Elevated CO₂ alleviates the negative impact of heat stress on wheat physiology but not on grain yield

Elevated CO₂ protects wheat photosynthesis from heat stress damage via increased electron transport and facilitates recovery of photosynthesis and biomass but not the yield due to heat-induced grain abortion.

Abstract

Hot days are becoming hotter and more frequent, threatening wheat yields worldwide. Developing wheat varieties ready for future climates calls for improved understanding of how elevated CO₂ (eCO₂) and heat stress (HS) interactively impact wheat yields. We grew a modern, high-yielding wheat cultivar (Scout) at ambient CO₂ (aCO₂, 419 μl l ⁻¹) or eCO₂ (654 μl l⁻¹) in a glasshouse maintained at 22/15 °C (day/night). Half of the plants were exposed to HS (40/24 °C) for 5 d at anthesis. In non-HS plants, eCO₂ enhanced (+36%) CO₂ assimilation rates (A_sat) measured at growth CO₂ despite down-regulation of photosynthetic capacity. HS reduced A_sat (–42%) in aCO₂- but not in eCO₂-grown plants because eCO₂ protected photosynthesis by increasing ribulose bisphosphate regeneration capacity and reducing photochemical damage under HS. eCO₂ stimulated biomass (+35%) of all plants and grain yield (+30%) of non-HS plants only. Plant biomass initially decreased following HS but recovered at maturity due to late tillering. HS equally reduced grain yield (–40%) in aCO₂- and eCO₂-grown plants due to grain abortion and reduced grain filling. While eCO₂ mitigated the negative impacts of HS at anthesis on wheat photosynthesis and biomass, grain yield was reduced by HS in both CO₂ treatments.

Introduction

Rising atmospheric CO₂ concentration is the primary cause of increasing global mean surface temperatures as well as increased frequency, duration, and intensity of heat waves. Heat stress (HS), defined as short-term temperature increases above the optimum range (Wahid et al., 2007), and other climate extremes such as droughts threaten global crop productivity, including wheat (Asseng et al., 2015). Wheat (Triticum aestivum L.) is the most widely grown crop (>218 Mha planted annually) in the world, the second most produced cereal globally (771 Mt in 2017) after maize (1134 Mt in 2017) (FAO, 2019), and a significant source of protein, providing ~20% of global calories for human consumption. Recent trends in climate and global crop production (Lobell et al., 2011) raise pertinent questions about the readiness of current crop genotypes to cope with future climate extremes, and highlight the need to evaluate the performance of current commercial, high-yielding crop genotypes under elevated CO₂ (eCO₂) and HS conditions.

Several studies have investigated the response of wheat to eCO₂ (Kimball, 1983; Hocking and Meyer, 1991; Kimball et al., 1995, 1999; Nie et al., 1995; Hunsaker et al., 1996, 2000; Miglietta et al., 1996; Osborne et al., 1998; Amthor, 2001). However, only a few studies have considered the interaction between eCO₂ and warming in wheat (Rawson, 1992; Delgado et al., 1994; Morison and Lawlor, 1999; Jauregui et al., 2015; Cai et al., 2016), and less frequently included HS (Wang et al., 2008, 2011; Shanmugam et al., 2013; Fitzgerald et al., 2016). Studies considering acute HS alone (Stone and Nicolas, 1994, 1996, 1998) or with eCO₂ focused mainly on biomass or yield and not the underlying physiological processes such as photosynthesis. Only a few studies have considered the interactive effects of eCO₂ and HS on wheat photosynthesis (Wang et al., 2008, 2011; Shanmugam et al., 2013; Macabuhay et al., 2018). These studies emphasize the need to determine the impacts of HS application at the vegetative and the important reproductive stage.

The FACE (free-air CO₂ enrichment) study by Fitzgerald et al. (2016) in wheat relied on natural heat waves during the reproductive stage and highlighted the need for controlled-environment experiments in order to carefully investigate the interactive effects of eCO₂ and HS on wheat productivity. The only FACE study with wheat by Macabuhay et al. (2018) involved controlled heat stress along with eCO₂ and concluded that eCO₂ may moderate some effects of HS on wheat grain yield, but such effects strongly depend on seasonal conditions and timing of HS. The limited number of studies highlight the gap in our understanding of how processes underlying wheat yield respond to the interactive effects of eCO₂ and HS. Such understanding is important to identify potential adaptive traits for future breeding in order to stay abreast of climate change.

HS may cause irreversible effects on plant growth and development (Wahid et al., 2007), and can inhibit both light and dark processes of photosynthesis via numerous mechanisms (Farooq et al., 2011). For example, temperatures >45 °C can damage PSII (Berry and Bjorkman, 1980; Sage and Kubien, 2007). Plants may acclimate and acquire thermal tolerance to HS by activating stress response mechanisms and expressing heat shock proteins to repair HS damage (Pan et al., 2018; Zhang et al., 2018). Acquired thermotolerance is cost intensive and compromises plant productivity (Wahid et al., 2007).

Elevated CO₂ reduces stomatal conductance and increases photosynthesis rates by stimulating carboxylation and suppressing oxygenation of Rubisco known as photorespiration (Ainsworth and Rogers, 2007; Leakey et al., 2009). Photosynthesis responds transiently to an instantaneous increase in temperature but may acclimate in response to long-term exposures (>1 d) to high temperature (Yamasaki et al., 2002). Above the thermal optimum (T_opt), high temperature reduces photosynthesis by increasing photorespiration and decreasing Rubisco activation (Eckardt and Portis, 1997). The maximal rate of RuBP carboxylation (V_cmax) responds positively to temperatures as high as 40 °C, but the maximal rate of ribulose bisphosphate (RuBP) regeneration or electron transport (J_max) generally decreases at lower temperatures in the range of 33 °C (Medlyn et al., 2002). The relative effect of eCO₂ on net photosynthesis is greater at high temperatures due to suppression of photorespiration (Long, 1991). Elevated CO₂ may also increase the T_opt of photosynthesis (Borjigidai et al., 2006; Alonso et al., 2008; Ghannoum et al., 2010). At eCO₂, the response of photosynthesis to temperature becomes increasingly limited by J_max and Rubisco activation (Sage and Kubien, 2007). Therefore, the T_opt of photosynthesis will reflect that of J_max in plants grown at eCO₂. Above T_opt, acclimation of photosynthesis to high temperatures is associated with increased electron transport and/or heat stabilty of Rubisco activase (Sage and Kubien, 2007). Hence, even though J_max decreases with short-term increases in temperature, prolonged exposure to high temperaure may trigger photosynthetic acclimation and increase J_max. Consequently, we predict that eCO₂ will increase the T_opt of photosynthesis, and mitigate negative effects of HS on photosynthesis via increased electron transport (Hypothesis 1).

