Drought exerts a greater influence than growth temperature on the temperature response of leaf day respiration in wheat (Triticum aestivum)

Abstract

We assessed how the temperature response of leaf day respiration (R _d) in wheat responded to contrasting water regimes and growth temperatures. In Experiment 1, well‐watered and drought‐stressed conditions were imposed on two genotypes; in Experiment 2, the two water regimes combined with high (HT), medium (MT) and low (LT) growth temperatures were imposed on one of the genotypes. R _d was estimated from simultaneous gas exchange and chlorophyll fluorescence measurements at six leaf temperatures (T _leaf) for each treatment, using the Yin method for nonphotorespiratory conditions and the nonrectangular hyperbolic fitting method for photorespiratory conditions. The two genotypes responded similarly to growth and measurement conditions. Estimates of R _d for nonphotorespiratory conditions were generally higher than those for photorespiratory conditions, but their responses to T _leaf were similar. Under well‐watered conditions, R _d and its sensitivity to T _leaf slightly acclimated to LT, but did not acclimate to HT. Temperature sensitivities of R _d were considerably suppressed by drought, and the suppression varied among growth temperatures. Thus, it is necessary to quantify interactions between drought and growth temperature for reliably modelling R _d under climate change. Our study also demonstrated that the Kok method, one of the currently popular methods for estimating R _d, underestimated R _d significantly.

Summary Statement

Leaf day respiration (R _d) plays an essential role in maintaining the primary metabolic and physiological functions of plants. Here, we estimated R _d under various water and growth temperature regimes to explore how R _d acclimates to environmental variables and evaluated if a simple method can estimate R _d for photorespiratory conditions by using gas exchange and chlorophyll fluorescence data.

1. INTRODUCTION

The ongoing global climate change has resulted in frequent and intense extreme climatic events, such as heat waves, cold snaps and drought spells (IPCC, ²⁰²¹; Lloret et al., ²⁰¹²; Perkins‐Kirkpatrick & Lewis, ²⁰²⁰; Solomon et al., ²⁰⁰⁹). Understanding how these climatic events affect crop physiological processes, particularly photosynthesis and respiration, will be critical for global food security and modelling crop productivity as well as for carbon budgets of agricultural ecosystems in response to climate change (Heskel et al., 2013; Lobell & Gourdji, ²⁰¹²; Yin & Struik, ²⁰¹⁷). Respiration plays an essential role in maintaining primary metabolic and physiological functions of plants and costs ca. 40% of gross photosynthetic assimilates of whole plants (Amthor, 2010; Gifford, ¹⁹⁹⁵). Therefore, it strongly affects not only the daily net carbon gain, nutrient acquisition and growth of individual plants but also the carbon fluxes at the ecosystem level (Tcherkez & Atkin, 2021; Tcherkez et al., ^2017a), in an ever‐changing environment.

Respiration occurring in leaves, the metabolically most active plant organs, accounts for a large part of the whole plant respiration (Atkin et al., 2007). Leaf respiration is sensitive to short‐term (minutes to hours) fluctuations in leaf temperature (T _leaf) and also acclimates to long‐term (days) growth temperature changes (Atkin & Tjoelker, 2003). The response of respiration to short‐term changes in temperature is often quantified by the parameter called 'activation energy' (E _a) of the Arrhenius model or by the Q ₁₀ factor (Atkin & Tjoelker, 2003). The thermal acclimation of leaf respiration has been widely investigated (e.g., Atkin et al., ²⁰⁰⁶; Coast et al., ²⁰²⁰; Way et al., ²⁰¹⁹). The degree to which leaf respiration acclimates to growth temperature differs among species and developmental stages of leaves (Atkin & Tjoelker, 2003; Atkin et al., ²⁰⁰⁵). Often, leaf respiration acclimates to a sustained warmer growth temperature by decreasing its rate at a reference temperature and/or its thermal sensitivity (i.e., E _a or Q ₁₀, the slope of the response curve), while acclimation to cooler temperature increases the values of these parameters (Atkin & Tjoelker, 2003). However, this acclimation is not always observed. A recent study that was conducted on the fast‐growing species Eucalyptus globulus showed an upregulation in basal rates measured at 25°C and E _a of leaf respiration in warm‐grown plants, although the underlying mechanisms are speculative (Crous et al., 2017).

Leaf respiration is also modulated by soil water availability, and drought‐induced changes in respiration could be associated with changes in the availability of substrates (e.g., soluble sugars and other carbohydrates), demand for respiratory products (e.g., ATP and NADH) and capacity of respiratory enzymes (Atkin & Macherel, 2009). The impact of drought stress (DS) on leaf respiration varies with species, drought severity and drought duration (Flexas et al., 2005). In ca. two‐thirds of the studies reviewed by Atkin and Macherel (2009), leaf respiration was reduced by drought, while in the remaining studies it was unaffected or occasionally increased. Particularly, the resilience of leaf respiration in drought was observed under cool and moderate measurement temperatures (Gauthier et al., 2014; Gimeno et al., ²⁰¹⁰). Furthermore, the effect of drought on leaf respiration can also interact with the effects of other environmental factors, for example, those of short‐ or long‐term temperature changes and elevated atmospheric CO₂, especially under field conditions (Ayub et al., 2011; Crous et al., ²⁰¹¹, ²⁰¹²; Gauthier et al., ²⁰¹⁴). These interactions will lead to more unpredictable responses of leaf respiration to drought.

There is growing evidence that the metabolic pathways of leaf respiration vary between illuminated and nonilluminated leaves, as a result of the inhibition of respiration in the light (Tcherkez et al., ^2017a, ^2017b; Tcherkez & Atkin, ²⁰²¹). Unlike leaf respiration in the dark (R _dk), leaf day respiration (R _d) occurs simultaneously with photosynthetic CO₂ assimilation and other physiological processes, such as photorespiration, reassimilation and photoinhibition, in the daytime (Yin et al., ^2020a). R _d is an important parameter in modelling net photosynthetic CO₂ assimilation (Farquhar et al., 1980) and can influence the estimation of other key photosynthetic parameters, such as V _cmax, the maximum rate of ribulose 1,5‐bisphosphate carboxylase/oxygenase (Rubisco) carboxylation (De Kauwe et al., 2016). As the model of Farquhar et al. (1980), "FvCB model" hereafter is widely used as the basic model for predicting leaf photosynthesis that is to be scaled up to the ecosystem level, R _d is also crucial in modelling ecosystem gross CO₂ efflux. However, the different metabolic pathways of R _d and R _dk may result in different responses to environmental variables (Gulías et al., 2002; Way et al., ²⁰¹⁹). Compared with the abundance of studies that have explored the environmental impacts on R _dk, the experimental data on how R _d responds to various environments is generally more lacking, probably because it is difficult to measure R _d.

