Seasonal Changes in Temperature Dependence of Photosynthetic Rate in Rice Under a Free-air CO₂ Enrichment

Abstract

• Background and Aims Influences of rising global CO₂ concentration and temperature on plant growth and ecosystem function have become major concerns, but how photosynthesis changes with CO₂ and temperature in the field is poorly understood. Therefore, studies were made of the effect of elevated CO₂ on temperature dependence of photosynthetic rates in rice (Oryza sativa) grown in a paddy field, in relation to seasons in two years.

• Methods Photosynthetic rates were determined monthly for rice grown under free-air CO₂ enrichment (FACE) compared to the normal atmosphere (570 vs 370 µmol mol⁻¹). Temperature dependence of the maximum rate of RuBP (ribulose-1,5-bisphosphate) carboxylation (V_cmax) and the maximum rate of electron transport (J_max) were analysed with the Arrhenius equation. The photosynthesis–temperature response was reconstructed to determine the optimal temperature (T_opt) that maximizes the photosynthetic rate.

• Key Results and Conclusions There was both an increase in the absolute value of the light-saturated photosynthetic rate at growth CO₂ (P_growth) and an increase in T_opt for P_growth caused by elevated CO₂ in FACE conditions. Seasonal decrease in P_growth was associated with a decrease in nitrogen content per unit leaf area (N_area) and thus in the maximum rate of electron transport (J_max) and the maximum rate of RuBP carboxylation (V_cmax). At ambient CO₂, T_opt increased with increasing growth temperature due mainly to increasing activation energy of V_cmax. At elevated CO₂, T_opt did not show a clear seasonal trend. Temperature dependence of photosynthesis was changed by seasonal climate and plant nitrogen status, which differed between ambient and elevated CO₂.

INTRODUCTION

Global atmospheric CO₂ concentration has risen from approx. 280 µmol mol⁻¹ in pre-industrial times to approx. 370 µmol mol⁻¹ now and may reach 570 µmol mol⁻¹ by 2050. Most global climate models predict that global surface temperature will increase by 3 °C, associated with increasing greenhouse gas emissions (IPCC, 2001). Influences of increasing CO₂ and temperature on plant growth and ecosystem function have become a major area of concern in recent decades (Mitchell et al., 1995; Norby and Luo, 2004).

Photosynthesis, a key determinant of the rate of plant growth, is influenced by both CO₂ and temperature. Photosynthetic rates increase with a short-term increase in CO₂ concentration and are related parabolically to leaf temperature (von Caemmerer, 2000). These responses are mechanistically described by the biochemical model of photosynthesis (Farquhar et al., 1980). The model has two major parameters, the potential rate of electron transport (J_max) and the maximum rate of RuBP (ribulose-1,5-bisphosphate) carboxylation (V_cmax).

The model of Farquhar et al. (1980) has contributed substantially to modelling gas exchange rates of plants and terrestrial ecosystems under changing environments. However, many modelling studies have ignored the effects of growth conditions on photosynthetic characteristics (long-term response). Photosynthesis often shows down-regulation under a long-term increase in CO₂ concentration (CO₂ acclimation; Sage, 1994; Ziska et al., 1996; Seneweera et al., 2002; Ainsworth et al., 2003; Chen et al., 2005). In many species, a long-term increase in temperature leads to an increase in the optimal temperature for maximal photosynthetic rate (temperature acclimation; Slatyer et al., 1977; Berry and Björkman, 1980; Badger et al., 1982; Ferrar et al., 1989; Hikosaka et al., 1999; Hikosaka et al., 2006). Some recent studies have investigated responses in V_cmax and J_max to growth temperature (Hikosaka et al., 1999; Bunce, 2000; Hikosaka, 2005; Yamori et al., 2005) and to seasonal environment (Medlyn et al., 2002a; Han et al., 2004; Onoda et al., 2005b). However, no study, as far as we know, has investigated seasonal change in temperature dependence of V_cmax and J_max under elevated CO₂ concentrations. We have investigated the effects on photosynthetic rate and seasonal acclimation of rice leaves grown in the field under current and increased CO₂ concentrations.