The effects of HS on plant biomass and grain yield depend on the magnitude and duration of HS. HS at the vegetative stage reduces biomass and grain yield mainly by speeding up plant development and reducing the time available to capture resources, and by reducing photosynthetic rates (Lobell and Gourdji, 2012). At the flowering or anthesis stage, HS reduces grain number due to pollen abortion, while at the grain-filling stage, HS reduces grain weight by limiting assimilate translocation and shortening the grain-filling duration (Wahid et al., 2007; Farooq et al., 2011; Prasad and Djanaguiraman, 2014). Elevated CO₂ may alleviate the negative impact of HS on biomass and grain yield through stimulation of photosynthesis, improvement in plant water status due to reduced transpiration, and protection of the photosynthetic apparatus from HS damage. Furthermore, increased levels of sucrose and hexoses in plants grown at eCO₂ are associated with increased spike biomass and fertile florets (Dreccer et al., 2014) and osmotic adjustment (Wahid et al., 2007) which can improve HS tolerance (Shanmugam et al., 2013). Taken together, we hypothesize that HS applied at anthesis will negatively impact plant biomass and grain yield less in eCO₂ than in ambient CO₂ (aCO₂) (Hypothesis 2).

The primary objective of this study was to test the performance of a current wheat champion genotype under future climate extremes. The chosen wheat cultivar, Scout, is a high yielding variety with very good grain quality and contains a putative high transpiration efficiency gene which can increase water use efficiency (https://www.pacificseeds.com.au/images/Icons/Products/Wheat/SNSWVICSA/ScoutVICSA.pdf). Considering the limitations of field conditions, we undertook our study under controlled environments to unravel the physiological underpinnings of the responses to eCO₂ and HS. Consequently, we grew wheat (cultivar Scout) plants in a glasshouse at current aCO₂ and future eCO₂, and exposed half of the plants to a 5 d HS at 50% anthesis (Zadoks scale DC65). We investigated the interactive effects of eCO₂ and HS on photosynthesis, biomass, and grain yield in Scout plants.

Materials and methods

Plant culture and treatments

The experiment was conducted in a glasshouse located at the Hawkesbury campus of Western Sydney University, Richmond, New South Wales. The commercial wheat cultivar Scout, which has a putative transpiration use efficiency gene (Condon et al., 2004), was selected for the current experiment. Seeds were sterilized using 1.5% NaOCl₂ for 1 min followed by incubation in the dark at 28 °C for 48 h in Petri plates. Sprouted seeds were planted in germination trays using seed raising and cutting mix (Scotts, Osmocote^®) at ambient CO₂ (419 μl l⁻¹, day time average), temperature 22.3/14.8 °C (day/night average), relative humidity (RH; 62%, day time average), and natural light (see Supplementary Fig. S1a–e at JXB online). Day and night time averages were calculated from 10.00 h to 16.00 h and from 20.00 h to 06.00 h, respectively. Two-week-old seedlings were transplanted into individual cylindrical pots (15 cm diameter and 35 cm height) filled with sieved soil collected from the local site. Two glass house chambers were used for plant growth treatments, one with aCO₂ and the other with eCO₂. Each chamber had two bays with 50 plants in each bay. Fifty pots with one plant per pot were placed close to each other with a density of 24 plants m^–2. At the transplanting stage, pots were randomly distributed into aCO₂ and eCO₂ chambers. Transplanted plants were grown under the current aCO₂ (419 μl l⁻¹, daytime average) and eCO₂ (654 μl l⁻¹, day time average) with 62 % (day time average) RH, 22.3/14.8 °C (day/night average) growth temperature, and natural light (800 μmol m⁻² s⁻¹, average daily maximum) (Supplementary Figs S1, S2). Half of the aCO₂- and eCO₂-grown plants were exposed to a 5 d HS treatment at 50% anthesis (13 weeks after planting, WAP). HS was applied by moving plants to a separate neighbouring chamber maintained at 40/24 °C (day/night average) air temperature and 71% (daytime average) RH during the 5 d HS treatment (Supplementary Figs S1a–e, S3). Plants were well watered throughout the experiment to separate HS and water stress effects. Thrive all-purpose fertilizer (Yates) was applied monthly throughout the experiment. To minimize chamber effects, pots were randomized regularly within and among the glasshouse chambers. Ten plants per treatment were used for physiological and biomass measurements.

Temperature response of leaf gas exchange at five leaf temperatures

The response of the light-saturated CO₂ assimilation rate (A_sat) to variations in substomatal CO₂ mole fraction (C_i) was measured at five leaf temperatures (15, 20, 25, 30, and 35 °C) in both aCO₂- and eCO₂-grown plants before HS. Saturating light of 1800 μmol m⁻² s⁻¹ photosynthetic photon flux density (PPFD) was used for measurements. Six plants per treatment were used to measure one A–C_i curve (see below for details) at each temperature. Plants were transferred into a growth cabinet (Sanyo) with temperature and light control to achieve the desired leaf temperature by controlling air temperatures. Leaf temperature sequence started at 25 °C decreasing to 15 °C and then increased up to 35 °C. Dark respiration (R_d) was measured by switching the light off for 20 min at the end of each temperature curve.

Single leaf gas exchange measurements

Instantaneous steady-state leaf gas exchange measurements were performed before (9 WAP), during (13 WAP), after (13 WAP), and at the recovery stage (17 WAP) of the HS cycle using a portable open gas exchange system (LI-6400XT, LI-COR, Lincoln, NE, USA, equipped with a leaf fluorometer). Measurements were performed at a PPFD of 1800 μmol m⁻² s⁻¹ with two CO₂ concentrations (400 μl l⁻¹ and 650 μl l⁻¹) and two leaf temperatures (25 °C and 35 °C). Measuring plants at common 25 °C gives an idea about photosynthetic acclimation, while measuring plants at common 35 °C indicates the effects of HS relative to control plants. Plants were moved to a neighbouring growth chamber to achieve the desired leaf temperature.

Parameters measured were light-saturated assimilation rate (A_sat), stomatal conductance (g_s), the ratio of intercellular to ambient CO₂ (C_i/C_a), dark respiration (R_d), and maximum light use efficiency of PSII of dark- and light-adapted leaves (F_v/F_m and F_v'/F_m', respectively). These parameters were also measured in control plants before HS (13 WAP) and at the recovery stage (17 WAP) following HS. Photosynthetic down-regulation or acclimation was examined by comparing the measurements at common CO₂ (aCO₂- and eCO₂-grown plants measured at 400 μl CO₂ l⁻¹) and growth CO₂ (aCO₂-grown plants measured at 400 μl CO₂ l⁻¹ and eCO₂-grown plants measured at 650 μl CO₂ l⁻¹). R_d was measured after a 15–20 min dark adaptation period. Photosynthetic water use efficiency (PWUE), also termed intrinsic water use efficiency, was calculated as A_sat (μmol m⁻² s⁻¹)/g_s (mol m⁻² s⁻¹). The response of the A_sat to variations in C_i (A–C_i response curve) was measured at 17 WAP in eight steps of CO₂ concentrations (50, 100, 230, 330, 420, 650, 1200. and 1800 μl l⁻¹) at a leaf temperature of 25 °C. Measurements were taken around mid-day (from 10.00 h to 15.00 h) on attached fully expanded flag leaves (last leaves) of the main stems. Before each measurement, the leaf was allowed to stabilize for 10–20 min until it reached a steady state of CO₂ uptake and stomatal conductance. Ten replicate plants per treatment were measured.