Although techniques to quantify R _d have been implemented for decades, there is still debate about the best technique to quantify R _d (Tcherkez et al., ^2017a, ^2017b; Tcherkez & Atkin, ²⁰²¹). Either direct or indirect techniques have been developed (Berghuijs et al., 2019; Gong et al., ²⁰¹⁵; Haupt‐Herting et al., ²⁰⁰¹; Kok, ¹⁹⁴⁸; Laisk, ¹⁹⁷⁷; Yin et al., ²⁰⁰⁹). Direct measurement of R _d (e.g., Gong et al., ²⁰¹⁵; Haupt‐Herting et al., ²⁰⁰¹) requires sophisticated devices, which are often unavailable. Indirect estimation of R _d by gas exchange measurements in ecophysiological studies is mostly based on either the Kok method (Kok, 1948) or the Laisk method (Laisk, 1977). The Kok method exploits the Kok effect, which is the abrupt decrease in the slope of photosynthetic CO₂ assimilation rate (A) against irradiance at around the light compensation point (10–40 μmol m⁻² s⁻¹). This abrupt switch is interpreted as the consequence of light inhibition to leaf respiration, and thus R _d can be calculated as the intercept of this linear relationship using points above the breakpoint, while the intercept of the linear relationship below the breakpoint is interpreted as R _dk (Heskel et al., 2013; Tcherkez et al., ^2017b; Yin et al., ^2020a). In the Laisk method, the response of A to low intercellular CO₂ concentration (C _i) is obtained at several (commonly three) levels of irradiances. These curves theoretically intersect at a common point, where the value of A represents R _d and the value of C _i represents the CO₂ compensation point in the absence of respiration. The R _d estimate by the Kok method is often somewhat lower than the estimate by the Laisk method (Villar et al., 1994; Yin et al., ²⁰¹¹), probably because the Kok method assumes that the PSII electron transport efficiency (Φ ₂) is constant across the light levels used for the gas exchange measurements. A modified Kok method, now known as the Yin method (see Tcherkez et al., ^2017a) was developed to overcome this weakness of the Kok method, by incorporating the information from chlorophyll fluorescence that accounts for the decline of Φ ₂ with light intensity (Yin et al., 2009, ²⁰¹¹). The Yin method gave estimates of R _d that were comparable with those from the Laisk method (Yin et al., 2011), with the benefit that the measurements are easier and less time‐consuming to implement than those with the Laisk method provided that a fluorometer is available for the concurrent gas exchange and chlorophyll fluorescence measurements.

Although the Kok method, and sometimes also the Yin method, have been applied to common photorespiratory (PR) conditions, theoretically both methods require measurements that are undertaken under nonphotorespiratory (NPR) conditions. This is because both methods using simple linear regression to estimate R _d implicitly assume that the chloroplast CO₂ partial pressure (C _c) is constant across light levels, whereas a modelling study demonstrated that under PR conditions, C _c sharply decreased (thus the relative amount of photorespiration increased) with increasing irradiance (Farquhar & Busch, 2017). The significant change of C _c with irradiance (even if the ambient CO₂ level is maintained constant) is the result of stomatal and mesophyll regulation of CO₂ diffusion inside the leaf, ensuring some reassimilation of CO₂ released by photorespiration and respiration. Using leaf anatomical data combined with two‐dimensional (2‐D) modelling that accounts for CO₂ diffusion, and thus reassimilation, Berghuijs et al. (2019) were able to estimate R _d based on gas exchange and chlorophyll fluorescence data for PR conditions; they showed that applying the Kok and the Yin methods to PR conditions causes an underestimation of R _d. If the 2‐D modelling of the CO₂ diffusion can reliably estimate R _d, a simpler method, that is, the coupled FvCB and g _m (mesophyll conductance) model, can also be explored to estimate R _d by fitting the coupled model to gas exchange and chlorophyll fluorescence data obtained under PR conditions, because the coupled model can implicitly consider the reassimilation of (photo)respirated CO₂ (Yin et al., 2021). The coupled model has a nonrectangular hyperbolic (NRH) form (von Caemmerer, ²⁰⁰⁰), and this NRH model was exploited by Yin and Struik (2009) to estimate g _m from the combined gas exchange and chlorophyll fluorescence data, which was shown to be much more reliable than the well‐known variable J method of Harley et al. (1992) for estimating g _m. Here, we will test if this model can well estimate R _d under PR conditions, even if g _m is unknown beforehand.

Wheat (Triticum aestivum L.), as one of the most important staple food crops, is grown worldwide. There is little practical data on how R _d in C₃ Poaceae species like wheat varies with environmental variables, in contrast to the case for tree species, for which studies under either climate‐controlled or field conditions have been extensively published in the recent past (e.g., Crous et al., ²⁰¹¹, ²⁰¹²; Kumarathunge et al., ²⁰²⁰; Way et al., ²⁰¹⁹). To date, to our knowledge, no single study provides a set of data on R _d and its sensitivity to T _leaf in response to a combination of contrasting water regimes and growth temperatures in wheat. Understanding how R _d responds to these environmental variables is critical for quantifying wheat productivity in response to current and future climate change scenarios.

The main objective of this study is to identify the impacts of soil water deficit and growth temperatures on the instantaneous temperature response of R _d in wheat. We hypothesize that R _d and its sensitivity to T _leaf would acclimate to contrasting water regimes and to different growth temperatures, and the thermal acclimation of R _d may differ between two contrasting water treatments. We estimated R _d using the established Yin method that requires measurements under NPR conditions. We also aimed to estimate R _d for common PR conditions; so, the NRH method was explored to estimate R _d as well. We then evaluated if acclimation differed between PR and NPR conditions. Moreover, we also compared the R _d estimated by the Kok, Yin and NRH methods.

2. MATERIALS AND METHODS

2.1. Plant materials and growth conditions

To determine any interactive effect of water deficit and genotype on R _d of wheat plants, an experiment was conducted in a climate‐controlled glasshouse at Wageningen University & Research in 2019 (EXP2019). Two winter wheat (T. aestivum L.) genotypes, Thésée and Récital, known to have different drought tolerance, were used in this experiment. Four batches of seeds were sown at 10‐day intervals, creating four replicates. In each replicate, seeds were germinated on a moist filter paper in Petri dishes (one night at room temperature followed by 24 h at 4°C), and then the seedlings were transplanted to multicell seedling trays in a glasshouse. When the first leaf fully emerged (ca. 1 week after sowing), plants were moved to a 4°C cold room (12 h day length, 50 µmol m⁻²​​ s⁻¹ photon flux density) to vernalize for 7 weeks. After vernalization, plants were transplanted to 7‐L pots (three plants per pot) filled with a mixture of black soil and peat in a 2:1 (v:v) ratio. The soil mixture was 6 kg per pot and mixed with 1 g N, 1  g P and 1 g K. All potted plants were then moved to a glasshouse and the position of pots were rotated daily and randomly. The climate condition in the glasshouse compartment was set as: 400 ± 5 ppm atmospheric CO₂ concentration, 22/16 ± 2°C day/night air temperature (20°C daily average), 75 ± 5% relative humidity (corresponding to a vapour pressure deficit (VPD) of 0.66/0.46 kPa for day/night), 16 h photoperiod, and photon flux density at >400 µmol m⁻² s⁻¹ supplied by sunlight plus supplementary sodium lamps. To avoid any nutrient deficiency, 0.5  g N and 0.25 g N were applied to each pot at the tillering and the stem‐elongating stages, respectively.