Field-grown plants were exposed to natural diurnal, seasonal and year-to-year fluctuations in leaf temperature in a free-air CO₂ enrichment (FACE) system that raises atmospheric CO₂ concentration in the field with minimal artefacts (Long et al., 2004). Seasonal changes in photosynthetic characteristics were measured for two seasons. Temperature dependence of photosynthetic rates were analysed based on the model of Farquhar et al. (1980). Questions addressed here are: (1) does temperature dependence of photosynthesis change seasonally and, if so, how different is it between ambient and elevated CO₂? (2) What biochemical mechanisms are involved in the change in temperature dependence of photosynthesis? (3) Does growth temperature explain the seasonal change in the photosynthetic characteristics?

MATERIALS AND METHODS

The rice field was located at Shizukuishi in northern Honshu, Japan (39°38′N, 140°57′E, 200 m a.s.l.). Mean annual temperature and precipitation in 1976–2004 were 9·3 °C and 1540 mm, respectively. Elevated atmospheric CO₂ concentration (C_a) was created with a FACE system (Okada et al., 2001), consisting of octagonal 12-m diameter CO₂ emission structures (‘rings’) established within the paddy. The target C_a at the centre of the rings was 200 µmol mol⁻¹ above ambient CO₂. The experiment was conducted over two years (2003 and 2004). The seasonal averages of C_a in the ambient CO₂ plots and in the elevated CO₂ plots were 384 ± 14 and 606 ± 29 µmol mol⁻¹ in 2003, and 366 ± 12 and 548 ± 28 µmol mol⁻¹ in 2004, respectively. Two ambient CO₂ (X and Z) and two elevated CO₂ plots (B and D) were used (for description of these plots see Okada et al., 2001). Mean temperature and photosynthetic photon flux (PPF) during the experiment are shown in Table 1.

Table 1.

Climate conditions and leaf characteristics (order and age) for the photosynthetic measurements

Year	Measurement date	Tg (°C)	PPF (mol m−2 d−1)	Leaf order	Mean leaf age (d)
2003	18–24 June	17·7	34·9	8th	13
	15–25 July	18·5	25·3	11th	18
	20–30 August	20·9	23·7	14th	17
	11–15 September	19·1	22·0	15th	40
2004	23–28 June	18·1	36·9	9th	15
	21–26 July	21·7	22·6	12th	14
	21–23 August	21·9	31·3	13th	16
	14–18 September	19·7	24·9	14th	31

T_g and PPF are the mean daily growth temperature and mean daily photosynthetic photon flux, respectively, in the 2 weeks prior to measurements. Leaf order was numbered from the first leaf after germination.

Rice (Oryza sativa L. ‘Akitakomachi’) plants were grown following the agronomic techniques typical of the local area (Kobayashi et al., 2001; Anten et al., 2003; Kim et al., 2003). On 21 May 2003 and 20 May 2004, 25 d after emergence, seedlings were transplanted into paddies. Seedlings raised in a greenhouse under ambient CO₂ were planted in the ambient CO₂ plots, and those raised in another greenhouse under elevated CO₂ were planted in the elevated CO₂ plots. Distances between plants (‘hills’) and rows were 17·5 and 30 cm, respectively (equivalent to 19·1 hills m⁻²). The amounts of fertilizers added were: 8 g N m⁻² (25 % ammonium sulfate and 75 % LP-70) on 16 May in both 2003 and 2004, 30 g P₂O₅ m⁻² on 24 April 2003 and 19 April 2004, 15 g K₂O m⁻² on 25 April 2003 and 19 April 2004, respectively.

Photosynthetic measurements were made on the most recently fully expanded leaves in the experimental periods (leaf order and leaf age after emergence are given in Table 1). Photosynthetic rates were measured using an open gas exchange system (Model LI-6400, LiCor Inc., Lincoln, NE, USA), with an LED light source (LI-6400-02B, LiCor) and a dual Peltier device to regulate the PPF and temperature in the chamber (3 × 2 cm²).

