Effects of three patterns of elevated CO₂ in single and multiple generations on photosynthesis and stomatal features in rice

Abstract

Background and Aims

Effects of elevated CO₂ (E) within a generation on photosynthesis and stomatal features have been well documented in crops; however, long-term responses to gradually elevated CO₂ (E_g) and abruptly elevated CO₂ (E_a) over multiple generations remain scarce.

Methods

Japonica rice plants grown in open-top chambers were tested in the first generation (F1) under E_a and in the fifth generation (F5) under E_g and E_a, as follows: E_a in F1: ambient CO₂ (A) + 200 μmol mol⁻¹; E_g in F5: an increase of A + 40 μmol mol⁻¹ year⁻¹ until A + 200 μmol mol⁻¹ from 2016 to 2020; E_a in F5: A + 200 μmol mol⁻¹ from 2016 to 2020. For multigenerational tests, the harvested seeds were grown continuously in the following year in the respective CO₂ environments.

Key Results

The responses to E_a in F1 were consistent with the previous consensus, such as the occurrence of photosynthetic acclimation, stimulation of photosynthesis, and downregulation of photosynthetic physiological parameters and stomatal area. In contrast, multigenerational exposure to both E_g and E_a did not induce photosynthetic acclimation, but stimulated greater photosynthesis and had little effect on the photosynthetic physiology and stomatal traits. This suggests that E retained intergenerational effects on photosynthesis and stomatal features and that there were no multigenerational differences in the effects of E_g and E_a.

Conclusions

The present study demonstrated that projecting future changes induced by E based on the physiological responses of contemporary plants could be misleading. Thus, responses of plants to large and rapid environmental changes within a generation cannot predict the long-term response of plants to natural environmental changes over multiple generations, especially in annual herbs with short life cycles.

INTRODUCTION

Anthropogenic pressures are rapidly changing the global environment, including the atmospheric carbon dioxide (CO₂) concentration, which is now >400 μmol mol⁻¹ (IPCC, 2022). Crops are sensitive to elevated CO₂ (E), and relevant studies have become a hot topic over the last three decades because of its potentially profound impact on crop ecosystems and a strong concern for future food security (Ziska and Bunce, 2007; Lobell et al., 2011; Adachi et al., 2014; Cai et al., 2020; Ainsworth and Long, 2021).

Given the large number of E experiments implemented, changes in photosynthetic characteristics, such as stimulated photosynthesis and photosynthetic resource use efficiency, downregulated biochemical capacity, stomatal conductance and dark-adapted chlorophyll fluorescence parameters, have been investigated well and reviewed (Huxman et al., 1999; Ainsworth and Long, 2005; Leakey et al., 2009; Shanmugam et al., 2013; Zong et al., 2014; Yang et al., 2023a). Nevertheless, the response of stomatal features (stomatal density and area) to E varies among species (Bettarini et al., 1998; Ainsworth and Rogers, 2007; Haworth et al., 2015). Acclimation to E is defined as follows: the values (net photosynthetic rate and stomatal conductance) are significantly lower in plants grown under E than under ambient CO₂ when both are measured at the same CO₂ concentration (Drake et al., 1997; Lodge et al., 2001), and both photosynthetic acclimation and stomatal acclimation are still under discussion because of the importance of accurate prediction of future carbon uptake and storage (Smith and Dukes, 2013; Li et al., 2015; Cai et al., 2018).

Notably, almost all the above studies on crops are based on an abruptly elevated CO₂ (E_a) within a generation, such as ambient CO₂ + 200 μmol mol⁻¹ in one growing season. However, CO₂ levels are rising gradually; the average rate of CO₂ increase was 2.4 μmol mol⁻¹ year⁻¹ during the last decade (https://www.esrl.noaa.gov/). Thus, small and slow environmental changes across generations might be more relevant in predicting the response to future climate change (Donelson et al., 2018). However, long-term responses to gradually elevated CO₂ (E_g) and E_a over multiple generations have received little attention owing to the time, energy and funds required for initiating such projects.

Although a few studies have reported differences between E_a and E_g, they did not involve an intergenerational effect, because these studies were conducted in one growing season or on trees (Luo and Reynolds, 1999; Hui et al., 2002; Drake et al., 2016). Previous reports found that compared with exposure to E_a for a single generation, multigenerational exposure to E_a could enhance biomass further (Frenck et al., 2013; Li et al., 2019). A limited number of studies have shown that E_a decreases the response of the structural and functional community of mycorrhizal fungi; however, exposure to E_g over 21 generations, with an increase of 10 μmol mol⁻¹ during each generation, does not change the fungal diversity and functioning (Klironomos et al., 2005). Therefore, it is reasonable to believe that the responses of plants to long-term exposure to E_a and E_g over multiple generations might differ from those to short-term exposure to E_a within a generation.