Another experiment was conducted in a climate‐controlled growth chamber in 2020 (EXP2020) to examine whether R _d can be altered by different water regimes and growth temperatures. EXP2019 showed no significant difference between the two genotypes in R _d and its temperature response (see Section 3); thus, only one genotype, Thésée, was used in EXP2020. Considering varying crop durations in different growth temperature regimes, four batches of seeds were sown at 14‐, 17‐ and 20‐day intervals, respectively, creating four replicates. The same vernalization treatment and plant management practices as in EXP2019 were applied. After vernalization, the potted plants were moved to a climate chamber. The climate settings in the climate chamber were set as: atmospheric CO₂ concentration, 400 ppm; day/night temperature, 21/17°C; relative humidity, 65%; VPD, 0.87/0.68 kPa for day/night; photon flux density, ca. 410 µmol m⁻² s⁻¹ at soil level; photoperiod, 16 h.

2.2. Experimental treatments

In EXP2019, drought treatment was applied at anthesis in both genotypes. From sowing to anthesis all pots were constantly irrigated to 90% soil water holding capacity, with a gravimetric soil water content of ca. 42%. After anthesis, four replicates of each genotype were well irrigated as control plants (well‐watered treatment, WW), whereas the other four replicates of each genotype were subjected to drought stress (treatment DS) by withholding irrigation until the gravimetric soil water content reduced to ca. 16% (drought‐stressed treatment, DS). This drought level was maintained until the end of measurements (ca. 1 week).

In EXP2020, combinations of water deficit treatments and growth temperature were applied. At the booting stage, before flag leaves appeared, plants were allocated to three climate chambers with different day/night air temperature settings: high temperature (HT: 28/24°C, average 26.67°C), medium temperature (MT: 21/17°C, average 19.67°C) and low temperature (LT: 16/12°C, average 14.67°C). MT is considered as the control treatment since the temperature was the same as the growth temperature before the temperature treatments started. To minimize any confounding impact of varying VPD, VPD was set identically across chambers (0.87/0.68 kPa for day/night VPD); as a result, 77%, 65% and 52% relative humidity were applied for HT, MT and LT treatments, respectively. After anthesis, plants from each climate chamber were subjected to the same two soil water treatments as in EXP2019.

2.3. Gas exchange, chlorophyll fluorescence and leaf N measurements

In both experiments, simultaneous gas exchange and chlorophyll fluorescence measurements were carried out on flag leaves at six leaf temperatures (T _leaf; from 15°C to 40°C with 5°C intervals, except for the LT plants in EXP2020, which were measured at T _leaf from 12°C to 35°C), 10 days after the onset of the drought treatment, by using a portable photosynthetic system (Li‐Cor 6800; Li‐Cor Inc.) with an integrated fluorescence chamber head of 6 cm². Li‐Cor 6800 and plants were together moved to a climate cabinet during measurement to achieve the desired T _leaf. The VPD in the cuvette increased with an increase in T _leaf and ranged from 1.0 kPa (at 15°C) to 3.0 kPa (at 40°C) for all plants in EXP2019 as well as HT and MT plants in EXP2020. For LT plants in EXP2020, VPD ranged from 1.0 kPa (at 12°C) to 2.5 kPa (at 35°C). For a given T _leaf, incident‐irradiance response curves (A–I _inc) were assessed under both PR (i.e., 21% O₂ combined with 400 ppm ambient CO₂ (C _a)) and NPR (i.e., 2% O₂ combined with 1000 ppm C _a) conditions on the same leaf. For the measurements at NPR conditions, a gas cylinder containing a mixture of 2% O₂ and 98% N₂ was used. Gas from the cylinder was supplied to the Li‐Cor 6800, where CO₂ was blended with the gas. Photon flux densities in the measurement chamber were 200, 150, 120, 90, 60, 40 and 0 µmol m⁻² s⁻¹ (applied in that order; the value of A at 0 µmol m⁻² s⁻¹ of I _inc represents R _dk) with 5–6 min for each step. The measurements were conducted randomly in each treatment and T _leaf. The operating Φ ₂ was determined at each light step as (1 – F _s/F' _m) (Genty et al., 1989), where F _s is the steady‐state fluorescence and F' _m is the maximum fluorescence during the saturating light pulse determined by the multiphase flash method (Loriaux et al., 2013).

After the measurements, the portion of the flag leaves used for gas exchange and chlorophyll fluorescence measurements was cut to measure leaf N elemental content. The rectangle area of the leaf portion was calculated as length multiplied by width, which was measured by a vernier caliper. Then, the leaf material was weighed after drying in a forced‐air oven at 70°C to a constant weight. The concentration of total N in leaf material on mass basis (N _mass, mg g⁻¹) was analysed using an EA1108 CHN‐O Element Analyser (Fisons Instruments) based on the micro‐Dumas combustion method. From these data, leaf N content on area basis (N _area, g m⁻²) was calculated.

2.4. Estimation of day respiration under NPR conditions

R _d was estimated by the Yin method (R _d(Yin)) (Yin et al., 2009, ²⁰¹¹). The theoretical basis of this method is the model for photosynthetic rate A limited by the light‐dependent electron transport rate (Yin et al., 2004):

$$A=J_2\left(1-\frac{f_{\text{pseudo}}}{1-f_{\text{cyc}}}\right)\left(\frac{C_{\mathrm{c}}-{\mathrm{\Gamma }}_{*}}{C_{\mathrm{c}}+{2\mathrm{\Gamma }}_{*}}\right)/4-R_{\mathrm{d}},$$ (1)

where J ₂ is the total rate of e⁻ transport passing photosystem II (PSII), f _cyc and f _pseudo represent fractions of the total e⁻ passing PSI that follow cyclic and pseudocyclic pathways, respectively, C _c is the chloroplast CO₂ partial pressure, and Γ_* is the C _c‐based CO₂ compensation point in the absence of R _d. By definition, J ₂ can be replaced by ρ ₂ βI _inc Φ ₂, where ρ ₂ is the proportion of absorbed irradiance partitioned to PSII, β is the absorptance by leaf photosynthetic pigments, I _inc is the incident irradiance, and Φ ₂ is the quantum efficiency of PSII electron transport. Then Equation (1) becomes:

$$A={\rho }_2\beta I_{\text{inc}}{\Phi }_2\left(1-\frac{f_{\text{pseudo}}}{1-f_{\text{cyc}}}\right)\left(\frac{C_{\mathrm{c}}-{\mathrm{\Gamma }}_{*}}{C_{\mathrm{c}}+{2\mathrm{\Gamma }}_{*}}\right)/4-R_{\mathrm{d}}.$$ (2)

For NPR conditions, C _c is assumed infinite and/or Γ_* approaches zero, then Equation (2) becomes:

$$A=s(I_{\text{inc}}{\Phi }_2/4)-R_{\mathrm{d}},$$ (3)

where the lumped parameter calibration factor $s={\rho }_2\beta [1-f_{\text{pseudo}}/(1-f_{\text{cyc}}\left)\right]$. So, using data of the electron‐transport‐limited range (200–40 µmol m⁻² s⁻¹) under NPR conditions, linear regression plots of A against (I _inc Φ ₂/4) can be produced, in which Φ ₂ is based on chlorophyll fluorescence measurements. The slope of the regression gives the estimate of a calibration factor s (Table S1), and the intercept yields the estimate of R _d(Yin) under NPR conditions (Yin et al., 2009). Clearly, this approach requires that all points are within the linear range of the A versus (I _inc Φ ₂/4) curves. Here, curves of A against (I _inc Φ ₂/4) were inspected to exclude the points at high ends that might deviate from the linear pattern, especially for drought plants.