Determination of Rubisco content

Following gas exchange measurements, leaf discs (0.5 cm²) were collected using a cork borer from measured flag leaves, rapidly frozen in liquid nitrogen, and stored at –80 °C until analysed. Each leaf disc was extracted in 0.8 ml of ice-cold extraction buffer [50 mM EPPS–sodium hydroxide (pH 7.8), 5 mM DTT, 5 mM magnesium chloride, 1 mM EDTA, 10 µl of protease inhibitor cocktail (Sigma), and 1% (w/v) polyvinyl polypyrrolidone] using a 2 ml Tenbroeck glass homogenizer kept on ice. The extract was centrifuged at 15 000 rpm (21 130 rcf) for 1 min and the supernatant was used for the assay of Rubisco content. Samples were incubated in activation buffer [50 mM EPPS (pH 8.0), 10 mM MgCl₂, 2 mM EDTA, 20 mM NaHCO₃] for 15 min at room temperature. Rubisco content was estimated by the irreversible binding of [¹⁴C]CABP (2-C-carboxyarabinitol 1,5-bisphosphate) to the fully carbamylated enzyme (Sharwood et al., 2008).

Growth and biomass measurements

Plants were harvested at three time points: before HS (B), after recovery from HS (R), and at the final harvest after maturity (M). At each harvest, morphological parameters were measured and the biomass was harvested separately for roots, shoots, and leaves. Samples were dried for 48 h in the oven at 60 °C immediately after harvesting. Leaf area was measured before HS and at the recovery stage of HS using a leaf area meter (LI-3100A, LI-COR). Plant height, leaf number, tiller number, and spike (grain-bearing plant organ) number were also recorded. Leaf mass per area (LMA, g m⁻²) was calculated as total leaf dry mass/total leaf area.

Mesophyll conductance and temperature response

Mesophyll conductance (g_m) was determined by concurrent gas exchange and stable carbon isotope measurements using a portable gas exchange system (LI-6400-XT, LI-COR) connected to a tunable diode laser (TDL) (TGA100, Campbell Scientific, UT, USA) for Scout grown at ambient aCO₂ partial pressures. A_sat and ¹³CO₂/¹²CO₂ carbon isotope discrimination were measured 35 d after planting at five leaf temperatures (15, 20, 25, 30, and 35 °C) and saturating light (1500 µmol quanta m⁻² s⁻¹). Leaf temperature sequence started at 25 °C decreasing to 15 °C and then increased up to 35 °C. Response of A_sat to variations in C_i was measured at each leaf temperature. R_d was measured by switching the light off for 20 min at the end of each temperature curve. Measurements were made inside a growth cabinet (Sanyo) to achieve the desired leaf temperature. The photosynthetic carbon isotope discrimination (Δ) to determine g_m was measured as follows (Evans et al., 1986):

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{\Delta }}_{}=\frac{1000\epsilon (\delta {}^{13}{C}_{sam}-\delta {}^{13}{C}_{ref})}{1000+\delta {}^{13}{C}_{sam}-\epsilon (\delta {}^{13}{C}_{sam}-\delta {}^{13}{C}_{ref})}.}\end{array}}$$ (1)

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{W}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e},\epsilon =\frac{C_{ref}}{C_{ref}-C_{sam}}.}\end{array}}$$ (2)

C _ref and C_sam are the CO₂ concentrations of dry air entering and exiting the leaf chamber, respectively, measured by the TDL. g_m was calculated using correction for ternary and second-order effects (Farquhar and Cernusak, 2012; Evans and Von Caemmerer, 2013) following the next expression:

$${\displaystyle \begin{array}{l}{\displaystyle g_m=\frac{\frac{1+t}{1-t}(b-a_i-\frac{e{R}_d}{A+R_d})\frac{A}{C_a}}{({\mathrm{\Delta }}_i-{\mathrm{\Delta }}_o-{\mathrm{\Delta }}_e-{\mathrm{\Delta }}_f)}.}\end{array}}$$ (3)

Where, Δ _i is the fractionation that would occur if the g_m were infinite in the absence of any respiratory fractionation (e=0), Δ _o is observed fractionation, and Δ _e and Δ _f are fractionation of ¹³C due to respiration and photorespiration, respectively (Evans and Von Caemmerer, 2013).

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{\Delta }}_i=\frac{1}{1-t}{a}^{\prime }\frac{1}{(1-t)}((1+t)b-a^{\prime })\frac{C_i}{C_a}.}\end{array}}$$ (4)

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{\Delta }}_e=\frac{1+t}{1-t}(\frac{e{R}_d}{(A+R_d)C_a}(C_i-\mathrm{\Gamma }*)).}\end{array}}$$ (5)

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{\Delta }}_f=\frac{1+t}{1-t}\left(f\frac{\mathrm{\Gamma }*}{C_a}\right).}\end{array}}$$ (6)

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{W}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e},t=\frac{(1+a^{\prime })E}{2{g^t}_{ac}}.}\end{array}}$$ (7)

The constants used in the model were as follows: E denotes the transpiration rate; g^t_ac is total conductance to diffusion in the boundary layer (ab=2.9‰) and in air (a=4.4‰); a′ is the combined fractionation of CO₂ across the boundary layer and stomata; net fractionation caused by RuBP and phosphoenolpyruvate (PEP) carboxylation (b=27.3‰) (Evans et al., 1986); fractionation with respect to the average CO₂ composition associated with photorespiration (f=11.6‰) (Lanigan et al., 2008); and we assumed null fractionation associated with mitochondrial respiration in light (e=0).

Statistical analysis and curve fitting

The full factorial experimental design included measurement of 10 plants per treatment for gas exchange and biomass determination. Data analyses and plotting were performed using R computer software (R Core Team, 2017). The effect of treatments and their interactions were analysed using Student’s t-test and linear modelling with ANOVA. The homogeneity of variance was tested using Levene’s test from the car package. Significance tests were performed with ANOVA and post-hoc Tukey test using the ‘glht’ function in the multcomp package designed for multiple comparisons. Other packages were also used, including (but not limited to) lubridate (for effective use of dates in plots), sciplot (for plotting), doby (for calculating means and SEs), and visreg (for plotting). The significance levels for ANOVA were, *P<0.05, **P<0.01, and ***P<0.001. Coefficient means were ranked using post-hoc Tukey test.