Measurements were replicated using at least three leaves in each plot. CO₂ response curves of photosynthesis were determined at approx. 20, 25, 30 and 35 °C leaf temperature at PPF >1800 µmol m⁻² s⁻¹. The vapour pressure deficit (VPD) was kept at <1·5 kPa for 15–30 °C, and 1·5–2·5 kPa for 35 °C. Leaves were allowed to equilibrate for 5–10 min at each new temperature before measurement. For each CO₂ response curve, photosynthesis was first measured at the growth CO₂ concentration (ambient CO₂, 370 µmol mol⁻¹ or elevated CO₂, 570 µmol mol⁻¹; P_growth), and then the C_a was increased in eight steps from 50 to 1500 µmol mol⁻¹. In moving to a new CO₂ concentration, sufficient time was given (>5 min) to allow a steady-state to be attained prior to measurement of the photosynthetic rate. Immediately after gas exchange measurements were completed, the leaf was detached and 3-cm-long segments were excised (excluding the tip and base) and their width measured for calculation of area with an absolute digimatic caliper (Mitutoyo, CD-S15C, Kanagawa, Japan). The dry mass of leaf segments was determined after oven-drying at 70 °C for >72 h, and then the nitrogen content was determined using an NC analyser (Sumigraph NC-80,Shimadzu, Kyoto, Japan).

Models

The photosynthesis curve plotted against intercellular CO₂ concentration (A–C_i curve) was analysed to determine the maximum rate of RuBP carboxylation (V_cmax) and the maximum rate of electron transport (J_max) using the biochemical model of photosynthesis (Farquhar et al., 1980). When ribulose-1,5-bisphosphate (RuBP) is saturated, the photosynthetic rate is determined by:

(1)

where P_c is the photosynthetic rate limited by the Rubisco activity, C_i is the concentration of CO₂ at intercellular space, Γ^* is the CO₂ compensation point in the absence of day respiration (R_d), K_c and K_o are Michaelis constants of RuBP carboxylase for CO₂ and O₂, respectively, and O is the O₂ concentration. When RuBP regeneration limits photosynthesis, the photosynthetic rate is expressed as:

(2)

where P_r is the photosynthetic rate limited by RuBP regeneration. The photosynthetic rate is the minimum of P_c and P_r.

The temperature dependence of kinetic parameters is described by the Arrhenius equation (Harley and Tenhunen, 1991; Bernacchi et al., 2001):

(3)

where fis the value of a parameter. f(25) is f at 25 °C, E_a is the activation energy, R is the gas constant (8·314 J mol⁻¹ K⁻¹) and T_k is leaf temperature in K.

We calculated values of K_c using eqn (3), where K_c at 25 °C and E_a of K_c were assumed to be 404·9 µmol mol⁻¹ and 79·43 kJ mol⁻¹, respectively. Similarly, K_o and Γ^* values were calculated assuming that K_o and Γ^* at 25 °C were 278·4 mmol mol⁻¹ and 42·8 µmol mol⁻¹, and E_a of K_o and E_a of Γ^* were 36·38 kJ mol⁻¹ and 37·83 kJ mol⁻¹, respectively (Bernacchi et al., 2001). Using the calculated K_c, K_o and Γ^* values, eqn (1) was fitted to the C_i-response curves of photosynthesis at a lower range of CO₂ (C_i < 300 µmol mol⁻¹). R_d was assumed to be 0·02 of V_cmax (von Caemmerer, 2000). J_max was calculated by fitting eqn (2) to a higher range of CO₂ (C_i > 600 µmol mol⁻¹). E_a of V_cmax and of J_max were obtained from pooled data for each plot as a regression coefficient (eqn 3). Curve fitting was performed with Kaleida graph (Synergy Software, Reading, PA, USA).