2.5. Estimation of day respiration under PR conditions

Equation (2) could also be explored to estimate R _d under PR conditions if C _c is maintained constant across irradiance levels. However, it is practically difficult to control C _c because, before measurements are actually undertaken, one does not know the actual values of photosynthetic rate, stomatal conductance and mesophyll conductance required for calculating C _c. Here, the aforementioned NRH equation was used to estimate R _d (R _d(NRH)) for PR conditions. This equation is obtained by combining the well‐known FvCB model (Farquhar et al., 1980) for e⁻ transport‐limited A with the Fick's first law of diffusion for the relation between A, intercellular CO₂ partial pressure (C _i) and C _c:

$$A=0.5\{J/4-R_{\mathrm{d}}+g_{\mathrm{m}}(C_{\mathrm{i}}+2{\mathrm{\Gamma }}_{*})-\sqrt{{[J/4-R_{\mathrm{d}}+g_{\mathrm{m}}(C_{\mathrm{i}}+2{\mathrm{\Gamma }}_{*})]}^2-4g_{\mathrm{m}}[(C_{\mathrm{i}}-{\mathrm{\Gamma }}_{*})J/4-R_{\mathrm{d}}(C_{\mathrm{i}}+2{\mathrm{\Gamma }}_{*})]}\},$$ (4)

where g _m is the mesophyll conductance and J is the linear e⁻ transport rate through PSII, which can be calculated as: J = sI _inc Φ ₂ (Yin et al., 2009). The calibration factor s was adopted from the slope value of the linear regression for the Yin method from the data under NPR conditions (see above). Non‐linear curve fitting based on Equation (4) was used to estimate g _m in a previous study (Yin & Struik, 2009). Here, this NRH equation for A was introduced to simultaneously estimate R _d(NRH) under PR conditions. The data used in the NRH method was within the same range of light levels as used in the Yin method.

There are two caveats. First, values of R _d(NRH) estimated by Equation (4) are based on the assumption that g _m is constant within the range of data used. Many studies, especially those using chlorophyll fluorescence‐based methods, have shown that g _m may vary with CO₂ and irradiance levels (e.g., Flexas et al., ²⁰⁰⁷; Ma et al., ²⁰²¹; Stangl et al., ²⁰¹⁹; Yin et al., ²⁰⁰⁹). However, whether g _m is constant or variable is still under debate. Besides using Equation (4), we also tested a form of the NRH equation of Yin et al. (2009) that accounts for the variable g _m, to assess if the g _m mode assumed has any influence on the estimation of R _d. This form of the NRH model associated with the variable g _m is given in Figure S1.

Second, regardless of the constant or variable g _m assumption, the use of the NRH method combined with J = sI _inc Φ ₂ requires that the calibration factor s is obtained from strictly NPR conditions. There is no guarantee that this could be the case under extreme conditions (especially when high temperature is combined with drought stress), as the stomatal conductance and g _m are so low under such conditions that Γ_*/C _c could not be maintained at the required low level to achieve NPR conditions even when C _a was set at 1000 ppm. In other words, if an NPR condition cannot be ensured, the obtained s is not equal to ρ ₂ β[1 − f _pseudo/(1 − f _cyc)], but to ρ ₂ β[1 − f _pseudo/(1 − f _cyc)][(C _c − Γ_*)/(C _c + 2Γ_*)]; thus, s would be lowered by a factor (C _c − Γ_*)/(C _c + 2Γ_*). We introduced dummy variables to Equation (4) to avoid the confounding effect of measuring under conditions that were not truly NPR and examined whether or not the calibration factor s was underestimated. The following identity was verified:

$$R_{\mathrm{d}}=Z_1{R}_{\mathrm{d},\text{NPR}}+Z_2{R}_{\mathrm{d},\text{PR}},$$ (5)

where R _d,NPR and R _d,PR are leaf day respiration under NPR and PR conditions, respectively, and Z ₁ and Z ₂ are dummy variables, which were set in such a way that Z ₁ = 1 and Z ₂ = 0 correspond to the NPR condition and Z ₁ = 0 and Z ₂ = 1 correspond to the PR condition. Such a procedure allows to simultaneously estimate the common parameters (s and g _m) as well as the different parameters (R _d,NPR and R _d,PR) from fitting to the combined data obtained under both NPR and PR conditions. This procedure for estimating the common parameters is equivalent to the method of simultaneous fitting of the calibration factor and g _m (e.g., Pons et al., ²⁰⁰⁹) if one is not sure whether an NPR state is reached.

2.6. Comparing the Kok, Yin and NRH methods

As stated earlier, like the Laisk method, the Kok method is a popular method to estimate R _d indirectly (e.g., still used recently by Way et al., ²⁰¹⁹). As our data also allow implementation of the Kok method, here we compare the NRH, Yin and Kok methods in estimating R _d. As the Kok method has been applied to estimate R _d (R _d(Kok)) under both PR and NPR conditions (e.g., Tcherkez et al., ^2017a), for the comparative purpose we also apply the Yin method to the PR conditions by fitting a linear regression to data at the same range of irradiances (values of the slope factor of the A vs. (I _inc Φ ₂/4) linear regression under PR conditions (s′) were listed in Table S2). It is worthy to note that Equation (4), upon which the NRH method is based to estimate R _d, can only be applied to the PR conditions.

2.7. The temperature responses of parameters

The thermal response of R _dk and R _d was described by the Arrhenius equation normalized with respect to its value at 25°C:

$$X=X_{25}{e}^{\left(\frac{1}{298}-\frac{1}{273+T_{\text{leaf}}}\right)\frac{E_{\mathit{X}}}{R}},$$ (6)

where X ₂₅ represents the value of parameters estimated at 25°C (R _d25 and R _dk25), E _X is the activation energy of relevant parameters to temperature (E _Rd and E _Rdk; in kJ mol⁻¹), and R is the universal gas constant (0.008314 kJ K⁻¹ mol⁻¹).

As shown in Equation (4), applying the NRH method requires Γ_* as input, which is 0.5O ₂/S _c/o (where O ₂ is the level of oxygen and S _c/o is the relative CO₂/O₂ specificity factor for Rubisco; von Caemmerer, ²⁰¹³; Farquhar et al., ¹⁹⁸⁰). The temperature response of S _c/o can also be described by Equation (6), and equivalent parameters S _c/o25 and E _Sc/o are generally considered to be conserved among C₃ species (von Caemmerer et al., ²⁰⁰⁹). So we adopted the value of Cousins et al. (2010) for S _c/o25 (3.022 mbar μbar⁻¹), and the value of Bernacchi et al. (2002) for E _Sc/o, which mathematically equals to the negative of the activation energy for Γ_*: 24.46 kJ mol⁻¹. Then, the value of S _c/o at each T _leaf could be estimated. Sensitivity analysis showed that, unlike g _m, the estimated R _d and its temperature response varied little with variations of S _c/o25 and E _Sc/o within the physiologically relevant ranges.