The Farquhar–von Caemmerer–Berry (FvCB) photosynthesis model was fit to the A–C_i response curve or chloroplastic CO₂ mole fraction (C_c), which was estimated from the g_m measurements performed in a previous experiment as described above. g_m was measured in plants grown at aCO₂ and assumed similar for plants grown at eCO₂ due to the small effect of growth CO₂ on g_m (Singsaas et al., 2004). We employed the plantecophys R package (Duursma, 2015), which uses the FvCB model to perform fits using measured g_m and R_d values along with recently reported values for K_c, energy of activation (E_a) for K_c, K_o, and Γ* in wheat (Silva-Pérez et al., 2017). Different temperatures values for K_c, K_o, and Γ*were determined using the Arrhenius equation as follows,

$${\displaystyle \begin{array}{l}{\displaystyle f(Tk)=k_{25}\cdot \exp \left[\frac{E_a\cdot (Tk-298)}{R\cdot 298\cdot Tk}\right].}\end{array}}$$ (8)

Where E_a is the activation energy (in J mol⁻¹) and k₂₅ is the value of the parameter at 25 °C. R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), and Tk is the leaf temperature in K. The activation energy term E_a describes the exponential rate of rise of enzyme activity with the increase in temperature. Fitting the FvCB model using the plantecophys package resulted in estimates of maximal Rubisco carboxylation rate (V_cmax) and maximal electron transport rate (J_max). The temperature correction parameter (T_correct) was set to False while fitting A–C_i curves using the plantecophys package. Means of coefficients were calculated using summaryBy function (in the doBy package). Means of estimated V_cmax and J_max values at five leaf temperatures were then fit by Arrhenius and peaked functions, respectively (Medlyn et al., 2002), using the non-linear least square (nls) function in R to determine energy of activation for V_cmax (E_aV) and J_max (E_aJ), and entropy (ΔSJ). Temperature responses of V_cmax and R_d means were fit using Arrhenius Equation 8. The temperature coefficient Q₁₀, a measure of the rate of change of a parameter as a consequence of increasing the temperature by 10 °C, was also determined for R_d using the following equation:

$${\displaystyle \begin{array}{l}{\displaystyle R_d=R_{d}25\cdot {Q_{10}}^{[(T-25)/10]}}\end{array}}$$ (9)

A peaked function (Harley et al., 1992) derived Arrhenius function was used to fit the temperature dependence of J_max, and is given by the following equation:

$${\displaystyle \begin{array}{l}{\displaystyle f(Tk)=k_{25}\cdot \exp \left[\frac{H_a\cdot (Tk-298)}{R\cdot 298\cdot Tk}\right]\left[\frac{1+\exp \left(\frac{298\cdot \mathrm{\Delta }S-H_d}{298\cdot R}\right)}{1+\exp \left(\frac{Tk\cdot \mathrm{\Delta }S-H_d}{Tk\cdot R}\right)}\right]}\end{array}}$$ (10)

Where H_a is the activation energy and k₂₅ is the J_max value at 25 °C, H_d is the deactivation energy, and S is the entropy term. H_d and ΔS together describe the rate of decrease in the function above the optimum. H_d was set to constant 200 kJ mol⁻¹ to avoid overparameterization. The temperature optimum of J_max was derived from Equation 2 (Medlyn et al., 2002) and written as follows:

$${\displaystyle \begin{array}{l}{\displaystyle T_{opt}=\frac{H_d}{\mathrm{\Delta }S-R\cdot \ln \left[\frac{E_a}{(H_d-E_a)}\right]}}\end{array}}$$ (11)

The temperature response of A_sat was fit using a simple parabola equation (Crous et al., 2013) to determine the temperature optimum of photosynthesis:

$${\displaystyle \begin{array}{l}{\displaystyle A_{sat}=A_{opt}-b\cdot {(T-T_{opt})}^2}\end{array}}$$ (12)

where T is leaf temperature during measurement of A_sat, T_opt represents the temperature optimum, and A_opt is the corresponding A_sat at T_opt. Steady-state gas exchange parameters g_m, g_s, C_i, and the J_max to V_cmax ratio were fit using the nls function with the polynomial equation:

$${\displaystyle \begin{array}{l}{\displaystyle y+A+Bx+C{x}^2}\end{array}}$$ (13)

Results

In non-HS plants, photosynthetic acclimation to eCO₂ was stronger at the vegetative stage, while eCO₂ stimulated photosynthesis at 25 °C at all growth stages

To assess photosynthetic acclimation due to eCO₂, non-HS plants were measured at the peak growth period (13 WAP) and after 50% anthesis (17 WAP). At 13 WAP, growth under eCO₂ reduced A_sat measured at common CO₂ at both 25 °C (–12%, P=0.004) (Fig. 1a; Table 1; and Supplementary Table S1) and 35 °C (–13.3%, P=0.01) (Table 1; Supplementary Table S1). At 13 WAP, eCO₂ enhanced A_sat of non-HS plants measured at growth CO₂, at both 25 °C (+25%, P=0.003) (Fig. 1a; Table 1; Supplementary Table S1) and 35 °C (+39%, P<0.001) (Table 1; Supplementary Table S1).

Fig. 1.

Photosynthetic response of wheat cultivar Scout to eCO₂ measured 13 and 17 weeks after planting (WAP) at 25 °C leaf temperature and two CO₂ concentrations. Bar plot of means for light-saturated CO₂ assimilation rate (a, b, and c) and stomatal conductance (d, e, and f) calculated using two-way ANOVA. The error bars indicate the SE of the mean (n=9–10). Ambient and elevated CO₂-grown plants are depicted in black and grey, respectively. Grouping is based on measurement CO₂ (400 μl l⁻¹ or 650 μl l⁻¹). Bars sharing the same letter in the individual panels are not significantly different according to Tukey’s HSD test at the 5% level. Statistical significance levels (t-test) for eCO₂ effect are shown: *P<0.05; **P<0.01: ***P<0.001.

Table 1.