Statistical analysis

Data are presented as means ± s.e. Statistical tests were performed using SPSS 7·5·1 statistical software (SPSS Inc., Chicago, IL, USA). ANOVA (split-plot) was conducted to test the effects of year (main plot), CO₂ (subplot), months (sub-subplot) and their interactions on photosynthetic characteristics. Student's t-test was used for the effect of the CO₂ treatments.

RESULTS

The mean daily temperature (T_g) and mean daily PPF during a 2-week period prior to each measurement (Table 1) are considered as the ‘growth environment’ for the leaves; they varied seasonally. T_g was highest in August and lowest in June in both 2003 and 2004. When compared for the same month, T_g was slightly higher in 2004. PPF was highest in June and lowest in September in 2003 and in July in 2004.

Effects of elevated CO₂ and seasonal environment on photosynthetic characteristics

Seasonal changes in temperature dependence of the light-saturated photosynthetic rates per unit leaf area (P_growth) determined at the growth CO₂ concentration (Fig. 1) tended to increase to a maximum with increasing leaf temperature, and then either remained constant or decreased with further increase in leaf temperature. At any given temperature and month, P_growth was higher in leaves grown at elevated CO₂. P_growth decreased as the growing season progressed (Table 2).

Fig. 1.

Temperature dependence of the light-saturated rate of photosynthesis (P_growth) for rice (Oryza sativa) grown at (A, C) ambient CO₂ (370 µmol mol⁻¹, open symbols) and (B, D) at elevated CO₂ (570 µmol mol⁻¹, closed symbols) in 2003 (A, B) and 2004 (C, D). Measurements were made in June (circles), July (squares), August (triangles) and September (diamonds). Data from two plots are pooled and presented as mean ± s.e. (n = 6).

Table 2.

Summary of analysis of variance (ANOVA, presented as F-values) for the effects of month, CO₂, year, and their interactions on the photosynthetic rate at 25 °C under growth CO₂ concentration (P_growth25), leaf nitrogen content (N_area), stomatal conductance at 25°C (g_s), intercellular CO₂ concentration at 25 °C (C_i), the maximum rate of electron transport at 25 °C (J_max25), the maximum rate of RuBP carboxylation at 25 °C (V_cmax25), the J_max to V_cmax ratio at 25 °C (J/V), activation energy of J_max (E_aj), activation energy of V_cmax (E_av), and optimum temperature of photosynthesis predicted by the model (T_opt)

Source of variation	d.f.	Pgrowth25	Narea	gs	Ci	Jmax25	Vcmax25	J/V	Eaj	Eav	Topt
Year	1	10·11	0·23	19·92*	39·51*	22·58*	14·45	7·06	0·22	2·03	13·44
Main plot error	2	11·74	2·75	6·56	1·48	2·67	7·10	0·59	1·08	3·08	1·44
CO2	1	462·73**	4·88	25·45*	1215·25***	3·19	26·71*	3·72	11·16	4·01	78·87*
CO2 × Year	1	17·19	0·41	12·83	5·44	0·19	7·61	1·88	0·46	5·11	4·16
Subplot error	2	0·10	2·28	0·18	1·72	0·83	0·49	1·25	0·97	0·73	0·80
Month	3	216·26***	175·72***	4·09*	19·87***	229·24***	307·07***	22·33***	0·83	6·41**	3·74*
Month × CO2	3	3·50*	6·72**	0·54	1·61	1·93	2·89	0·23	0·65	0·28	1·34
Month × Year	3	11·01***	18·63***	4·64*	3·21	6·56**	7·91**	2·22	1·79	5·10*	5·66*
CO2 × Month × Year	3	1·03	0·26	0·68	0·59	0·86	1·88	0·41	7·71**	10·33**	5·27*
Sub-subplot error	12

P_growth25, N_area, g_s and C_i are measured values. J_max25, V_cmax25, J/V, E_aj, E_av and T_opt are calculated values. Significance levels: ^***, P < 0·001; ^**, P < 0·01; ^*, P < 0·05.