2.8. Model analyses and statistics

Simple linear regressions in the Yin method were performed using the LINEST function in Microsoft Excel. Non‐linear curve‐fitting procedures in the NRH method and the Arrhenius equation were carried out using the GAUSS method in PROC NLIN of SAS (SAS Institute Inc.). The SAS codes can be obtained upon request to the corresponding author. All regression fitting was first performed for each replicate to apply analysis of variances (ANOVA) to test significance (see below) between treatments. The ANOVA was used to analyse the main effects of water regimes and genotypes or growth temperatures, and their interactions on the Arrhenius equation parameters of R _d. Post hoc testing of each treatment was done by using Tukey's honest significance test. These analyses were conducted using R version 4.1.2 (R Foundation). The ANOVA showed little significant difference between replicates in fitted parameter values, and thus, we used pooled data of individual replicates to obtain parameter estimates to be presented as the final treatment‐specific values.

3. RESULTS

3.1. Pretesting of the two caveats

We compared the values of R _d(NRH) estimated for PR conditions by assuming constant or variable g _m modes and found that R _d(NRH) estimated by the variable g _m mode was only 3.3% lower than that estimated by the constant g _m model (Figure S1). Given that Equation (4) under the constant g _m mode is simpler than the equivalent equation under the variable g _m mode and g _m estimated from Equation (4) is easier to interpret than the parameter estimated from the variable g _m mode, only the results generated by the constant g _m mode were used for further analyses.

The approach combining the NRH method (Equation 4) and dummy variables was used to examine if there was a confounding effect of any underestimation of the calibration factor s on R _d(NRH) for PR conditions. Using methods either with or without dummy variables, we found that overfitting occurred for a few T _leaf values of drought plants due to the variability of data, which resulted in the failure of estimating g _m. For such cases, nevertheless, the estimations of s and R _d(NRH) were still reasonable. The results of each parameter (calibration factor s, g _m, R _d(NRH) for PR and R _d(Yin) for NPR conditions) from approaches with and without dummy variables were very similar (Figure S2). On average, s was underestimated only by ca. 1% (Figure S2B), suggesting that the gas mixture we used (2% O₂ combined with 1000 ppm C _a) for estimating the calibration factor did allow to reach a nearly NPR state, even for extreme conditions (drought combined with high temperatures). So, for the sake of simplicity, we only present and discuss the results obtained from the method without using the dummy variables.

3.2. Comparisons of R _d and its inhibition by light estimated by different methods

R _d(Kok) was generally lower than R _d(Yin) under both NPR (18.2% lower) and PR (13.9% lower) conditions (Figure 1a,b). Under PR conditions, values of R _d(Kok) were only 66.5% of R _d(NRH), and those of R _d(Yin) were 79.6% of R _d(NRH) (Figure 1c,d). These findings suggest that the Kok method underestimated R _d substantially, and the Yin method, if also applied to PR conditions, also underestimated R _d, but less so than the Kok method. Therefore, the light inhibition of leaf respiration, estimated as (1 − R _d /R _dk), depending on the methods used for estimating R _d. Under NPR conditions, the estimated average light inhibition by the Kok method was 18.7%, whereas that by the Yin method was only 1.4% (Figure 2a,b). Under PR conditions, the estimated average light inhibition by the Kok, Yin and NRH methods was 40.3%, 28.5% and 10.1%, respectively (Figure 2c–e), indicating that light inhibition was stronger under PR than NPR conditions. In general, values of R _dk and R _d under NPR conditions were greater than under PR conditions (Figure 3), which was in agreement with results from a previous report where both R _dk and R _d were higher at 2% O₂ than at 21% O₂ in mature leaves (Buckley et al., 2017).

Figure 1.

Correlations between leaf day respiration estimated by the Kok method (R _d(Kok)) and estimated by the Yin method (R _d(Yin)) under nonphotorespiratory (NPR; a) or photorespiratory (PR; b) conditions, and correlations between R _d(Kok) (c) or R _d(Yin) (d) and leaf day respiration estimated by the nonrectangular hyperbolic (NRH) method (R _d(NRH)) under PR conditions, across wheat genotypes Thésée and Récital, growth temperature and water treatments, and leaf temperatures during measurements. Note that the Yin method suits to estimate R _d for the NPR conditions only; as stated in the text, it is applied also to PR conditions here merely for the comparison purpose. The dashed diagonal represents the 1:1 relationship and the solid line represents fit to the data with the equation shown in each panel. The bold number (%) above the equation in each panel is the average of y‐axis relative to x‐axis values.

Figure 2.

Correlations between leaf day respiration estimated by the Kok (R _d(Kok)), Yin (R _d(Yin)) or nonrectangular hyperbolic (NRH) (R _d(NRH)) method and leaf respiration in the dark (R _dk) under nonphotorespiratory (NPR; a, b) and photorespiratory (PR; c–e) conditions, across wheat genotypes Thésée and Récital, growth temperature and water treatments, and leaf temperatures during measurements. Note that the Yin method suits to estimate R _d for the NPR conditions only; as stated in the text, it is applied also to PR conditions here merely for the comparison purpose. The dashed diagonal represents the 1:1 relationship and the solid line represents fit to the data with the equation shown in each panel. The bold number (%) above the equation in each panel is the average of y‐axis relative to x‐axis values.

Figure 3.

Correlations between leaf respiration in the dark (R _dk) under nonphotorespiratory (NPR) conditions versus R _dk under photorespiratory (PR) conditions (a), and correlations between leaf day respiration estimated by the Yin method (R _d(Yin)) under NPR conditions versus R _d(Yin) under PR conditions (b) or that estimated by the nonrectangular hyperbolic (NRH) method (R _d(NRH)) under PR conditions (c), across wheat genotypes Thésée and Récital, growth temperature and water treatments, and leaf temperatures during measurements. Note that the Yin method suits to estimate R _d for the NPR conditions only; as stated in the text, it is applied also to PR conditions here merely for the comparison purpose. The dashed diagonal represents the 1:1 relationship and the solid line represents fit to the data with the equation shown in each panel. The bold number (%) above the equation in each panel is the average of y‐axis relative to x‐axis values.

The obtained thermal responses of R _dk, R _d(Kok) and R _d(Yin) under PR and NPR conditions as well as R _d(NRH) under PR conditions were similar (Figures S3, S4 and S5 and Tables S3 and S4). Hereafter, we only use R _d(Yin) for NPR conditions and R _d(NRH) for PR conditions for further analyses because our study focuses on R _d and theoretically the Yin method and the NRH method work best to estimate R _d for NPR and PR conditions, respectively.

3.3. Impact of water regimes on R _d and its response to leaf temperature in two wheat genotypes

In EXP2019, R _d was estimated across two genotypes and two water regimes. As expected, for both NPR and PR conditions, the estimated values of R _d increased with rising T _leaf across genotypes and water regimes (Figure 4), as was widely found in previous studies (Atkin et al., 2006; Crous et al., ²⁰¹¹; Way et al., ²⁰¹⁹; Yin et al., ²⁰¹⁴). The temperature response of R _d did not vary much between Thésée and Récital, but significantly differed between WW and DS conditions. Generally, R _d was suppressed by drought stress. In agreement with the results of Crous et al. (2012), this suppression was slight or not significant at lower T _leaf (15 and 20°C), but became increasingly pronounced with increasing T _leaf, reaching up to 76.3% and 71.5% under NPR conditions and up to 74.2% and 64.7% under PR conditions in Thésée and Récital, respectively, for T _leaf = 40°C (Figure 4).