Summary of statistics for gas exchange parameters

Parameter (mean plant−1)	Measurement		13 WAP	17 WAP
	Temperature °C	CO2 (μl l−1)	Main effects	Main effects		Interaction
			CO2	CO2	HS	CO2×HS
A (µmol m−2 s−1)	25	400	**	*	*	**
		650	**	**	*	**
	35	400	**	NS	NS	*
		650	**	NS	NS	*
R d (µmol m−2 s−1)	25	400	NS	NS	NS	NS
	35	400	*
g s (mol m−2 s−1)	25	400	NS	*	*	NS
		650	NS	**	**	*
	35	400	NS	NS	NS	NS
		650	NS	NS	NS	NS
PWUE (A/gs)	25	400	NS	NS	NS	NS
		650	NS	NS	NS	NS
	35	400	NS	NS	*	*
		650	NS	NS	NS	NS
F v/Fm	25	400	NS	*	NS	NS
F v'/Fm'	25	400	NS	**	**	*
A (µmol m−2 s−1)	25	Growth CO2	**	***	NS	**
	35		***	**	NS	*
R d (µmol m−2s−1)	25		***
	35		*
g s (mol m−2 s−1)	25		NS	***	**	*
	35		NS	NS	NS	NS
PWUE (A/gs)	25		***	***	NS	*
	35		**	***	NS	NS

Summary of statistical analysis using two-way ANOVA for the effects of elevated CO₂ and heat stress (HS) on leaf gas exchange parameters measured at 13 and 17 weeks after planting (WAP). HS plants were measured at the recovery stage (n=9–10). Significance levels are **P<0.001; ** P <0.01; * P <0.05; NS, P>0.05.

Relative to 13 WAP, A_sat decreased after 50% anthesis (17 WAP), but was not affected by eCO₂ in non-HS plants measured at common CO₂ and 25 °C or 35 °C (Fig. 1b, c; Table 1; Supplementary Table S1). When non-HS plants were measured during anthesis at growth CO₂, eCO₂ increased A_sat at 25 °C (+36%, P<0.001) but not at 35 °C (Fig. 1b , c; Table 1; Supplementary Table S1). eCO₂ had no significant effect on g_s of non-HS plants (Morison, 1998) measured 13 or 17 WAP at common or growth CO₂ (Fig. 1d–f; Table 1; Supplementary Table S1).

In non-HS plants, thermal responses of leaf gas exchange differed between aCO₂ and eCO₂ at higher temperatures

A–C_i curves were measured at five leaf temperatures to characterize the thermal photosynthetic responses of wheat plants grown at aCO₂ and eCO₂ (Fig. 2; Table 2; Supplementary Fig. S4). In non-HS, aCO₂- and eCO₂-grown Scout, A_sat, g_s, and C_i increased with temperature up to T_opt of ~23.5 °C and decreased more under eCO₂ relative to aCO₂ at higher temperatures. Relative to aCO₂, plants grown under eCO₂ had higher A_sat up to T_opt but similar A_sat at higher temperatures (Supplementary Fig. S4a). R_d increased with temperature under both aCO₂ and eCO₂; however, the rate of increase was slower at higher temperatures under eCO₂, resulting in lower R_d under eCO₂ relative to aCO₂ at 30°C and 35 °C. Nevertheless, energy of activation (E_aR) and the Q₁₀ coefficient (rate of change due to an increase of temperature by 10 °C) of R_d were similar under aCO₂ and eCO₂ (Table 2; Supplementary Fig. S4b).

Fig. 2.

In vivo Rubisco properties and temperature response of V_cmax and J_max measured 13 weeks after planting (WAP). Maximum velocity of carboxylation, V_cmax (a), maximum velocity of RuBP regeneration, J_max (b), and J_max/V_cmax ratio (c) determined using the response of CO₂ assimilation to variation in chloroplastic CO₂ (C_c) at five leaf temperatures (15, 20, 25, 30, and 35 °C) in wheat cultivar Scout (n=6). The ratio of J_max/V_cmax (c) is plotted using the visreg package in R. Regression lines are means with 95% confidence intervals. The lower panel is a bar plot showing in vivo V_cmax at 25 °C (n=6) (d) and Rubisco sites (n=5) (e) measured in flag leaf discs harvested at the same time point. For (a), (c), and (d), values are means ±SE. Ambient and elevated CO₂-grown plants are shown in black and grey, respectively.

Table 2.

Summary of modelled parameters for temperature response of photosynthesis

Parameter	Constant	Ambient growth CO2	Elevated growth CO2
A sat (µmol m−2 s−1)	T opt (°C)	23.7±1.1 a	23.4±1.3 a
	A opt	25.5±1.3 a	30.9±2.7 b
V cmax (µmol m−2 s−1)	V cmax at 25 °C	149±6 a	121±12 a
	E a V (kJ mol−1)	51±4 a	38±10 a
J max (µmol m−2 s−1)	J max at 25 °C	200±12 a	190±22 a
	T opt (°C)	29.5±0.7 a	27.5±0.9 a
	J max at Topt	233±6	210±11 a
	E a J (kJ mol−1)	37±11 a	34±22 a
	ΔSJ (J mol−1 K−1)	648±5 a	651±8 a
	H d (kJ mol−1)	200
R d (µmol m−2 s−1)	R d at 25 °C	2.4±0.1 a	2.2±0.1 a
	E a R (kJ mol−1)	41±3 a	31±6 a
	Q 10	1.73±0.07 a	1.50±0.13 a

Summary of coefficients derived using non-linear least square fitting of CO₂ assimilation rates, maximal rate of carboxylation (V_cmax), and maximal rate of RuBP regeneration (J_max) determined using A–C_i response curves and dark respiration measured at five leaf temperatures (15, 20, 25, 30, and 35 °C). Values are means ±SEs. Derived parameters include temperature optima (T_opt) of photosynthesis (A_opt); activation energy for carboxylation (E_aV); activation energy (E_aJ)¸ entropy term (∆SJ), and T_opt and corresponding value for J_max with deactivation energy (H_d) assumed constant; and activation energy (E_aR) and temperature coefficient (Q₁₀) for dark respiration. Letters indicate significance of variation in means (n=6)

V _cmax and J_max were calculated by fitting the response of A_sat to variations in C_c (A–C_c response curve) using measured R_d and g_m. g_m increased up to 25 °C and remained relatively unchanged at higher temperatures (Supplementary Table S3). Within the range of measured leaf temperatures, V_cmax increased with leaf temperature, while J_max increased up to a T_opt of 28 °C and decreased thereafter. V_cmax and J_max decreased with eCO₂ at the two highest temperatures (Fig. 2). The J_max/V_cmax ratio was higher under eCO₂ relative to aCO₂ at lower temperatures and decreased with leaf temperature under aCO₂ or eCO₂ (eventually being similar at 35 °C) (Fig. 2). Despite variations in the temperature response, the overall fitted parameters were mostly similar in plants grown at aCO₂ or eCO₂, except for A_opt which was higher under eCO₂ (Table 2). There was no significant difference in V_cmax at 25 °C, J_max at 25 °C, in vitro measured Rubisco sites, or their activation energy under aCO₂ or eCO₂ (Fig. 2; Table 2).