Facing human-mediated climate change, plants might adapt over generations, and plants of the next generation are likely to be different from contemporary plants, leading to long-term trends in plant fitness (Collins and Bell, 2004; Burgess and Marshall, 2014; Yin et al., 2022). Maternal environment can have substantial effects on offspring (transgenerational effects). Transgenerational adaptation is a flexible evolutionary mechanism that can result from seed modification or epigenetic variation (Lau et al., 2008; Donohue, 2009; Vráblová et al., 2018). Additionally, the adaptation of physiological traits is generally more rapid than that of morphology and growth because of the relatively few genes (Herman and Sultan, 2011; Leakey and Lau, 2012; Yin et al., 2019).

The objective of this study was to investigate photosynthesis and stomatal features in two aspects: (1) whether multigenerational exposure to E, including E_a and E_g, altered these traits compared with E_a within a generation; and (2) whether there were multigenerational differences in the effects of E_a and E_g. Thus, in this study, photosynthesis and stomatal features of japonica rice were measured under three patterns of E in single and multiple generations from 2016 to 2020: E_a within a generation, E_g over multiple generations and E_a over multiple generations.

MATERIALS AND METHODS

Site description

The experiment was conducted at the Agricultural Meteorology and Ecology Experiment Station of the Nanjing University of Information Science and Technology (32.16°N, 118.86°E) during the rice-growing seasons from 2016 to 2020. The region has a subtropical humid climate, with a mean annual temperature of 15.6 °C, mean annual precipitation of ~1100 mm, and mean annual sunshine of >1900 h. The soil texture was a silty loam with clay, silt and sand contents of 9.5, 85.2 and 5.3 %, respectively. The tilled soil had a pH of 6.1, and soil organic carbon, total N, total P and total K contents of 9.52, 1.18, 0.85 and 18.17 g kg^–1, respectively, at a depth of 0–20 cm.

Experimental design

Twelve open-top chambers (OTCs) were established to simulate CO₂ enrichment, with four OTCs designated as replicates for each treatment. There were three treatments of CO₂ concentration, including one ambient atmospheric CO₂ concentration (A) and two elevated CO₂ concentrations (E): (1) maintaining A from 2016 to 2020, at ~400 μmol mol⁻¹; (2) a ‘ramp’ of A + 40 μmol mol⁻¹ steps, with each step lasting for one growing season from 2016 to 2020; that is, A + 40 μmol mol⁻¹ in 2016, A + 80 μmol mol⁻¹ in 2017, A + 120 μmol mol⁻¹ in 2018, A + 160 μmol mol⁻¹ in 2019 and A + 200 μmol mol⁻¹ in 2020, which could be considered as E_g; and (3) maintaining A + 200 μmol mol⁻¹ from 2016 to 2020, which could be considered as E_a.

This study included the following design: (1) E_a within a generation (E_a in F1), that is, seeds grown at A and E_a in one growing season in 2019 (Fig. 1A); (2) E_g over five generations (E_g in F5), that is, seeds continuously grown at A and E_g over five generations from 2016 to 2020 (Fig. 1B); and (3) E_a over five generations (E_a in F5), that is, seeds continuously grown at A and E_a over five generations from 2016 to 2020 (Fig. 1B).

Fig. 1.

Schematic representation of the experimental design and treatments. (A) abruptly elevated CO₂ within a generation (E_a in F1). (B) Gradually elevated CO₂ over five generations (E_g in F5) and abruptly elevated CO₂ over five generations (E_a in F5).

For the multigenerational tests, the harvested seeds were grown in the following year continuously in the respective CO₂ environments (Fig. 1B). Seeds were harvested around 20 October every year and collected from maternal plants of each of the four A, E_g and E_a plots. Harvested seeds were air-dried to a constant weight (14 % moisture) and stored in a refrigerator at 4 °C before being transferred to their respective OTCs for germination and experimental trait analyses.

The OTCs were regular octagonal prism-shaped structures constructed using an aluminum alloy frame (12 m² bottom area and 3 m height) with a highly transparent glass covering (3 mm thickness and 80 % light transmittance), and the light transmittance did not differ among the OTCs (data not shown). CO₂ (GMM222; Vaisala, Helsinki, Finland) and temperature sensors were suspended in the middle of the chamber, 1.5 m above the ground, to record the CO₂ concentration and temperature in the OTCs. The photosynthetic photon flux density was monitored using a micro-meteorological station (AWS800; Campbell Scientific, USA) installed in the experimental field. The details of the control system for CO₂ enrichment have been described elsewhere (Yang et al., 2021). A summary of the CO₂ concentrations, temperatures and photosynthetic photon flux density from 2016 to 2020 is presented in Supplementary data Table S1.