Stomatal conductance (g_s), determined at 25 °C, was lower in leaves grown at elevated CO₂ (Tables 2, 3). It differed significantly between months, although no seasonal trend was observed. The average intercellular CO₂ concentration (C_i) at 25 °C was 79·9 % of C_a at ambient CO₂ and 80·7 % of C_a at elevated CO₂, and increased during the growing season in both ambient and elevated CO₂ (Tables 2, 3). Leaf nitrogen content per unit area (N_area) was not affected by CO₂ during growth, but declined during the growing season irrespective of CO₂ treatment (Tables 2, 3).

Table 3.

Stomatal conductance for water vapour (g_s) and intercellular CO₂ concentration (C_i) at 25 °C under growth CO₂ concentration, and leaf nitrogen content per unit area (N_area) for rice grown at ambient CO₂ (370 µmol mol⁻¹) and at elevated CO₂ (570 µmol mol⁻¹)

		gs (mol m−2 s−1)		Ci (µmol mol−1)		Narea (g m−2)
Year	Measurement date	Ambient	FACE	Ambient	FACE	Ambient	FACE
2003	18–24 June	0·23 ± 0·05	0·23 ± 0·07	268 ± 16	435 ± 30***	1·91 ± 0·26	1·95 ± 0·17
	15–25 July	0·27 ± 0·09	0·16 ± 0·03*	283 ± 13	420 ± 15***	1·89 ± 0·19	1·78 ± 0·14
	20–30 August	0·28 ± 0·07	0·18 ± 0·03**	282 ± 12	441 ± 20***	2·12 ± 0·18	1·73 ± 0·14**
	11–15 September	0·27 ± 0·09	0·20 ± 0·07	298 ± 17	449 ± 30***	1·33 ± 0·07	1·25 ± 0·12
2004	23–28 June	0·34 ± 0·05	0·32 ± 0·04	292 ± 3	464 ± 9***	2·17 ± 0·17	2·18 ± 0·11
	21–26 July	0·32 ± 0·06	0·30 ± 0·04	293 ± 12	458 ± 18***	1·95 ± 0·23	1·94 ± 0·08
	21–23 August	0·41 ± 0·07	0·44 ± 0·10	318 ± 4	499 ± 10***	1·88 ± 0·10	1·60 ± 0·15**
	14–18 September	0·25 ± 0·03	0·20 ± 0·03*	329 ± 6	509 ± 11***	0·97 ± 0·08	0·96 ± 0·06

Means ± s.e. (n = 6) are shown. Asterisks indicate significant differences between CO₂ treatments: ^***, P < 0·0001; ^**, P < 0·01; ^*, P < 0·05.

J_max and V_cmax determined at 25 °C (J_max25 and V_cmax25, respectively) decreased during the season (Fig. 2A–D). Since the decrease in V_cmax25 was greater than that in J_max25, the J_max/V_cmax ratio increased (Fig. 2E, F). There was a significant effect of CO₂ on V_cmax25 (Table 2), but was not on J_max25. V_cmax25 tended to be lower at elevated CO₂ (Fig. 2C, D).

Fig. 2.

Seasonal change in (A, B) the maximum rate of electron transport (J_max25), (C, D) the maximum rate of caboxylation (V_cmax25), and (E, F) the J_max25/V_cmax25 ratio at 25 °C for rice (Oryza sativa) grown at ambient CO₂ (370 µmol mol⁻¹, open symbols) and at elevated CO₂ (570 µmol mol⁻¹, closed symbols) in 2003 (A, C, E) and 2004 (B, D, F). Data from two plots are pooled and shown as mean ± s.e. (n = 6). Asterisks indicate significant differences between CO₂ treatments at P < 0·05.

