water vapor, α⁺ is the equilibrium isotope fractionation at the leaf temperature (1.009; Majoube 1971), α_k is the kinetic isotope effect for the evaporation of water from the leaf into dry air (1.032; Cernusak et al., 2016), and h is relative humidity.

We initiated the measurement of the light response after 5 to 6 h at maximum illumination, when the δ¹⁸O of transpired water reached an asymptotic value. We assumed that the leaf was at steady state, and so the water entering by the petiole is equivalent to the water exiting the leaf, in this case, R_s = R_v.

We further assumed that the water at the site of evaporation is a good representation of bulk leaf water, such that R_e = R_leaf, where R_leaf is the isotopic ratio of leaf water. Here, the ¹⁸O/¹⁶O ratio of H₂O will be represented as , and so R_v = and R_leaf =

Equation 2 can now be rearranged and solved for the ratio of H₂¹⁸O/H₂¹⁶O in the evaporating liquid (leaf water) as a function of the ratio in the vapor, the kinetic and equilibrium O isotope fractionations, and the humidity:

(3)

(H₂¹⁸O/H₂¹⁶O)_v is the ratio of water vapor measured by the CRDS in the air exiting the chamber. This equation shows that the H₂¹⁸O/H₂¹⁶O of the liquid will be 1% to 2% higher than the ratio in the vapor, regardless of whether water entering the petiole is labeled. When calculating the gross rate of photosynthesis, this leaf water enrichment is taken into account.

The basic mass balance equation for O₂ and its major isotopologues (¹⁶O₂ and ¹⁸O¹⁶O) flowing through the cuvette is:

(4)

where GOP is gross O₂ production and U is O₂ uptake by respiration, photorespiration, and all other processes. To distinguish this flux from individual isotope fluxes, the production of each isotopologue will be represented by P.

An analogous equation can be written for ¹⁶O₂:

(5)

In calculating O₂ fluxes, we consider only two isotopologues, ¹⁶O₂ and ¹⁸O¹⁶O. The ratio of ¹⁸O¹⁶O/¹⁶O₂ is nominally 1:250. We do not consider ¹⁸O₂ because its abundance is so low (¹⁸O₂/¹⁶O₂ ∼ 1:250,000). Equation 5 can be expanded to:

(6)

The factor of 2 in the first term on the right-hand side reflects the fact that an O₂ molecule has two chances to acquire an ¹⁸O atom from water splitting. The subscript cuv indicates ambient air inside the cuvette. The term α_U represents the isotope effect associated with respiration and all other uptake processes (photorespiration, alternative pathway, Mehler reaction…). Calculated fluxes are essentially insensitive to α_U (0.982), because the δ¹⁸O of photosynthetic O₂ differs from that of ambient O₂ more than the δ¹⁸O of O₂ consumed by respiration.

Solving Equation 5 for U_16O2, substituting into Equation 6, and solving for P_16O2 gives:

(7)

In Equation 7, H₂¹⁸O/H₂¹⁶O_L = (H₂¹⁸O/H₂¹⁶O_VSMOW)(1 + δ¹⁸O_l-VSMOW/10³) and ¹⁸O¹⁶O/¹⁶O_2cuv = 2(H₂¹⁸O/H₂¹⁶O_VSMOW)(1 + δ¹⁸O_air-VSMOW/10³)(1 + δ¹⁸O_cuv-air/10³). The subscripts air and cuv refer to O₂ in ambient air and air within the cuvette, respectively, as above. VSMOW indicates that the isotopic composition is expressed with Vienna Standard Mean Ocean Water as the reference. ΔF is the flux out of the chamber minus the flux into the chamber. This flux is the flux of ¹⁶O₂ or ¹⁸O¹⁶O out of the chamber minus the flux in. These fluxes are calculated from the change in ¹⁶O₂/¹⁴N₂ and ¹⁸O¹⁶O/¹⁴N₂ as air passes through the cuvette, measured by the IRMS. In calculating these fluxes, we assume that the N₂ concentration of entering and exiting air is the same, the O₂ mixing ratio of air is 0.0201784, and the ¹⁸O¹⁶O/¹⁶O₂ ratio of ambient air is 0.0040104.

Considering that (1 + ¹⁸O/¹⁶O)² is very close to 1, the production of ¹⁸O₂ (and other isotopologues) is negligible and the GOP is:

(8)

NOP is simply the sum of the outgoing minus incoming fluxes of ¹⁶O₂ (Eq. 5) and ¹⁸O¹⁶O (Eq. 6):

(9)

The difference between outgoing and incoming fluxes of ¹⁶O₂ is determined using δO₂/N₂ (see below). Net O₂ uptake is the difference between GOP and NOP (GOP − NOP).