Figure 4.

Thermal responses of day respiration estimated by the Yin method (R _d(Yin)) for nonphotorespiratory (NPR) conditions (a) or by the nonrectangular hyperbolic (NRH) method (R _d(NRH) for photorespiratory (PR) conditions (b) in two wheat genotypes (Thésée and Récital) under well‐watered (WW) and drought‐stressed (DS) conditions in EXP2019. The filled points and solid lines represent the WW plants, and the open points and dashed lines represent the DS plants. Black symbols and lines refer to Thésée, while grey symbols and lines refer to Récital. Lines are the Arrhenius equation fitted to the data. Error bars indicate the standard error of the estimates (n = 4).

The temperature response of R _d was well described by the Arrhenius equation, although the estimated R _d deviated more under drought conditions as a result of higher variabilities of data among replicated plants. R _d and its response to T _leaf showed appreciable acclimation to drought. Under drought stress, R _d25 was apparently reduced, with the reduction ranging from 44.1% to 55.2% across genotypes and between PR and NPR conditions (Figure 4 and Table 1). The values of E _Rd under well‐watered conditions were similar, ranging from 58.87 to 68.64 kJ mol⁻¹ in the two genotypes under NPR and PR conditions (Table 1), which agreed with the estimate (64.18 kJ mol⁻¹) by Yin et al. (2014) for tomato. The estimated E _Rd was notably affected by water treatments. Under drought conditions the values of E _Rd decreased to less than 42 kJ mol⁻¹ (Table 1), reflecting that R _d was less sensitive to increasing temperature under water‐deficit conditions than in well‐watered conditions (Figure 4).

Table 1.

Values of modelled leaf day respiration at 25°C (R _d25) and activation energy for leaf day respiration ($E_{R_{\mathrm{d}}}$) estimated by the Arrhenius equation for two genotypes of wheat (Thésée and Récital) under WW and DS conditions in EXP2019.

	Treatment
Condition	Genotype	Water regime	R d25 (µmol m−2 s−1)	$E_{R_{\mathrm{d}}}$(kJ mol−1)	r 2
NPR conditions	Thésée	WW	1.45 (0.09)a	62.29 (3.72)a	0.991
(R d(Yin))		DS	0.81 (0.08)b	23.53 (7.56)b	0.746
	Récital	WW	1.25 (0.06)a	68.64 (3.05)a	0.996
		DS	0.56 (0.15)b	41.18 (17.74)ab	0.615
	p value fromANOVA
	Genotype		0.140	0.093
	Water regime		<0.0001	0.0003
	Genotype × Water regime		0.823	0.362
PR conditions	Thésée	WW	1.27 (0.04)a	58.87 (2.00)a	0.997
(R d(NRH))		DS	0.58 (0.15)b	40.90 (16.93)b	0.716
	Récital	WW	1.04 (0.05)a	60.67 (2.72)a	0.995
		DS	0.52 (0.10)b	40.94 (12.87)b	0.765
	p value from ANOVA
	Genotype		0.021	0.702
	Water regime		<0.0001	0.0001
	Genotype × Water regime		0.176	0.630

Note: Data of leaf day respiration used for fitting the Arrhenius equation was estimated by either the Yin method (R _d(Yin)) under NPR conditions or the NRH method (R _d(NRH)) under PR conditions. Standard errors of the estimates are given within parentheses. Different letters represent statistical differences among treatments based on post hoc testing (p < 0.05, Tukey's honest significance test).

Abbreviations: ANOVA, analysis of variance; DS, drought‐stressed; NPR, nonphotorespiratory; PR, photorespiratory; WW, well‐watered.

3.4. Impact of the combined water and growth temperature regimes on R _d and its temperature response

In EXP2020, treatments were designed to investigate the interactive effect of growth temperature and water regime on R _d. The response of R _d to T _leaf was described by the Arrhenius equation in each treatment, although it did not fit well the data for DS plants grown at LT (r ² = 0.329 and 0.395 for NPR and PR conditions, respectively).

Under WW conditions, R _d estimated at each respective T _leaf was rather consistently lower in plants grown at HT and MT than in plants grown at LT in both NPR and PR conditions (Figure 5a,b). This was also reflected in the higher estimation of R _d25 in LT plants under WW conditions (Table 2). However, the estimated E _Rd for WW plants herein was found to be lowest at LT and highest at HT (Table 2), although the difference was not significant, especially at PR conditions. Our results are contradictory to the lower sensitivity of R _d to T _leaf at warmer growth temperature reported previously (Atkin et al., 2005 and references therein).

Figure 5.

Thermal responses of day respiration estimated by the Yin method (R _d(Yin)) for nonphotorespiratory (NPR) conditions (left panels) or by the nonrectangular hyperbolic (NRH) method (R _d(NRH) for photorespiratory (PR) conditions (right panels) in winter wheat Thésée grown at three growth temperatures (HT: high temperature; MT: medium temperature; LT: low temperature) under well‐watered (WW; upper panels) and drought‐stressed (DS; lower panels) conditions in EXP2020. The filled points and solid lines represent the well‐watered plants, and the open points and dashed lines represent the drought‐stressed plants. Lines are the Arrhenius equation fitted to the data. Error bars indicate the standard error of the estimates (n = 4).

Table 2.

Values of modelled leaf day respiration at 25°C (R _d25) and activation energy for leaf day respiration ($E_{R_{\mathrm{d}}}$) estimated by the Arrhenius equation for wheat Thésée grown at three growth temperatures (HT, MT and LT) under WW and DS conditions in EXP2020.

	Treatment
Condition	Growth temperature	Water regime	R d25 (µmol m−2 s−1)	$E_{R_{\mathrm{d}}}$(kJ mol−1)	r 2
NPR conditions	HT	WW	1.26 (0.15)b	70.69 (7.15)a	0.978
(R d(Yin))		DS	0.78 (0.13)c	28.56 (11.86)bc	0.672
	MT	WW	1.36 (0.16)ab	64.58 (7.44)ab	0.969
		DS	1.13 (0.11)c	26.96 (7.27)bc	0.816
	LT	WW	1.60 (0.18)a	58.00 (10.35)ab	0.942
		DS	1.16 (0.08)c	8.10 (6.09)c	0.329
	p value from ANOVA
	Growth temperature		0.0071	0.166
	Water regime		<0.0001	<0.0001
	Growth temperature × Water regime		0.106	0.902
PR conditions	HT	WW	1.22 (0.14)ab	60.90 (7.39)a	0.967
(R d(NRH))		DS	0.51 (0.08)c	47.10 (9.86)a	0.902
	MT	WW	1.27 (0.03)ab	59.17 (1.50)a	0.998
		DS	0.95 (0.14)bc	36.09 (10.35)ab	0.816
	LT	WW	1.52 (0.14)a	55.83 (8.39)a	0.944
		DS	0.78 (0.07)c	12.93 (8.55)b	0.395
	p value from ANOVA
	Growth temperature		0.041	0.014
	Water regime		<0.0001	<0.0001
	Growth temperature × Water regime		0.068	0.024

Note: Data of leaf day respiration used for fitting the Arrhenius equation was estimated by either the Yin method (R _d(Yin)) under nonphotorespiratory (NPR) conditions or the NRH method (R _d(NRH)) under photorespiratory (PR) conditions. Standard errors of the estimates are given within parentheses. Different letters represent statistical differences among treatments based on post hoc testing (p < 0.05, Tukey's honest significance test).