J_max and V_cmax increased exponentially with leaf temperature: an example is shown in Fig. 3, with the curve fitted using the Arrhenius equation. The activation energy (E_a) is a measure of temperature dependence of photosynthetic rate. Since deactivation at high temperatures was not observed for either J_max or V_cmax, we did not use a model characterized by an optimum (peak) (Medlyn et al., 2002a, b). ANOVA suggested that the activation energy of J_max (E_aj) was not different between leaves grown in different CO₂ concentrations (Table 2). However, the seasonal change in E_aj was not consistent across years and CO₂ conditions (Fig. 4). For example, at elevated CO₂, E_aj increased seasonally in 2003 (P = 0·004, Fig. 4A), while it decreased in 2004 (P = 0·003, Fig. 4B). E_av was not affected by growth CO₂ but was significantly different among months (Table 2, Fig. 4C, D).

Fig. 3.

An example of the temperature dependence of the maximum rate of electron transport (J_max) and the maximum rate of caboxylation (V_cmax) of rice (Oryza sativa) grown at elevated CO₂ (570 µmol mol⁻¹) in June 2004. Data are fitted with the Arrhenius equation (eqn. 3). Activation energy of J_max (E_aj) was 19·42 kJ mol⁻¹ and activation energy of V_cmax (E_av) was 52·93 kJ mol⁻¹.

Fig. 4.

(A, B) Seasonal change in the activation energy of J_max (E_aj), and (C, D) activation energy of V_cmax (E_av) for rice grown at ambient CO₂ (370 µmol mol⁻¹) and at elevated CO₂ (570 µmol mol⁻¹) in 2003 (A, C) and 2004 (B, D). One point denotes one plot.

Modelling of temperature dependence of photosynthetic rate at growth CO₂ conditions

Using the above parameters C_i, J_max, V_cmax, E_aj and E_av, we reconstructed the temperature dependence of photosynthetic rate at the CO₂ concentrations during leaf growth. There was a strong correlation between measured and estimated rates of photosynthesis (y = 0·95x; r = 0·97, P < 0·0001) with the regression was very close to the 1 : 1 line (Fig. 5). This indicates that the present photosynthesis model gave a fairly good quantitative description of the effect of years and CO₂ on growth. At ambient CO₂, photosynthesis at optimal temperature was limited by P_c in both years, while at elevated CO₂ photosynthesis at optimal temperature was limited by P_r in earlier stages (June, July and August), and by P_c in the latest stage (September; data not shown). Temperature dependence of (relative) photosynthetic rate differed between ambient and elevated CO₂ (Fig. 6). The optimal temperature of photosynthesis (T_opt, the value where the photosynthetic rate was maximum) was significantly higher at elevated CO₂ (Table 2): it ranged from 22 to 34·5 °C with an average value of 28·9 °C at ambient CO₂, and from 29·5 to 37 °C with an average value of 33·5 °C at elevated CO₂. Temperature dependence of photosynthesis also showed a large seasonal change. There was a significant effect of month on T_opt (Table 2).

Fig. 5.

Comparison between measured and estimated photosynthesis of plants grown at ambient CO₂ (370 µmol mol⁻¹, open symbols) and at elevated CO₂ (570 µmol mol⁻¹, closed symbols) in 2003 (circles) and 2004 (squares). The solid line is the regression y = 0·95x (r = 0·97, P < 0·0001). The broken line indicates equivalence between measured and estimated photosynthesis.

Fig. 6.

Modelled temperature dependence of photosynthesis at (A, C) ambient CO₂ (370 µmol mol⁻¹), and (B, D) elevated CO₂ (570 µmol mol⁻¹)in 2003 (A, B) and 2004 (C, D) growing seasons (June, July, August and September). Values are normalized to 1 at 25 °C. Circles indicate the maximum photosynthesis.

Relationship between growth temperature (T_g) and photosynthetic characteristics

There was no significant difference in E_av between leaves grown at the two CO₂ concentrations (Table 2). E_av was positively correlated with T_g across both CO₂ concentrations (Fig. 7A, P = 0·025). However, E_aj was not correlated with T_g (data not shown). There was a significant correlation between T_opt and T_g at ambient CO₂ (P = 0·018), but not at elevated CO₂ (P = 0·122; Fig. 7B).