Rice cultivation

The japonica rice (Oryza sativa L.) cultivar ‘Nanjing 9108’ was grown in OTC field plots. In 2019 and 2020, pre-germinated seeds were sown on 20 May, and the rice seedlings were transplanted on 20 June. Nitrogen fertilizer was applied at a rate of 25 g m⁻² using compound chemical fertilizer (N:P:K = 15 %: 15 %: 15 %) for supplying basal N and urea (N, 46 %) for top-dressing. The basal N and two top-dressings were applied at a ratio of 6:2:2 pre-transplantation (20 June 2019 and19 June 2020), early tillering (7 July 2019 and 7 July 2020) and panicle initiation (26 July 2019 and 26 July 2020). Field management, such as water supply and disease control, was performed following local practices. The rice cultivation from 2016 to 2018 was almost consistent with that during 2019 and 2020.

Gas exchange measurements

Before gas exchange measurements, a chlorophyll meter (SPAD-502; Konica Minolta Optics, Japan) was used to measure the chlorophyll content (SPAD) of eight to ten leaves per OTC in the fully expanded youngest leaves (flag leaves at the filling stage). Three leaves (in 2019) or two leaves (in 2020) with representative SPAD (values near the mean) were selected for the subsequent gas exchange measurements with an LI-Cor 6400XT Portable Photosynthesis System (Li-Cor, Lincoln, NE, USA) from 09.00 to 14.00 h on sunny days in each OTC. The measurements were conducted at the jointing stage (1–3 August 2019 and 9–13 August 2020) and the filling stage (22–25 September 2019 and 26–30 September 2020). Basal CO₂ leakage in LI-Cor 6400XT usually occurs at the interface between the two sealing gaskets of the clamping chamber (Flexas et al., 2007). Therefore, we used the latex gaskets of LI-Cor 6800 (spare part number 6568-566), which has better elasticity and plasticity, to replace the foamed gaskets of LI-Cor 6400 (white, spare part number 6400-30; and black, spare part number 6400-33).

During the CO₂ response curve (A–C_i, C_i: intercellular CO₂ concentration) measurements in 2019 and 2020, the photosynthetic photon flux density was maintained at 1500 μmol m⁻² s⁻¹ and the leaf temperature was set to 33 °C at the jointing stage and 28 °C at the filling stage. The vapour pressure deficit was 1.77 ± 0.19 kPa in 2019 and 1.78 ± 0.25 kPa in 2020, and the flow rate for all measurements was 500 μmol s⁻¹ in both years. For the A–C_i curve measurement, the CO₂ concentration (C_a) steps were 400, 300, 200, 100, 50, 400, 500, 600, 800, 1000 and 1200 μmol mol⁻¹ under A and 600, 500, 400, 300, 200, 100, 50, 600, 800, 1000 and 1200 μmol mol⁻¹ under E.

The maximum rates of carboxylation (V_cmax) and electron transport (J_max) were determined using the Farquhar, von Caemmerer and Berry models (Farquhar et al., 1980) and were rescaled to 25 °C (V_cmax25 and J_max25) using the temperature response functions given by Bernacchi et al. (2001) and Bernacchi et al. (2003), respectively.

Leaf dark-adapted chlorophyll fluorescence

After gas exchange, leaf dark-adapted chlorophyll fluorescence parameters were estimated using a plant efficiency analyzer (Pocket PEA; Hansatech Instruments, King’s Lynn, UK) at a saturating light pulse of 3500 µmol m⁻² s⁻¹. The dark-adapted leaves were measured at predawn (~05.00 h) at the same leaf locations where the gas exchange was determined. The maximum quantum yield of photosystem II (PSII) (F_v/F_m) was calculated as (F_m − F_o)/F_m, where F_o and F_m are the minimum and maximum fluorescence intensity, respectively (Schreiber and Armond, 1978). The quantum yield for the energy dissipation (φD_o) was calculated as F_o/F_m, and the quantum yield for electron transport (φE_o) was calculated as (1 − F_o/F_m) × (1 − V_J), where V_J represents the relative variable fluorescence in phase J of the fluorescence induction curve (Strasser and Strasser, 1995; Tsimilli-Michael and Strasser, 2008). The performance index (PI_ABS) is the index that integrates different phenomena related to PSII activity (Wang et al., 2016).