Photosynthesis and PSII efficiency decreased during HS at both CO₂ treatments but recovered only under eCO₂

In this study, we successfully implemented a 5 d HS cycle at 50% anthesis as evidenced by the higher leaf temperature of the HS relative to the control plants (Supplementary Fig. S1f). Overall, HS reduced photosynthesis and was more damaging in aCO₂ than in eCO₂ plants (Fig. 4; Supplementary Table S1). Before HS (15 WAP), eCO₂ increased both A_sat (+43%, P<0.001) and g_s (+20%, P=0.032) measured at growth CO₂. HS reduced A_sat measured during and after HS in both CO₂ treatments. HS increased g_s measured during HS and reduced g_s after HS. One week after HS, both A_sat and g_s had completely recovered in eCO₂-grown plants but not in aCO₂-grown plants, which showed significant reductions in A_sat (–42%, P=0.017) and g_s (–32%, P=0.006) (Fig. 4a, c; Table 1; Supplementary Table S1).

Fig. 4.

Photosynthesis and chlorophyll fluorescence response of aCO₂- and eCO₂-grown wheat cv. Scout measured before, during, after, and at the recovery stage of the heat stress cycle. CO₂ assimilation rates (a), of the F_v/F_m ratio in dark-adapted leaves (b), stomatal conductance (c), and of the F_v'/F_m' ratio in light-adapted leaves (d) measured at growth CO₂ (aCO₂-grown plants measured at 400 μl l⁻¹ and eCO₂-grown plants measured at 650 μl l⁻¹). Values are means ±SE (n=9–10). Ambient and elevated CO₂-grown plants are depicted in black and grey, respectively. Filled and open circles represent control and heat-stressed plants, respectively. The circle and star symbols depict CO₂ assimilation rates measured at 25 °C and 35 °C, respectively.

The lasting negative effect of HS on photosynthesis at the recovery stage was associated with reduced V_cmax (–53%, P=0.002) in aCO₂- but not in eCO₂-grown plants. Conversely, the photosynthetic recovery after HS was associated with increased J_max (+37%, P=0.001) in eCO₂- but not in aCO₂-grown plants (Fig. 3d). Interestingly, HS significantly increased the J_max/V_cmax ratio in both aCO₂- and eCO₂-grown plants, but the ratio was not affected by growth CO₂.

Fig. 3.

Response of V_cmax and J_max to growth at eCO₂ and heat stress (HS) measured 13 and 17 weeks after planting (WAP) at the recovery stage of the HS cycle. Bar plot of means ±SE for V_cmax (a and b), J_max (c and d), and V_cmax/J_max (e and f) using two-way ANOVA. Leaf gas exchange was measured at 25 °C in ambient (black) and elevated (grey) CO₂-grown plants exposed (HS) or not exposed (Control) to a 5 d HS. Bars sharing the same letter in the individual panels are not significantly different according to Tukey’s HSD test at the 5% level. The error bars indicate the SE of the mean (n=9–10). Statistical significance levels (t-test) for eCO₂ effect are shown: *P<0.05; **P<0.01: ***P<0.001.

Chlorophyll fluorescence measurements confirmed the persistent HS damage to photosynthesis in aCO₂- relative to eCO₂-grown plants. HS reduced light-adapted F_v'/F_m' measured after and at the recovery stage of HS in aCO₂- (–29%, P=0.019) but not in eCO₂-grown plants (Fig. 4d). HS reduced dark-adapted F_v/F_m in aCO₂- more than in eCO₂-grown plants; and F_v/F_m failed to recover in aCO₂ plants after HS (Fig. 4b).

At maturity, total plant biomass, but not grain yield, recovered from HS in both CO₂ treatments

Elevated CO₂ stimulated growth rate, biomass, and grain yield. A faster growth rate was evident from the larger number of ears (+127%, P<0.001) in eCO₂- relative to aCO₂-grown plants harvested 13 WAP (before HS) (Fig. 5). Elevated CO₂ significantly stimulated the total biomass harvested throughout the growing period (Fig. 5; Table 3; Supplementary Table S2). Total biomass stimulation was contributed by the overall increase in root, stem, and leaf biomass along with an increase in leaf area, leaf number, tiller number, and spike number (Table 3; Supplementary Table S2). At the final harvest, eCO₂-grown plants had 35% (P<0.001) more biomass and 30% higher grain yield (P=0.001) than aCO₂-grown plants under control conditions (Fig. 5; Table 3; Supplementary Table S2). The increase in grain yield of control plants under eCO₂ was due to an increased number of tillers and consequently ears (+22%, P<0.001), while the main shoot grain yield was not stimulated (Supplementary Table S2).

Fig. 5.

Response of biomass and ears (or tillers) to eCO₂ and HS across the life cycle of wheat cv. Scout. Response of total biomass (a) and spike number (b) to eCO₂ and HS at three time points; before HS (B), after recovery from HS (R), and at the final harvest after maturity (M). Ambient and elevated CO₂-grown plants are depicted in black and grey, respectively. Solid and dotted lines represent control and heat-stressed plants, respectively. Filled and open circles represent control and heat-stressed plants, respectively. Vertical black dotted lines show the timing of HS. Symbols are means per plant ±SE (n=9–10).

Table 3.

Summary of statistics for plant dry mass (DM) and morphological parameters

Time point	Parameter (mean per plant)	Main effects		Interaction
		CO2	HS	CO2×HS
13 WAP (T1)	Tiller number	**
	Leaf number	***
	Leaf area (cm2)	***
	Ear number	***
	Ear DM (g)	***
	Leaf DM (g)	***
	Stem DM (g)	***
	Roots DM (g)	NS
	Shoot DM (g)	***
	Total DM (g)	***
17 WAP (T2)	Tiller number	***	***	***
	Leaf number	***	***	***
	Leaf area (cm2)	***	NS	**
	Ear number	***	***	***
	Ear DM (g)	***	***	NS
	Leaf DM (g)	***	***	***
	Stem DM (g)	***	***	***
	Roots DM (g)	***	**	*
	Shoot DM (g)	***	***	NS
	Total DM (g)	***	***	NS
25 WAP (T3)	Ear number	***	***	*
	Ear DM (g)	***	***	NS
	Roots DM (g)	NS	***	NS
	Shoot DM (g)	***	***	**
	Total DM (g)	***	NS	NS
	Main stem grain yield (g)	NS	***	NS
	Grain yield (g)	***	***	NS
	Grain number	**	***	NS
	Grains per ear	NS	***	NS
	Grain size (mg per grain)	**	***	NS
	Harvest index	**	***	*

Summary of statistical analysis using two-way ANOVA for the effects of elevated CO₂ and heat stress (HS) on biomass and morphological parameters for plants harvested at various time points (n=9–10). Significance levels are *** P <0.001; **P<0.01; *P<0.05; NS, P>0.05

HS reduced the biomass of aCO₂ plants (–30%, P<0.001) more than eCO₂ plants (–10%, P=0.09) harvested at 17 WAP following the HS (Fig. 5; Supplementary Table S2). By the final harvest, HS plants recovered and had similar biomass relative to control plants grown under both aCO₂ and eCO₂. This recovery in biomass was driven by the HS-induced stimulation of additional late tillers and consequently new ears (Fig. 5).