Abbreviations: ANOVA, analysis of variance; DS, drought‐stressed; MT,  medium temperature; HT, high temperature;  LT, low temperature; NPR, nonphotorespiratory; PR, photorespiratory; WW, well‐watered.

Again, drought stress reduced R _d across various growth environments and between PR and NPR conditions and this reduction was more pronounced at higher T _leaf (Figure 5), which led to lower sensitivity to T _leaf, and, thus, lower E _Rd in DS plants than in WW plants (Table 2). However, this drought‐induced reduction of R _d differed among various growth regimes. DS plants grown at LT maintained relatively higher R _d at lower T _leaf (below 25°C) than HT‐ and MT‐grown plants, while at higher T _leaf (above 30°C) higher R _d values were observed in DS plants grown at MT as compared with those grown at HT and LT (Figure 5c,d). This resulted in a downward shift in the temperature response curve and a lower estimated R _d25 (0.78 and 0.51 µmol m⁻² s⁻¹ for NPR and PR, respectively) in DS plants grown at HT, and a nearly horizontal curve with an extremely low estimate of E _Rd (8.10 and 12.93 kJ mol⁻¹ for NPR and PR, respectively) in DS plants grown at LT, as compared to those grown at HT and MT (Table 2).

4. DISCUSSION

The present study investigated the impacts of long‐term drought and growth temperature treatments on the short‐term thermal response of R _d in two winter wheat genotypes (Thésée and Récital). While the two genotypes are known to have different drought tolerance in yield performance, little difference in R _d was observed between them under the two contrasting water regimes in EXP2019. Thus, we only used Thésée in EXP2020.

4.1. Impact of growth temperature on R _d under well‐watered conditions

Thermal acclimation of respiration is usually assessed by how changes with growth temperature in the response of respiration rate to measurement T _leaf sustained over time (Crous et al., 2011). In many previous studies, plants acclimated to a sustained warmer climate by reducing their respiration rate at a given measurement temperature (e.g., Crous et al., ²⁰¹¹). Moreover, the extent of thermal acclimation of respiration is often lower in pre‐existing, fully expanded leaves that are shifted to a new growth temperature than in leaves that develop under various growth temperatures (Atkin & Tjoelker, 2003). Here, although the plants were subjected to the growth temperature treatment before flag leaves emerged, our results showed that under WW conditions rates of R _d at any given T _leaf were nearly identical for plants grown at HT and MT (i.e., no thermal acclimation), regardless of estimation methods or PR conditions, but R _d slightly acclimated to the LT with an upward shift of the temperature response curve and a higher R _d25 (Figure 5a,b and Table 2). This was consistent with findings in previous studies across various R _d estimation methods and species, indicating that little or only partial thermal acclimation of R _d is more common (Atkin et al., 2006; Way et al., ²⁰¹⁹).

Plant respiration is sensitive to short‐term changes in temperature, while the effect of growth temperature on the thermal sensitivity of respiration varies among species (Atkin et al., 2005). Previous works have reported that the sensitivity of respiration to T _leaf declines with increasing growth temperature, as a result of the acclimation of respiration to a warmer temperature (e.g., Cai et al., ²⁰²⁰). In contrast, we found that the estimates of $E_{R_{\mathrm{d}}}$were lower in LT plants than in MT and HT plants under WW conditions (Table 2), implying that plants grown at lower temperatures are likely to be less sensitive to rising T _leaf than those grown at warmer temperatures. Coincidentally, Crous et al. (2017) also observed a higher value of$E_{R_{\mathrm{d}}}$ in a fast‐growing tree species (Eucalyptus globulus) grown in a warmer climate, although the mechanisms behind this remain unclear. The possible explanation for this upward trend of sensitivity to T _leaf in warm climates could be linked to the higher leaf N content on a mass basis at higher growth temperatures (Figure S6E,F); plant N status is highly associated with factors such as enzyme capacity, substrate supply and respiratory products that may affect the metabolic activities in plant tissues (O'Leary et al., ²⁰¹⁹ and references therein).

4.2. Impact of drought stress on R _d and its interaction with growth temperature

The response of leaf respiration to limited water availability seems to be equivocal and elusive as decreased (Ayub et al., 2011), unaffected (Gimeno et al., 2010) and even increased (Gauthier et al., 2014) leaf respiration rate under water deficit have been reported. The various responses could be linked to differences in species used, the severity and duration of soil dehydration and/or other environmental factors, for example, temperature (Flexas et al., 2005). In our study, inhibition by a severe drought of R _d was slight or not significant under lower T _leaf across all treatments (Figures 4 and 5), consistent with Gauthier et al. (2014), who reported that the resilience of leaf respiration in the dark (R _dk) was observed in low to moderate ranges of T _leaf. This finding might explain the unaffected and even slightly increased leaf respiration under drought treatment in some cases where R _d or R _dk was measured at a set of common or lower temperatures rather than a short‐term change of measurement temperature (Gimeno et al., 2010; Sperlich et al., ²⁰¹⁶). At T _leaf > 20°C, however, our results showed that the inhibition of R _d by drought was substantial (Figures 4 and 5), which was in agreement with previous studies (Crous et al., ²⁰¹²; Dahal & Vanlerberghe, ²⁰¹⁷; Haupt‐Herting et al., ²⁰⁰¹). As a result, there was low sensitivity of R _d to T _leaf in drought plants (Figures 4 and 5). However, several previous studies reported contradictory results: drought may exacerbate R _dk at moderate or higher temperatures (Slot et al., 2008; Zagdańska, ¹⁹⁹⁵) and even lead to a 'respiratory burst' in an extremely high (>40°C) short‐term measurement temperature range (Gauthier et al., 2014), although the underlying mechanisms remain speculative. In some of these studies (e.g., Gauthier et al., ²⁰¹⁴), plants experienced two phases of drought with a rewatering treatment in between, and the temperature response of leaf respiration was measured at the end of the second period of drought. This means that the increased leaf respiration under drought stress could be due to drought priming. Moreover, in these early studies, drought treatment was applied at the seedling or sapling stage (Bartoli et al., 2005; Gauthier et al., ²⁰¹⁴; Zagdańska, ¹⁹⁹⁵), instead of at the postanthesis stage in which assimilated carbohydrates and nitrogenous compounds, those related to substrate supply and respiratory capacity (Tjoelker et al., 1999), are being translocated from source (leaves) to sink (grains) (Shao et al., 2021), and energy demand for sucrose synthesis and/or phloem loading is in decline (Atkin & Macherel, 2009). The above may explain the difference between their results and our experiments.