Fig. 7.

Relationship between growth temperature (T_g, the mean daily temperature in the 2 weeks prior to measurements) and photosynthetic characteristics: (A) activation energy of V_cmax (E_av) and (B) the optimal temperature of photosynthesis at growth CO₂ (T_opt). Plants were grown at ambient CO₂ (370 µmol mol⁻¹, open symbols) and at elevated CO₂ (570 µmol mol⁻¹, closed symbols) in 2003 (circles) and 2004 (squares). Regression lines: (A) y = 23·96 + 1·98x (r = 0·40, P = 0·025) and (B) y = 4·86 + 1·21x (r = 0·58 P = 0·018, ambient), and y = 20·92 + 0·64x (r = 0·40, P = 0·122, FACE). One point denotes one plot.

DISCUSSION

The biochemical model of photosynthesis developed by Farquhar et al. (1980) is useful for predicting carbon exchange by plants under global environmental change because it represents a mechanism for the effects of elevated CO₂ on photosynthetic rates. The model is also useful for analysing temperature dependence of photosynthesis. As photosynthesis–temperature curves are parabolic with a broad peak, many data points are needed to obtain the optimal temperature (e.g. Cunningham and Read, 2002). However, as most of the parameters in the model of Farquhar et al. (1980) follow the Arrhenius equation, it is possible to describe photosynthetic response to temperature with a relatively small number of data points. The similarity between measured and estimated values (Fig. 5) suggests that the model gave a fairly good quantitative description of photosynthetic rates. Several studies have shown changes in temperature dependence of model parameters (such as E_av and E_aj) under a seasonal environment (Medlyn et al., 2002a; Han et al., 2004), but information is still insufficient when large acclimational change and interspecific differences are considered (Leuning, 2002; Medlyn et al., 2002b). The present study is the first report showing a seasonal change in temperature dependence of photosynthetic parameters at elevated CO₂.

Absolute photosynthetic rate (P_growth)

Elevated CO₂ significantly increased P_growth (Fig. 1). This is simply ascribed to higher C_i (Table 3). However, a slight but significant decrease in V_cmax25 at elevated CO₂ (Fig. 2, Table 2) partly offset the effect of increased C_i. This down-regulation may be caused by sugar accumulation (Rey and Jarvis, 1998; Seneweera et al., 2002; Rogers et al., 2004) or by accelerated leaf senescence with advanced plant development (Rogers et al., 1996; Ludewing and Sonnewald, 2000; von Caemmerer et al., 2001; Seneweera et al., 2002).

At both CO₂ concentrations, P_growth25 (P_growth at 25 °C) decreased as the plants grew (Fig. 1), consistent with previous studies for rice (Hasegawa et al., 1996 for ambient CO₂; Seneweera et al., 2002). This is attributed to the seasonal decrease in J_max and V_cmax (Fig. 2), which is associated with the reduction in N_area (Table 3). Seasonal reduction in N_area may be related to plant ontogeny rather than environmental change. As plant mass increases, nutrient supply from the soil may become relatively insufficient, leading to a nitrogen deficiency in the plant body. In the later stages of the life cycle, reallocation of nitrogen to reproductive organs may also decrease nitrogen in vegetative parts. Mae and Ohira (1981) showed that about half of the nitrogen in vegetative organs was retranslocated to reproductive organs in rice.