Measurement of stomatal parameters

A section of the abaxial epidermis measuring ~1 cm² was imprinted from the middle of the leaf using clear nail varnish. Stomatal density (the number of stomata per unit leaf area) was observed using a microscope (XD-MDI; Xindi, Shanghai, China), and the area of the stomatal apparatus (the size of the guard cells and the pore) was calculated using the image of stomata with ImageJ, a publicly available Java-based image processing program developed at the US National Institutes of Health (http://rsb.info.nih.gov/ij/). The stomatal density of each segment was determined for the five fields. The areas of five stomatal apparatuses were observed in each field. The stomatal apparatus area index (the ratio of stomatal area to unit leaf area) in each field was calculated as the stomatal density multiplied by the stomatal apparatus area. The measurement of stomatal parameters within generation mentioned above refers to our recent study (Yang et al., 2021).

Leaf mass and nitrogen content per unit area

Leaf area measurements were performed using an area meter (Li-3100; Li-Cor, Lincoln, NE, USA). Subsequently, the leaves were oven-dried at 105 °C for 30 min, then at 80 °C to a constant weight, and weighed. The leaf nitrogen concentration was measured using a CHNOS Elemental Analyzer (MOD Vario El III, Germany). Finally, leaf mass per unit leaf area (LMA; in grams per square metre) and nitrogen content per unit leaf area (N_a; in grams per square metre) were calculated using the above data.

Photosynthetic resource use efficiency

Water use efficiency (WUE) was defined as the net photosynthetic rate at growth CO₂ concentration (A_growth) at saturating light intensity (1500 μmol m⁻² s⁻¹) divided by the transpiration rate (T_r). Likewise, photosynthetic nitrogen use efficiency (PNUE) was defined as A_growth divided by N_a.

Statistical analysis

An ANOVA with Tukey’s HSD post-hoc test was applied to analyse the difference in a given item between the treatments. The A–C_i curves were fitted using the ‘plantecophys’ package of R software. Linear regression was used to quantify the relationships between variables. Statistical analyses were performed using IBM SPSS 20.0 (SPSS, Chicago, IL, USA) and Microsoft Excel 2016. Statistical significance was set at P < 0.05.

RESULTS

Gas exchange parameters

In all treatments (E_a in F1 and E_g and E_a in F5), E dramatically stimulated the net photosynthetic rate at growth CO₂ concentrations (A_growth) at the jointing and filling stages (P < 0.05; Table 1). The stimulation of the average A_growth was 21, 32 and 33 % at E_a in F1, E_g in F5 and E_a in F5, respectively, across the two stages.

Table 1.

Gas exchange parameters measured at growth CO₂ concentration under E_a in F1 and E_g and E_a in F5

Generation	Stage	CO2	A growth (μmol m−2 s−1)	g s (mol m−2 s−1)	T r (mmol m−2 s−1)	V cmax25 (μmol m−2 s−1)	J max25 (μmol m−2 s−1)
F1	Jointing	A	29.6 (0.5) b	0.338 (0.014) a	5.52 (0.21) a	115.7 (2.3) a	139.6 (3.5) a
		E a	33.1 (0.8) a	0.254 (0.020) b	4.69 (0.32) b	103.2 (2.3) b	131.0 (2.4) b
	Filling	A	17.3 (0.4) b	0.148 (0.007) a	2.60 (0.09) a	82.0 (1.8) a	145.3 (3.4) a
		E a	22.3 (0.4) a	0.133 (0.007) a	2.31 (0.08) b	75.2 (1.0) b	134.0 (3.6) b
F5	Jointing	A	25.3 (0.4) b	0.211 (0.012) a	4.57 (0.27) a	102.9 (2.1) a	129.8 (2.9) a
		E g	32.0 (0.6) a	0.203 (0.014) a	3.84 (0.17) a	100.5 (1.3) a	128.0 (3.0) a
		E a	31.9 (0.5) a	0.207 (0.019) a	3.97 (0.22) a	96.1 (2.4) a	127.0 (2.9) a
	Filling	A	15.9 (0.5) b	0.140 (0.005) a	2.22 (0.09) a	69.4 (2.0) a	123.7 (3.2) a
		E g	21.9 (0.9) a	0.121 (0.010) a	2.04 (0.15) a	69.3 (2.3) a	129.8 (3.8) a
		E a	22.1 (0.4) a	0.116 (0.007) a	1.95 (0.09) a	69.3 (1.7) a	133.6 (2.9) a

Abbreviations: A_growth, net photosynthetic rate at growth CO₂ concentration; g_s, stomatal conductance to CO₂; J_max25, the maximum electron transport rate at 25 °C; T_r, transpiration rate; V_cmax25, the maximum rate of Rubisco carboxylation at 25 °C. Values are the mean (s.e.m.), n = 4. Different lowercase letters indicate significant differences among the CO₂ treatments (P < 0.05).