Mass Spectrometric Measurements of GOP and NOP

To determine GOP and NOP, we used a Thermo Delta plus XP IRMS (Thermo Finnigan) configured to simultaneously measure masses 28, 29, 32, 33, 34, 38, 40, and 44. The changeover valve of the mass spectrometer was connected to the gas-exchange system using two 25-μm-i.d. glass capillaries (SGE Analytical Science). The connections of the capillaries with the chamber and the changeover valve were secured with vespal ferrules (Agilent Technologies). To equalize sample and reference side ion currents, the capillary length was set to 149 cm on the sample side and to 145 cm on the reference side. Each measurement consisted of a block of 16 to 25 cycles (reference/sample/reference changeovers). For each cycle, δO₂/N₂ (32/28) and δ¹⁸O of O₂ (34/32) were calculated from the difference in isotope ratios of air leaving and entering the cuvette. Both elemental and isotopic ratios were expressed in the standard δ notation as:

(10)

Ratio_sample and Ratio_reference are the measured ion current ratios of ¹⁶O₂/¹⁴N₂ and ¹⁸O¹⁶O/¹⁶O₂.

Light Response Curves

Light response curves of net CO₂ assimilation, NOP, and GOP were obtained by decreasing PAR measured at the top of the leaf from high light to darkness. Typical data collected from an experiment at 21% O₂ are presented in Figure 1, based on results from 10 decreasing light levels (1,100, 800, 500, 200, 100, 80, 70, 60, 40, and 0 μmol m⁻² s⁻¹). Similar experiments were done using air containing 2% O₂ to minimize photorespiration. For the 2% O₂ experiments, maximal light intensity was reduced to 700 μmol m⁻² s⁻¹ and fluxes were measured at a greater number of lower light intensities (900, 700, 420, 300, 150, 100, 90, 80, 70, 60, 40, 35, and 0 μmol m⁻² s⁻¹). Typical data collected from an experiment at 2% O₂ are presented in Figure 1. Light response curve data were fitted with the nonrectangular quadratic function of Ögren and Evans (1993) using the curve-fit routine of SciPy (Oliphant, 2007; Millman and Aivazis, 2011; van der Walt et al., 2011). This routine uses nonlinear least squares for the specified functional form.

Mesophyll Conductance

g_m was evaluated using the common approach of adopting a single resistance (g_m) to characterize the transport of CO₂ between the intercellular airspace and the site of carboxylation (Tholen et al., 2012). g_m was calculated using the O₂ equations developed by Renou et al. (1990). O₂ uptake (U) is given by the sum of the rate of Rubisco oxygenation (V_o), the rate of oxygen consumption by the oxidation of glycolate during the C₂ cycle (0.5V_o), and mitochondrial respiration and other processes (Mehler reaction, plastoquinone terminal oxidase…; designated as R_o):

(11)

A similar equation was developed from Farquhar et al. (1980) for CO₂ assimilation:

(12)

where V_c is the velocity of carboxylation and R_c is the rate of decarboxylation in the light associated with mitochondrial respiration. Solving Equation 11 for V_o, combining with Equation 12, and solving for V_c gives:

(13)

In addition, Farquhar and von Caemmerer (1982) defined C_c as:

(14)

Γ* is the CO₂ compensation point in the absence of respiration in the light. By combining Equations 13 and 14, we obtain:

(15)

The limitation of photosynthesis through g_m can be written, following Fick’s law, as:

(16)

And so:

(17)

Here, Γ* was not measured but assumed to be 42.75 μmol mol⁻¹ at 25°C, as estimated by Bernacchi et al. (2001). A standardized Arrhenius function and assumed activation energy of 37.8 kJ mol⁻¹ were used to account for the effect of temperature on Γ*, as in Medlyn et al. (2002).

In the case where R_c = R_o, as assumed by Renou et al. (1990), Equation 17 can be simplified as in Flexas et al. (2012):

(18)

where R is respiration rate in the light.

Determining Terms Needed to Calculate g_m

The rate of respiration in the light (R) was determined by extrapolating the linear portion of the A_net versus irradiance curve to zero irradiance (Kok, 1948). It is important to note that this extrapolation assumes that it is possible to extrapolate from conditions where C_c < C_a, above the break point in the light response curve, to conditions where C_c > C_a, below the light compensation point. This approach to assessing the respiration rate in the light was challenged recently (Farquhar and Busch 2017; but see Buckley et al., 2017) but is provisionally retained.