Despite the recovery in biomass, the grain yield was similarly reduced by HS in both aCO₂- (–38%, P<0.001) and eCO₂- (–41%, P<0.001) grown plants due to grain abortion in old ears and insufficient grain filling in new ears (Fig. 6a, b; Supplementary Table S2). HS reduced grain yield of tillers (–77%, P<0.001) more than the main shoot (–45%, P<0.001), which developed earlier. In addition, HS reduced grain yield of tillers in plants grown at aCO₂ (–71 % P<0.001) less than in those grown at eCO₂ (–81%, P<0.001) due to their higher tiller number (Fig. 6c, d). This phenomenon is well recognized as growth stimulation at eCO₂ may limit grain yield due to trade-off between vegetative and reproductive components, including grains (Dias de Oliveira et al., 2015). HS caused grain abortion, leading to empty ears without grains, or damaged and shrunken grains (Supplementary Fig. S5) evident from the reduction in grain per spike (–53%, P<0.001) and average grain weight (–25%, P<0.001) under both CO₂ treatments.

Fig. 6.

Response of plant total biomass and grain yield to elevated CO₂ and heat stress (HS) at the final harvest. Bar plot of means ±SE for total biomass (a), grain yield (b), grain yield of tillers (c), and grain yield of the main shoot (d) using two-way ANOVA measured in ambient (black) and elevated (grey) CO₂-grown plants exposed (HS) or not exposed (Control) to a 5 d HS. Bars sharing the same letter in the individual panels are not significantly different according to Tukey’s HSD test at the 5% level. Values are means ±SE (n=9–10). Statistical significance levels (t-test) for eCO₂ effect are shown: *P<0.05; **P<0.01: ***P<0.001.

Discussion

Although field experiments are crucial to understand plant- and canopy-level responses to the environment, mechanistic and interactive analysis of climate change variables such as eCO₂ and temperature remains challenging in the field. In this study, we investigated the interactive effects of eCO₂ and HS on photosynthesis, biomass, and grain yield of Scout, a high-yielding modern wheat cultivar, under controlled environments. Wheat was grown at ambient or elevated CO₂ and exposed to a 5 d HS at 50% anthesis. eCO₂ stimulated photosynthesis, biomass, and grain yield, while HS reduced photosynthetic rates under aCO₂ and eCO₂. eCO₂ improved the recovery of photosynthesis and biomass following HS, but the HS-induced reduction in grain yield was similar under both CO₂ treatments due to grain abortion and inadequate grain filling. Our study demonstrated the interactive effects between eCO₂ and HS, providing insights into the mechanisms underlying the interactions, and identified a major discrepancy between the response of wheat photosynthesis and biomass versus grain yield to eCO₂×HS. The novelty of this study is in linking photosynthesis, biomass, and grain yield responses to the interactive effects of future climate variables. Our results and modelled parameters will be useful in developing a mechanistic modelling approach based on photosynthesis and parametrization of crop models that can predict yield under future extreme climate. Our results suggest that grain filling and translocation to the grain at high temperature are key weaknesses in modern high-yielding wheat varieties. Moreover, plastic tillering in response to eCO₂ and HS adversely impacted the final grain yield. These traits should be prioritized in current breeding programmes to sustain staple food production under future climate extremes.

Elevated CO₂ reduced photosynthetic electron transport capacity at high temperature

Elevated CO₂ modulates the instantaneous temperature response of photosynthesis (Sage and Kubien, 2007; Ghannoum et al., 2010). Growth at eCO₂ slowed down the rate of increase in V_cmax and accentuated the decrease in J_max above T_opt in Scout (Fig. 2c), possibly due to reduced Rubisco activation and limitation in electron transport capacity at high temperature and eCO₂ (Sage and Kubien, 2007). Contrary to our hypothesis that eCO₂ will increase T_opt (Long, 1991), photosynthetic T_opt was similar under aCO₂ and eCO₂ (Table 2). Lower V_cmax and J_max at higher temperatures may have prevented the increase in T_opt under eCO₂, because the temperature dependence of J_max determines the shift of optimal temperature of photosynthesis at eCO₂ (Hikosaka et al., 2006). Our results are consistent with a previous study by Alonso et al. (2008) where they found decreased V_cmax under eCO₂. In other wheat studies, eCO₂ increased V_cmax and J_max at supraoptimal temperatures, and reduced J_max at suboptimal temperatures (Alonso et al., 2009). The discrepant V_cmax and J_max responses to our studies could be due to an unusual increase in V_cmax under eCO₂ observed in the study by Alonso et al. (2009). Modelled values of V_cmax at 25 °C (150 μmol m⁻² s⁻¹ and 121 μmol m⁻² s⁻¹ at aCO₂ and eCO₂, respectively) are similar to in vitro (137 μmol m⁻² s⁻¹ and 123 μmol m⁻² s⁻¹ at aCO₂ and eCO₂, respectively) values measured in the current study and previously reported in wheat (117 μmol m⁻² s⁻¹) by Silva-Pérez et al. (2017). Modelled values of E_aV (51 kJ and 38 kJ at aCO₂ and eCO₂, respectively) were lower than E_aV (63 kJ) reported in the wheat study by Silva-Pérez et al. (2017).

Elevated CO₂ protected wheat photosynthesis by stimulating electron transport potential following HS

Generally, HS reduces net photosynthetic rates in wheat (Wang et al., 2008), while the extent of the response depends on the cultivar (Sharma et al., 2014). However, acclimation to long-term eCO₂ can modulate the photosynthetic responses to HS during the vegetative or anthesis stage (Wahid et al., 2007). In Scout, photosynthetic rates recovered following HS under eCO₂ but not under aCO₂ (Fig. 4a), indicating that HS transiently reduced photosynthesis without eliciting permanent damage to the photosynthetic apparatus of eCO₂-grown plants. The recovery of photosynthesis under eCO₂ was associated with the recovery of the electron transport rate (Fig. 4d) and photochemical efficiency (Fig. 4a), maintenance of V_cmax (Fig. 3b), and increased J_max (Fig. 3d) relative to non-HS, eCO₂-grown plants, thus validating our hypothesis that eCO₂ will protect photosynthesis via increased electron transport. Higher J_max may have protected the photosynthetic apparatus from HS damage by increasing electron sinks, and hence photochemical quenching (Sage and Kubien, 2007). Higher J_max is also associated with higher Rubisco activation (Perdomo et al., 2017), which may have helped recovery of photosynthesis under eCO₂.