However, the remaining gas exchange parameters exhibited distinct CO₂ responses in F1 and F5. In F1, E_a significantly decreased g_s at the jointing stage and T_r, V_cmax25 and J_max25 at two stages (P < 0.05), with only a non-significant reduction in g_s at the filling stage (P > 0.05). In contrast, E_g and E_a in F5 did not affect these traits at the two stages (P > 0.05), despite a tendency for a reduction in g_s and T_r (Table 1). Furthermore, all gas exchange parameters in F5 did not differ between E_g and E_a (Table 1).

Photosynthetic and stomatal acclimation

In F1, the net photosynthetic rate (A_n) was significantly lower under E_a than under A at different C_i conditions (P < 0.05), with the exception of the lowest C_i concentration at the jointing and filling stages (Fig. 2A, B). However, in F5, A_n did not differ at all C_i concentrations under E_g and E_a at two stages (P > 0.05; Fig. 2C, D). The results clearly showed the occurrence of photosynthetic acclimation under E_a in F1, but not under E_g and E_a in F5.

Fig. 2.

The A–C_i curves of E_a in F1 (A, B), E_g and E_a in F5 (C, D) at the jointing (A, C) and filling stages (B, D). Each point represents the mean of four replicates, and bars represent the s.e.m. *P < 0.05 and **P < 0.01; ns, no significant difference. Panels A and B were taken from our previous paper (Yang et al., 2021).

Compared with A in F1, g_s was lower under E_a only when C_i was <400 μmol mol⁻¹ at the jointing stage (P < 0.05; Supplementary data Fig. S1A), whereas g_s did not differ at high C_i at the jointing stage and at all C_i levels at the filling stage between the CO₂ treatments (P > 0.05; Supplementary data Fig. S1A, B). Likewise, E_g and E_a in F5 did not alter g_s in all C_i conditions at two stages (P > 0.05), except for two C_i concentrations at the filling stage (Supplementary data Fig. S1C, D). This suggests that stomatal acclimation rarely occurs under E, regardless of the E pattern (E_gvs E_a) and generation (F1 vs F5).

Leaf traits

The LMA was not significantly affected by E_a in F1 at the jointing and filling stages, and a slightly downward trend was observed, whereas E_g and E_a in F5 significantly or slightly increased LMA at the two stages (Table 2). Compared with A in F1, N_a was lower under E_a, with significant or non-significant differences; however, N_a was slightly higher but non-significant under E_g and E_a than under A in F5 at the two stages (Table 2). Additionally, LMA and N_a in F5 did not show significant difference between E_g and E_a.

Table 2.

Effects of elevated CO₂ on leaf traits in F1 and F5

Generation	Stage	CO2	LMA (g m−2)	N a (g m−2)
F1	Jointing	A	56.0 ± 0.9 a	2.11 ± 0.04 a
		E a	55.8 ± 0.7 a	2.06 ± 0.02 a
	Filling	A	66.2 ± 1.1 a	1.68 ± 0.04 a
		E a	65.4 ± 1.2 a	1.58 ± 0.04 b
F5	Jointing	A	49.3 ± 0.7 a	1.63 ± 0.04 a
		E g	50.7 ± 1.0 a	1.67 ± 0.03 a
		E a	51.8 ± 0.8 a	1.67 ± 0.03 a
	Filling	A	61.0 ± 2.0 b	1.37 ± 0.05 a
		E g	67.7 ± 1.4 a	1.47 ± 0.03 a
		E a	66.9 ± 1.0 a	1.42 ± 0.03 a

Abbreviations: LMA, leaf mass per unit leaf area; N_a, leaf nitrogen content per unit leaf area. Values are the mean ± s.e.m. (n = 4). Different lowercase letters indicate significant differences among CO₂ treatments (P < 0.05). The values of F1 were taken from our previous paper (Yang et al., 2021).

Photosynthetic nitrogen and water use efficiency

Independent of single and multiple generations, E significantly promoted PNUE (Fig. 3A, B) and WUE (Fig. 3C, D). The average stimulation in PNUE was 26, 26 and 29 % under E_a in F1, E_g in F5 and E_a in F5, respectively, across the two stages. Accordingly, the average stimulation in WUE was 41, 51 and 52 % under E_a in F1, E_g in F5 and E_a in F5, respectively.

Fig. 3.

Photosynthetic nitrogen use efficiency (PNUE) and water use efficiency (WUE) of E_a in F1 (A, C) and E_g and E_a in F5 (B, D) at the jointing and filling stages. Different lowercase letters indicate significant differences among the bars (values are shown as the mean + s.e.m., P < 0.05, n = 4).