Furthermore, our results showed that indeed there was an interactive impact of drought stress and growth temperature on the response of R _d to T _leaf, with a much lower$E_{R_{\mathrm{d}}}$ value in DS plants grown at LT and a lower R _d25 in DS plants grown at HT than in WW plants (Table 2). A low$E_{R_{\mathrm{d}}}$ in DS plants grown at LT means a reduced R _d at higher T _leaf. When T _leaf was <25°C, LT treatment appeared to alleviate the negative impact of drought on R _d rates as compared with HT and MT treatments, whereas this was not the case in the higher T _leaf range (>25°C) where DS plants grown at MT exhibited the highest rates of R _d (Figure 5c,d). The possible explanation for the mitigated drought effect on LT‐grown plants could be linked to the upregulated alternative oxidase (Dahal & Vanlerberghe, 2017; Searle et al., ²⁰¹¹) and mitochondrial uncoupling proteins (Barreto et al., 2017; Nantes et al., ¹⁹⁹⁹).

4.3. The theoretical basis of the NRH method, compared with the Kok method and the Yin method

Compared with the Kok method, the Yin method exploits the additional information from chlorophyll fluorescence data. Our results showed that R _d(Kok) was lower than R _d(Yin) under both NPR and PR conditions (Figure 1a,b), which was in line with previous studies (e.g., Yin et al., ²⁰¹¹). The lower estimates of R _d(Kok) were due to the neglect of a decrease in Φ ₂ with increasing light intensity, which occurs even with limiting light levels. Theoretically, the Yin method works under NPR conditions (Berghuijs et al., 2019; Yin et al., ²⁰¹¹). It works for PR conditions only if C _c is maintained constant across light levels, which is technically difficult to achieve in measurements because g _m is unknown beforehand. Berghuijs et al. (2019) pointed out that the linear regression as used in the Yin method will underestimate R _d if applied to PR conditions; this is also confirmed by our data (Figure 1d). The theoretical underpinning is that under PR conditions C _c is not constant but increases with decreasing light intensity, leading to an apparent Kok effect (Farquhar & Busch, 2017; Yin et al., ^2020a). The increase of C _c with decreasing light intensity is the consequence that the high flux of leaf respiration, relative to photosynthesis, at low irradiances can result in an accumulation of CO₂ if the respired CO₂ cannot escape totally to the atmosphere as a result of stomatal and mesophyll resistances. This high C _c means that part of the respired CO₂ can be reassimilated. The combined FvCB and g _m model, Equation (4), can generate the increase of C _c with decreasing irradiance (Farquhar & Busch, 2017), and thus, in principle, can account for the reassimilation, as shown by Yin et al. (2021). In fact, a fraction of (photo)respired CO₂ being reassimilated can be calculated from stomatal and mesophyll resistance components (Tholen et al., 2012; Yin et al., ^2020b, ²⁰²¹). Equation (4) was previously used by Yin and Struik (2009) to estimate g _m, in line with the assertion that the chlorophyll fluorescence‐based estimate of g _m relies on the reassimilation of photorespired CO₂ (Laisk et al., 2006; Yin et al., ^2020b). Here, we use Equation (4) to estimate g _m and R _d simultaneously, by exploring both gas exchange and chlorophyll fluorescence data across a range of low light intensities. The principle is in analogy to the procedure of Brooks and Farquhar (1985) that corrects for the decrease in C _i, and of Ayub et al. (2011) that further corrects for the decrease in C _c, with increasing irradiance, but with the benefit that the NRH fitting method is easier to implement. It is also in analogy to the 2‐D modelling of Berghuijs et al. (2019) that accounts for the reassimilation of (photo)respired CO₂, but with the benefit that Equation (4) is considerably simpler than the 2‐D model. In contrast, the linear regression‐based Kok and Yin methods do not account for reassimilation, and therefore, underestimate real (or gross) R _d, if applied to PR conditions. They can be reliably used for NPR conditions because reassimilation of respired CO₂, if any, contributes little to total assimilation under NPR conditions created by high ambient [CO₂] or/and low [O₂].

Based on the above discussion, differences in R _d values estimated in our study could be related to (1) PR versus NPR conditions, (2) reassimilation, and (3) assumptions behind the methodology. If these methods are appropriately applied (i.e., the NRH method applied to PR conditions and the Yin method to NPR conditions), the estimates refer to gross respiration. So, when these two methods are compared, the difference refers to the difference of R _d between PR and NPR conditions. However, when the Yin method is applied to PR conditions, then the underestimation by the Yin method, relative to the NRH method, refers to the difference caused by reassimilation. The Kok method always underestimates R _d, regardless of PR and NPR conditions, so, the difference in its estimated R _d from other methods is due to the methodology.

The estimation of g _m can be very sensitive to measurement errors (Yin & Struik, 2009). Here, we failed to estimate g _m in some cases due to the variation in data from stressed plants (Table S5). Nevertheless, the estimates of R _d could still be reliable as reflected by the low standard error of estimates of R _d. The NRH method has the following advantages. First, this method can estimate R _d under PR conditions by implicitly considering the reassimilation or variation in C _c with irradiance as it corrects for the error of the linear regression methods assuming a constant C _c under PR conditions. Second, this method only requires gas exchange and chlorophyll fluorescence data, but does not require sophisticated and expensive isotopic devices as required by direct R _d measuring methods (e.g., Gong et al., ²⁰¹⁵; Haupt‐Herting et al., ²⁰⁰¹) or leaf anatomical data as required by the 2‐D modelling method of Berghuijs et al. (2019). Third, it provides additional estimates for parameters g _m, which could be recognized as indicators of physiological processes in response to environmental variables.

Our results showed that light inhibition of leaf respiration was higher under PR conditions than under NPR conditions (Figure 2), consistent with the result of Yin et al. (2020a) that apparent light inhibition of respiration and thus the Kok effect is not obvious under low O₂ or high CO₂ conditions or their combinations. Moreover, our data also showed that the underestimation of R _d by the Kok and Yin methods under PR conditions may lead to an overestimation of light inhibition of leaf respiration (Figure 2c,d). Berghuijs et al. (2019) suggested that R _d estimated by the Yin and the Kok methods at NPR conditions cannot represent the real R _d at PR conditions. Here, our results showed that values of R _d for NPR conditions were generally higher than those for PR conditions (Figure 3b,c), which was in agreement with the result of Yin et al. (2020a) that real light inhibition on respiration increases with increasing amount of photorespiration. Again, the reason for the greater light suppression of R _d under PR conditions remains to be elucidated.

5. CONCLUDING REMARKS

In contrast to the plethora of studies that have explored the responses of R _dk to contrasting environments, we assessed the extent to which R _d of wheat leaves acclimated to drought and growth temperature. We proved a simple method that can estimate R _d for PR conditions by using gas exchange and chlorophyll fluorescence data. It was demonstrated that R _d and its temperature response for both PR and NPR conditions acclimated more to drought than to growth temperature. Understanding this acclimation of R _d is needed to support the modelling of R _d, and thus of crop productivity and of carbon cycling in agricultural ecosystems, under future climate change.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

Supporting information

Supporting information.

Fang, L. , Yin, X. , van der Putten, P.E.L. , Martre, P. & Struik, P.C. (2022) Drought exerts a greater influence than growth temperature on the temperature response of leaf day respiration in wheat (Triticum aestivum). Plant, Cell & Environment, 45, 2062–2077. 10.1111/pce.14324

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.