In contrast, aCO₂-grown Scout suffered permanent loss of photosynthesis and photochemical efficiency (F_v/F_m) after HS. Reduced F_v/F_m is a sign of stress (Sharkova, 2001; Haque et al., 2014) and indicates lower quantum efficiency of PSII (Baker, 2008). Damage to the photosynthetic apparatus was also evident from the reduction in V_cmax at aCO₂ (Fig. 3b), although J_max was not affected by HS (Fig. 3d). Consequently, the J_max/V_cmax ratio was equally increased by HS in both CO₂ treatments, suggesting increased resource allocation to RuBP regeneration or electron transport (Hikosaka et al., 2006) in response to HS irrespective of growth CO₂. An enhanced J_max/V_cmax ratio by exposure to HS may potentially play a role in avoiding photoinhibition (Walker et al., 2014).

In line with our results, photosynthesis and F_v/F_m were inhibited by HS (3 d at 40 °C) applied after anthesis in two wheat cultivars grown at aCO₂ but not at eCO₂ (Shanmugam et al., 2013). Protection from HS damage of photosynthesis as a result of improved photochemical quenching or electron transport appears to be a universal mechanism in crops exposed to eCO₂. In tomato, HS (42 °C) reduced A_sat (–57%), V_cmax, and J_max (–45%) under aCO₂, while eCO₂ increased A_sat (+96%), V_cmax, and J_max after 24 h of recovery from HS (Pan et al., 2018). In Arabidopsis, photosynthesis and chlorophyll fluorescence were less inhibited by HS (38 °C) in eCO₂ than in aCO₂ 8 d after recovery (Zinta et al., 2014). The study concluded that eCO₂ mitigated HS stress impacts through up-regulation of antioxidant defence metabolism and reduced photorespiration, resulting in lowered oxidative pressure (Zinta et al., 2014). In other studies investigating the interactive effects of eCO₂ and HS (reviewed by Wang et al., 2011), eCO₂ enhanced the thermal tolerance of photosynthesis in both cool- and warm-season species, indicating that the mitigating effects of CO₂ were independent of the plant habitat (Hogan et al., 1991; Wang et al., 2008).

Following HS, plant biomass recovered in all plants due to late tillering, while grain yield declined even under eCO₂

Despite the initial negative impacts of HS on plant growth in Scout, total plant biomass recovered at maturity, and this was associated with positive source (photosynthesis) and sink (tiller) responses. As discussed earlier, HS caused irreversible photosynthetic damage at aCO₂, while growth at eCO₂ mitigated the negative impact of HS on photosynthesis. Moreover, the biomass of HS plants recovered under both CO₂ treatments due to late tiller and ear development (Bányai et al., 2014). When grain development is stalled under certain conditions (e.g. HS), the crop develops new grains by producing additional late tillers. This is considered a non-harmful acclimation response to HS which creates additional sinks. Hence, grain abortion due to HS was compensated by the production of additional late tillers contributing to the recovery in biomass at the final harvest. An equal decrease in biomass under aCO₂ and eCO₂ following exposure to HS has been reported in a study using the C₃ crop Sinapis alba (white mustard), which also concluded that interactive effects of CO₂ and HS depend on species, magnitude of HS, and growth conditions (Coleman et al., 1991). In cases where HS causes persistent reduction in biomass at aCO₂, eCO₂ often alleviates the negative impacts of HS (Zinta et al., 2014). It is worth noting that the development of additional late ears and tillers following HS is expected to increase sinks for the translocation of assimilates. Greater sink strength may partly explain photosynthetic recovery in HS plants (Paul and Foyer, 2001). However, photosynthesis recovered in eCO₂ plants only, while late tillering was observed under both CO₂ treatments. Similarly, in wheat grown using growth chambers and exposed to moderate HS (32 °C) after anthesis, grain yield decreased under both ambient and elevated CO₂ (Zhang et al., 2018). Althogh Scout biomass recovered in all plants exposed to HS, grain yield was equally reduced in both CO₂ treatments due to grain abortion in the old ears and insufficient time for grain filling in the new ears. In response to HS, some ears had completely lost grains, and ears with developing grains could not fill, leading to shrunken and damaged grains (Supplementary Fig. S5), and hence a significant loss of grain yield consistent with previous studies (Stone and Nicolas, 1996, 1998; Spiertz et al., 2006; Prasad and Djanaguiraman, 2014).

Observed HS damage to grain yield was higher in tillers than in the main shoot (Fig. 6c, d) due to high senstivity of wheat at heading and anthesis stages (Prasad and Djanaguiraman, 2014). When HS was applied, the main shoots may have been past anthesis while tillers were in the heading or anthesis stages, and thus more exposed to HS impacts (Prasad and Djanaguiraman, 2014). FACE studies (Fitzgerald et al., 2016; Macabuhay et al., 2018) in wheat involving interactive effects eCO₂ and HS found that eCO₂ can buffer against heat waves, and eCO₂ may moderate some effects of HS in wheat depending on seasonal conditions and HS timing.

In conclusion, eCO₂ stimulated photosynthesis, biomass, and grain yield in a modern, high-yielding wheat variety. In non-HS plants, photosynthetic stimulation by eCO₂ was observed despite reduction of V_cmax at all temperatures and J_max at higher temperatures. In heat-shocked plants, eCO₂ stimulated J_max and maintained photochemical efficiency, hence providing photosynthetic protection against HS damage. Consequently, HS reduced photosynthesis under aCO₂ more than under eCO₂. Plant biomass completely recovered from HS under both CO₂ treatments due to the development of additional late tillers and ears; yet these did not fully develop and fill grains. Therefore, HS applied at anthesis equally reduced grain yield under aCO₂ and eCO₂ due to grain abortion. In the field, late tillers would not necessarily produce higher grain yield either, because plants will run out of soil water and there is not enough time for grain filling. The current study demonstrates the interactive impacts of eCO₂ and severe HS applied at 50% anthesis on wheat yield. HS can occur over a wide window from booting to late grain-filling stage, thus affecting yield in variable ways and limiting the generalization of our results. Nonetheless, our study provides insights into the interactive effects of eCO₂ and HS on the thermal responses of wheat photosynthesis which apply over a wide range of scenarios, and hence can form the basis for crop models to incorporate the interactive effects of eCO₂ and HS.

Supplementary data

Supplementary data are available at JXB online.

Table S1. Response of leaf gas exchange parameters to elevated CO₂ and heat stress.

Table S2. Response of plant dry mass and morphological parameters to elevated CO₂ and heat stress.

Table S3. Temperature response of mesophyll conductance in Scout.

Fig. S1. Glasshouse growth conditions and heat stress cycle.

Fig. S2. Radiation over time during the experiment.

Fig. S3. Experimental design depicting plant growth plotted over time.

Fig. S4. Temperature response of spot gas exchange parameters.

Fig. S5. Response of grain size and morphology to heat stress.