Stomatal features

In F1, stomatal apparatus area (SA) and stomatal apparatus area index (SAI) were significantly lower under E_a than under A at the two stages, whereas stomatal density (SD) did not change between CO₂ treatments (Table 3). In contrast to E_a in F1, E_g and E_a in F5 had no effects on SD or on SA and SAI at the two stages, with the exception of decreased SA under E_a at the jointing stage (Table 3; Fig. 4). Notably, all stomatal parameters in F5 were similar between E_g and E_a, and the magnitude of change was always larger under E_a than under E_g (Table 3; Fig. 4).

Table 3.

Effects of elevated CO₂ on stomatal parameters in F1 and F5

Generation	Stage	CO2	SD (mm−2)	SA (μm2)	SAI (%)
F1	Jointing	A	532 ± 5 a	302.5 ± 1.2 a	16.1 ± 0.1 a
		E a	538 ± 3 a	273.8 ± 0.7 b	14.7 ± 0.1 b
	Heading	A	724 ± 6 a	258.5 ± 0.8 a	18.7 ± 0.2 a
		E a	719 ± 4 a	252.2 ± 0.9 b	18.1 ± 0.1 b
F5	Jointing	A	514 ± 9 a	133.1 ± 2.1 a	6.9 ± 0.2 a
		E g	528 ± 10 a	127.5 ± 1.7 ab	6.8 ± 0.2 a
		E a	532 ± 10 a	123.0 ± 1.8 b	6.6 ± 0.1 a
	Filling	A	702 ± 12 a	111.8 ± 1.7 a	7.8 ± 0.2 a
		E g	706 ± 12 a	111.3 ± 1.6 a	7.8 ± 0.2 a
		E a	710 ± 11 a	111.0 ± 1.4 a	7.9 ± 0.2 a

Abbreviations: SA, stomatal apparatus area; SAI, stomatal apparatus area index; SD, stomatal density. Values are the mean ± s.e.m. (n = 4). Different lowercase letters indicate significant differences among CO₂ treatments (P < 0.05).

Fig. 4.

The response ratio of stomatal density (SD), stomatal apparatus area (SA) and stomatal apparatus area index (SAI) under E_a in F1 and under E_g and E_a in F5 at the jointing (A) and filling (B) stages. *P < 0.05 (n = 4). SAI = SD × SA ×10⁻⁴ ×100 %. Response ratio = (E − A)/A, where A (ambient CO₂) and E (elevated CO₂) are from the same generation (F1 or F5).

Chlorophyll fluorescence of dark-adapted leaves

Although the magnitude of the chlorophyll fluorescence parameters did not show significant differences among the CO₂ treatments in F1 and F5, the directions of these parameters differed between F1 and F5 (Supplementary data Table S2). For example, the maximum quantum yield of PSII (F_v/F_m), quantum yield for electron transport (φE_o) and performance index (PI_ABS) had negative effects, whereas the quantum yield for energy dissipation (φD_o) had a positive effect under E_a in F1 at the two stages. However, these traits did not exhibit consistent directions under E_g and E_a in F5.

DISCUSSION

Almost all growth chamber, greenhouse and field experiments have investigated the response of crops to E_a only within a generation; for example, ambient CO₂ + 200 μmol mol⁻¹ in one growing season (Wand et al., 1999; Ainsworth and Long, 2005; Leuzinger and Hättenschwiler, 2013; Ainsworth and Long, 2021), as shown in Fig. 1A in the present study. However, multigenerational exposure to E might alter plant physiology and growth in ways that cannot be predicted from single-generation studies, especially for C₃ species that are most sensitive to changes in the atmospheric CO₂ concentration (Ward and Kelly, 2004). Hence, our study provides new insights into the effects of E over multiple generations on photosynthesis and stomatal features in japonica rice. Two patterns of E were used in this study: E_g and E_a.