The zero-irradiance intercepts of the light response curves of NOP were calculated from net C assimilation rates measured between 40 and 110 μmol m⁻² s⁻¹. These values were then equated to both R_o and R_c. To account for the dependence of A_net or NOP on changes in internal CO₂ concentration, R_o and R_c were recalculated according to Kirschbaum and Farquhar (1987). This correction uses an estimate of the rate of ribulose 1,5-bisphosphate regeneration, V_j, calculated from A_net or NOP. It constrains R_c and R_o when the intercept of the linear regression of V_j versus light is zero. Uncorrected values are reported as R_o-apparent and R_c-apparent. Corrected values are reported as R_o-corrected and R_c-corrected. R_o-corrected and R_c-corrected values were used in Equations 11 and 12 and in all further calculations. The parameters V_j and R_o were used to correct NOP for variable C_i; V_j and R_c were used to correct A_net.

Typical Experimental Protocol

Before and during each experiment, the reference gas (incoming air) was analyzed against itself to assess the magnitude and drift in the zero enrichment (shaded areas in Supplemental Figs. S6 and S7). The zero enrichment is the artifactual difference in the composition of incoming and outgoing air when there is no source or sink for O₂ within the chamber. Leaves were initially placed in the dark for 30 min. The light intensity was then raised to maximize transpiration and accelerate leaf labeling. Leaves were maintained in this condition until δ¹⁸O of the water vapor reached isotopic steady state. At this point, the δ¹⁸O of transpired water leaving the chamber was constant (Supplemental Figs. S6 and S7D), and the measure of O₂ and CO₂ fluxes began, along with their light response, as described above. For the duration of an experiment, leaves were exposed to either 21% O₂ (Supplemental Fig. S6) or 2% O₂ (Supplemental Fig. S7) in air. Air temperature was maintained at 20°C ± 1°C. Averaged leaf area was 46.23 ± 1.3 and 81.06 ± 2.44 cm² at 21% O₂ and 2% O₂, respectively. Average flow rate was measured at 2.47 ± 0.1 and 2.64 ± 0.1 L min⁻¹ at 21% O₂ and 2% O₂, respectively, under our typical conditions (1 atm pressure and 20°C ± 1°C).

Comparative Estimates of Photosynthetic Properties

Alternative formulations from the literature were used to calculate J_g and g_m in order to evaluate the accuracy of our calculations from GOP measurements. The theoretical electron transport rate, J_g, was calculated as by Flexas et al. (2012; Eq. 12.18):

(19)

C_c was arbitrarily calculated from ambient CO₂ exiting the chamber, C_a, as C_c = 0.4 C_a following Roupsard et al. (1996).

To determine the accuracy of g_m, theoretical g_m was calculated using Equation 7 from Harley et al. (1992), which considers a variable J (for review, see Flexas et al., 2012, and refs. therein):

(20)

Usually, J_a is obtained indirectly using fluorescence measurements. In the absence of such measurements, we assumed J_a = 4*GOP, as in von Caemmerer (2000), since four electrons are produced for each molecule of O₂ released by PSII. This estimate of g_m is referred to as g_m Harley (Supplemental Figs. S2 and S3).

Statistical Analysis

Linear regressions were used to determine the linearity in the light response of NOP, GOP, and A_net. These regressions also were used to calculate the Kok effect. Uncertainties correspond to a confidence interval of 95%. Linear regressions also were used to determine the correlation of O₂ versus CO₂. Independent Student’s t tests were used to assess the statistical difference between dark respiration and R_L values and between R_L values from A_net and NOP. Differences between means were considered significant at p-value < 0.05 and highly significant at p-value < 0.01.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. The variation of C_a and C_i with light intensity for French bean leaves exposed to 21% O₂ and 2% O₂.

• Supplemental Figure S2. Correlation between g_m calculated using the O₂ method and theoretical g_m calculated using the variable J method of Harley et al. (1992).

• Supplemental Figure S3. g_m versus irradiance for French bean leaves exposed to 21% O₂.

• Supplemental Figure S4. δO₂/N₂, NOP, δ¹⁸O, and GOP for each of three replicate incubations versus irradiance for French bean leaves exposed to 21% O₂.

• Supplemental Figure S5. δO₂/N₂, NOP, δ¹⁸O, and GOP for each of three replicate incubations versus irradiance for French bean leaves exposed to 2% O₂.

• Supplemental Figure S6. Measurements for a French bean leaf exposed to 21% O₂.

• Supplemental Figure S7. Measurements for a French bean leaf exposed to 2% O₂.