Temperature dependence of photosynthesis

The optimal temperature of the photosynthetic rate determined by the model (T_opt) was higher at elevated than at ambient CO₂ (Figs 6, 7B). In earlier stages (June, July and August), this is attributed to the difference in the limiting step of photosynthesis: photosynthesis at T_opt was limited by P_c at ambient CO₂ and by P_r at elevated CO₂ (data not shown). In many species, P_c has a lower optimal temperature than P_r (Kirschbaum and Farquhar, 1984; Hikosaka, 1997; Hikosaka et al., 1999; Onoda et al., 2005b). This is because the increase in the carboxylation rate with increasing temperature is partly offset by the increase in photorespiration rate (Kirschbaum and Farquhar, 1984). In September, on the other hand, photosynthesis at T_opt was limited by P_c at both CO₂ concentrations. The increase in T_opt at elevated CO₂ is thus attributed to the effect of C_i on temperature dependence of P_c, which is directly influenced by the balance between carboxylation and photorespiration. Increasing CO₂ concentration decreases the contribution of photorespiration, which makes photosynthesis more temperature-dependent and increases the optimal temperature (Kirschbaum and Farquhar, 1984; Long, 1991).

T_opt showed a significant difference between months (Table 2). This may be partly explained by the increase in growth temperature at ambient CO₂ (Fig. 7B). Since P_c limited photosynthesis at ambient CO₂, the change in T_opt was attributable to the change in E_av. According to the model, an increase in E_av by 10 kJ mol⁻¹ leads to an increase in the optimal temperature of P_c by 5·4 °C (Hikosaka et al., 2006). E_av actually increased with growth temperature (Fig. 7A), which was consistent with other studies (Hikosaka et al., 1999; Onoda et al., 2005b; Yamori et al., 2005). The increase in E_av with increasing growth temperature is a common response in C₃ species (Hikosaka et al., 2006).

In contrast, at elevated CO₂, T_opt showed neither a clear seasonal trend nor a dependence on T_g (Fig. 7B). This is because E_aj did not change with time (Table 2). In many species the temperature dependence of J_max changes with growth temperature (Armond et al., 1978; Badger et al., 1982; Hikosaka et al., 1999; Ziska, 2001; Yamasaki et al., 2002), but in some species it does not (Sage et al., 1995). The difference in T_g of less than 5 °C in our study might have been too small to detect a significant change in E_aj, or alternatively E_aj of rice was not affected by growth temperature. On the other hand, different seasonal trends in the dependence of E_aj between ambient and elevated CO₂ and between the two years (Fig. 4) suggest that factors other than temperature are involved in the change in E_aj.

At elevated CO₂ the limiting step of photosynthesis changed between the early (June, July and August) and the late stage (September), caused by a higher J_max/V_cmax ratio in September. We found a positive correlation between the J_max/V_cmax ratio and T_g (data not shown). However, this result is inconsistent with earlier studies: in some species the J_max/V_cmax ratio increased at low temperature (Hikosaka et al., 1999; Hikosaka, 2005; Onoda et al., 2005a; Yamori et al., 2005) but in others it did not with growth temperature (Bunce, 2000; Hikosaka and Hirose, 2001; Medlyn et al., 2002a; Onoda et al., 2005b). For rice grown under controlled conditions, the J_max/V_cmax ratio was higher at low temperatures (Makino et al., 1994). The inconsistency between earlier studies and ours may be caused by an alteration of the J_max/V_cmax ratio due to factor(s) other than T_g. Seneweera et al. (2002) found that flag leaves of rice had a lower V_cmax per unit Rubisco than earlier leaves. A decrease in internal conductance of CO₂ diffusion may be involved in the seasonal change in the J_max/V_cmax ratio (von Caemmerer, 2000; Onoda et al., 2005b).

CONCLUSIONS

There was an increase in the absolute value of P_growth and in the optimal temperature of the P_growth–temperature curve caused by elevated CO₂ concentration during growth and seasonal environment. Seasonal decrease in P_growth was associated with decrease in nitrogen status with plant growth, which decreased N_area and thus J_max and V_cmax. The seasonal change in the T_opt differed between the two CO₂ concentrations. At ambient CO₂, T_opt increased with increasing growth temperature due mainly to increasing activation energy of V_cmax. At elevated CO₂, T_opt did not show clear seasonal changes. This was partly caused by the seasonal increase in the J_max/V_cmax ratio. Thus, the temperature dependence of photosynthesis was influenced by seasonal environment and reduction in nitrogen with plant growth, which was different between ambient and elevated CO₂.