Three influential meta-analyses have reported that E stimulates photosynthesis by 13–35 % in crops (Ainsworth and Long, 2005; Ainsworth and Rogers, 2007; Leakey et al., 2009). Our results agree with the conclusion, as observed under E_a in F1 with a 21 % stimulation in A_growth (Table 1). Likewise, E_g and E_a in F5 markedly increased A_growth by 32 and 33 %, respectively (Table 1). This suggests that E-induced stimulation of photosynthesis is a certain fact, regardless of the pattern of E (E_gvs E_a) and the generation (F1 vs F5) (Saban et al., 2019). Interestingly, in contrast to F1, in which the remaining gas exchange parameters decreased under E_a, which was consistent with the previous consensus (Ainsworth and Long, 2005; Wang et al., 2013; Hu et al., 2022), E_g and E_a in F5 did not affect these traits (Table 1). This indicates that transgenerational effects can attenuate the negative impact of E on photosynthetic physiological traits. Additionally, the larger stimulation in A_growth in F5 than in F1 was attributed to the lower reduction in g_s, owing to less CO₂ limitation at constant external CO₂ concentrations (Supplementary data Fig. S2). The reduction in leaf nitrogen can be explained by the decreased transpiration-driven mass flow under E (McDonald et al., 2002; McGrath and Lobell, 2013), which matched the relationships between T_r and N_a under E_a in F1. Likewise, the absence of reduction in T_r and N_a also occurred simultaneously under E_g and E_a in F5 (Tables 1 and 2).

Stomatal structure is an embodiment of the ability of plants to adapt to the environment (Kondamudi et al., 2016). An early study using fossil leaves showed that the evolution of leaf-form with the development of stomatal characteristics is causally associated with the change in atmospheric CO₂ over time (Beerling et al., 2001). In the present study, stomatal features exhibited different responses not only between F1 and F5 but also between E_a and E_g. To be specific, E_a in F1 decreased SA and SAI at the two stages, E_a in F5 merely decreased SA at the jointing stage, and E_g in F5 did not affect stomatal traits at two stages (Fig. 4). This suggests that compared with a single generation, multigenerational exposure to E had little effect on stomatal features, especially under the E_g pattern. Thus, generations regulate the adaptation of stomata to a CO₂-enriched environment. Furthermore, SA determines the response of g_s to E (Ainsworth and Rogers, 2007; Yang et al., 2021), and this relationship was also observed under E_g and E_a over multiple generations (Supplementary data Fig. S3).

Photosynthetic and stomatal acclimation in CO₂-enriched environments is a major problem that increases the complexity of models for predicting future carbon uptake and storage (Martinez-Carrasco et al., 2005; Li et al., 2012; Cai et al., 2018; Smith and Keenan, 2020). The vast majority of experiments have proved the existence of photosynthetic acclimation under E (Sage, 1994; Leakey et al., 2009; Hu et al., 2022), as shown under E_a in F1. However, photosynthetic acclimation was not detected under E_g and E_a in F5 (Fig. 2C, D). Previous studies have demonstrated that photosynthetic acclimation is accompanied by a reduction in V_cmax, J_max, N_a, F_v/F_m and φE_o (Hymus et al., 2001; Ainsworth and Long, 2005; Chen et al., 2005; Lawlor and Tezara, 2009; Wilhelm and Selmar, 2011). This was consistent with the present findings for E_a in F1 but not for V_cmax and J_max under E_g and E_a in F5 (Tables 1 and 2; Fig. 2; Supplementary data Table S2). Furthermore, stomatal acclimation rarely occurred in all treatments (Supplementary data Fig. S1). These results suggest that compared with a single generation, multigenerational exposure to E, including E_g and E_a, altered the ability for acclimation of photosynthesis, but not for acclimation of stomata.

PNUE and WUE were promoted by E (Fig. 3); however, the mechanism differed between generations. In F1, the stimulated PNUE (or WUE) was attributed to increased A_growth and decreased N_a (or T_r) (Tables 1 and 2). In contrast, in F5, only the increased A_growth, not the decreased N_a (or T_r), accounted for the stimulated PNUE (or WUE). A limited study found that multigenerational exposure to E_a could modulate g_s, thereby attenuating the negative impact of drought stress on WUE in wheat (Li et al., 2017). In the present study, the larger WUE stimulation over multiple generations than for a single generation supported and broadened this finding in both E_a and E_g conditions.

To our knowledge, the present study is the first to compare and illustrate the extent and direction of the changes in photosynthesis and stomatal features under E_a and E_g over multiple generations relative to those under E_a within a generation. As discussed above, the effects of E_a on plants within a generation have been well documented and are consistent with our results. To date, only a single study has reported that fungal species richness and mycorrhizal fungal taxa decreased under E_a, but not under E_g over 21 generations (Klironomos et al., 2005). This observation seemed to support our finding that E_g in F5 had little effect on plant physiological traits. Several studies supported the results of E_a in F5, which showed that there was a lack of photosynthetic acclimation under E_a over multiple generations (Bezemer and Jones, 2012; Haworth et al., 2015). Furthermore, some studies of natural CO₂ springs (surrounding volcanic degassing vents) might correspond to the treatment of E_a in F5 (Miglietta et al., 1993). For instance, no difference in leaf nitrogen for monocotyledonous species and a trend of reduced photosynthetic acclimation was found in natural CO₂ springs (long-term exposure, >10 years; Onoda et al., 2007, 2009; Watson-Lazowski et al., 2016), possibly because multigenerational plants increased nitrogen uptake from inorganic nitrogen pools in the soil (Ueda et al., 2017).

A meta-analysis of responses to E in natural CO₂ springs showed that photosynthetic physiological parameters decreased in CO₂-enriched environments (Saban et al., 2019); however, this result was not observed in the present study. The discrepancy between natural CO₂ springs and the present study might be based on the fact that sufficient water and nitrogen are available in agricultural ecosystems. Thus, transgenerational adaptation might differ between natural and human-managed agricultural ecosystems. Here, we considered the changes in g_s as an example of a plant adaptation response. Owing to sufficient water availability in the paddy field in the present study, the reduction in g_s induced by E was alleviated over generations, primarily as an adaptive response to maintain photosynthesis at relatively high levels. In another study, g_s was decreased further by maternal E relative to maternal A over 7 years in a desert, primarily as an adaptive response to conserve water (Grossman and Rice, 2014).

Unexpectedly, multigenerational exposure to E_g and E_a resulted in similar responses in almost all tested traits. The maternal environment is generally assumed to change seed quality (Potvin and Tousignant, 1996; Huxman et al., 1998; Lau et al., 2008). Such differentiation between E_g and E_a across generations did occur, such as seed nitrogen content (Yang et al., 2023b); however, the difference in seed quality did not explain the similar responsiveness. Nevertheless, we speculate that the mechanisms of E_g and E_a over multiple generations might be different. Under E_g in F5, plants were likely to adapt gradually to stepwise changes in CO₂ levels from one generation to the next; that is, owing to the smaller increase in CO₂ level, an increase of 40 μmol mol⁻¹ based on A could not result in a significant effect in the first generation (Liu et al., 2020). Likewise, an increase of 40 μmol mol⁻¹ based on A + 40 μmol mol⁻¹ could not result in a significant effect in the second generation either, and the adaption continued to the fifth generation, wherein the CO₂ level reached A + 200 μmol mol⁻¹. But under E_a in F5, plants could ‘remember’ the maternal environment and use environmental ‘memory’ to adapt when experiencing the recurrent environments (Bruce et al., 2007; Avramova, 2015; Liu et al., 2019); that is, when experiencing E_a in F1, such a large and rapid change in the CO₂ level altered photosynthetic and stomatal traits; however, when several subsequent generations were continuously grown in the same CO₂ environment, plants could ‘remember’ the maternal CO₂ environment and adjust to adapt. Such transgenerational ‘memory’ of the ancestral environment might result from epigenetic variation, such as DNA methylation, contributing to phenotype expression in offspring (Heard and Martienssen, 2014), as shown in a recent study suggesting that DNA methylation might contribute to plant adaptation to multigenerational exposure to future atmospheric CO₂ (Saban et al., 2020). Thus, DNA methylation for plants across generations requires further investigation in CO₂-enriched environments.

Our results under E_a in F1 (most experiments conducted) and E_g and E_a in F5 (actual cases in future climate) provide experimental evidence that E retains intergenerational effects on photosynthesis and stomatal features, whereas there were no multigenerational differences in the effects of E_g and E_a. Briefly, the responses to E_a in F1 were consistent with the previous consensus, such as the occurrence of photosynthetic acclimation, stimulation of photosynthesis and downregulation of photosynthetic physiological parameters and stomatal area. In contrast, E_g and E_a in F5 did not induce photosynthetic acclimation but stimulated greater photosynthesis and had little effect on photosynthetic physiology and stomatal traits. Hence, projecting future changes induced by E based on the physiological responses of contemporary plants could be misleading. Furthermore, the results from large and rapid environmental changes within a generation cannot accurately predict the effects of natural environmental changes on plants over multiple generations, especially in annual herbs with short life cycles.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Table S1: basic information on CO₂ concentration, temperature and photosynthetic photon flux density from 2016 to 2020. Table S2: the responses of dark-adapted chlorophyll fluorescence parameters to E_a in F1 and to E_g and E_a in F5. Fig. S1: the responses of stomatal conductance to intercellular CO₂ under E_a in F1 and E_g and E_a in F5 at the jointing and filling stages. Fig. S2: relationship of the response ratio of net photosynthetic rate against the response ratio of stomatal conductance at growth CO₂ concentration. Fig. S3: correlation between stomatal conductance and stomatal apparatus area under E_g and E_a in F5.

FUNDING

This work was supported by the National Natural Science Foundation of China (41530533 and 42171064) and the Strategic Priority Research Program of the Chinese Academy of Sciences, China (XDA26010103).