Effect of Rubisco Activase Deﬁciency on the Temperature Response of CO 2 Assimilation Rate and Rubisco Activation State: Insights from Transgenic Tobacco with Reduced Amounts of Rubisco Activase

The activation of Rubisco in vivo requires the presence of the regulatory protein Rubisco activase. To elucidate its role in maintaining CO 2 assimilation rate at high temperature, we examined the temperature response of CO 2 assimilation rate at 380 m L L 2 1 CO 2 concentration ( A 380 ) and Rubisco activation state in wild-type and transgenic tobacco ( Nicotiana tabacum ) with reduced Rubisco activase content grown at either 20 (cid:1) C or 30 (cid:1) C. Analyses of gas exchange and chlorophyll ﬂuorescence showed that in the wild type, A 380 was limited by ribulose 1,5-bisphosphate regeneration at lower temperatures, whereas at higher temperatures, A 380 was limited by ribulose 1,5-bisphosphate carboxylation irrespective of growth temperatures. Growth temperature induced modest differences in Rubisco activation state that declined with measuring temperature, from mean values of 76% at 15 (cid:1) C to 63% at 40 (cid:1) C in wild-type plants. At measuring temperatures of 25 (cid:1) C and below, an 80% reduction in Rubisco activase content was required before Rubisco activation state was decreased. Above 35 (cid:1) C, Rubisco activation state decreased slightly with more modest decreases in Rubisco activase content, but the extent of the reductions in Rubisco activation state were small, such that a 55% reduction in Rubisco activase content did not alter the temperature sensitivity of Rubisco activation and had no effect on in vivo catalytic turnover rates of Rubisco. There was a strong correlation between Rubisco activase content and Rubisco activation state once Rubisco activase content was less that 20% of wild type at all measuring temperatures. We conclude that reduction in Rubisco activase

The activation of Rubisco in vivo requires the presence of the regulatory protein Rubisco activase. To elucidate its role in maintaining CO 2 assimilation rate at high temperature, we examined the temperature response of CO 2 assimilation rate at 380 mL L 21 CO 2 concentration (A 380 ) and Rubisco activation state in wild-type and transgenic tobacco (Nicotiana tabacum) with reduced Rubisco activase content grown at either 20°C or 30°C. Analyses of gas exchange and chlorophyll fluorescence showed that in the wild type, A 380 was limited by ribulose 1,5-bisphosphate regeneration at lower temperatures, whereas at higher temperatures, A 380 was limited by ribulose 1,5-bisphosphate carboxylation irrespective of growth temperatures. Growth temperature induced modest differences in Rubisco activation state that declined with measuring temperature, from mean values of 76% at 15°C to 63% at 40°C in wild-type plants. At measuring temperatures of 25°C and below, an 80% reduction in Rubisco activase content was required before Rubisco activation state was decreased. Above 35°C, Rubisco activation state decreased slightly with more modest decreases in Rubisco activase content, but the extent of the reductions in Rubisco activation state were small, such that a 55% reduction in Rubisco activase content did not alter the temperature sensitivity of Rubisco activation and had no effect on in vivo catalytic turnover rates of Rubisco. There was a strong correlation between Rubisco activase content and Rubisco activation state once Rubisco activase content was less that 20% of wild type at all measuring temperatures. We conclude that reduction in Rubisco activase content does not lead to an increase in the temperature sensitivity of Rubisco activation state in tobacco.
The catalytic sites of Rubisco must be activated for CO 2 fixation to take place. This requires the carbamylation of a Lys residue at the catalytic sites to allow the binding of Mg 2+ and ribulose 1,5-bisphosphate (RuBP; Andrews and Lorimer, 1987). Rubisco activase facilitates carbamylation and the maintenance of Rubisco activity by removing inhibitors such as tight-binding sugar phosphates from Rubisco catalytic sites in an ATP-dependent manner (Andrews, 1996;Spreitzer and Salvucci, 2002;Portis, 2003;Parry et al., 2008). The activity of Rubisco activase is regulated by the ATP/ ADP ratio and redox state in the chloroplast (Zhang and Portis, 1999;Zhang et al., 2002;Portis, 2003).
In many plant species, Rubisco activation state decreases at high temperature in vivo (Crafts-Brandner and Salvucci, 2000;Salvucci and Crafts-Brandner, 2004b;Cen and Sage, 2005;Yamori et al., 2006b;Makino and Sage, 2007). However, it is unclear what the primary mechanisms underlying the inhibition of Rubisco activation are and whether Rubisco deactivation limits CO 2 assimilation rate at high temperature. It has been proposed that Rubisco activation state decreases at high temperature, because the activity of Rubisco activase is insufficient to keep pace with the faster rates of Rubisco inactivation at high temperatures (Crafts-Brandner and Salvucci, 2000;Crafts-Brandner, 2004a, 2004c;Kim and Portis, 2006). In in vitro assays using purified Rubisco and Rubisco activase, the activity of Rubisco activase was sufficient for the activation of Rubisco at the optimum temperature but not at high temperatures (Crafts-Brandner and Salvucci, 2000;Crafts-Brandner, 2004a, 2004c and Rubisco activase more readily dissociates into inactive forms at high temperature, causing a loss of Rubisco activase capacity (Crafts-Brandner and Law, 2000;Salvucci and Crafts-Brandner, 2004b). Moreover, the rates of inhibitor formation by misprotonation of RuBP during catalysis increased at higher temperatures (Salvucci and Crafts-Brandner, 2004c;Kim and Portis, 2006). CO 2 assimilation rates and plant growth were improved under heat stress in transgenic Arabidopsis expressing thermotolerant Rubisco activase isoforms generated by either geneshuffling technology (Kurek et al., 2007) or chimeric Rubisco activase constructs (Kumar et al., 2009). These results support the view that the reduction of Rubisco activase activity limits the Rubisco activation and, therefore, the CO 2 assimilation rates at high temperatures.
It has also been suggested that the decrease in CO 2 assimilation rate at high temperatures is caused by a limitation of RuBP regeneration capacity (e.g. electron transport capacity) rather than by Rubisco deactivation per se Wise et al., 2004;Cen and Sage, 2005;Makino and Sage, 2007;Kubien and Sage, 2008). These groups suggest that Rubisco deactivation at high temperature may be a regulatory response to the limitation of one of the processes contributing to electron transport capacities. For example, at high temperature, protons can leak through the thylakoid membrane, impairing the coupling of ATP synthesis to electron transport (Pastenes and Horton, 1996;Bukhov et al., 1999Bukhov et al., , 2000. As the electron transport capacity becomes limiting, ATP/ADP ratios and the redox potential of the chloroplast decline, causing a loss of Rubisco activase activity and, in turn, a reduction in the Rubisco activation state (Zhang and Portis, 1999;Zhang et al., 2002;Sage and Kubien, 2007). Based on this understanding, the decline in the Rubisco activation state at high temperature may be a regulated response to a limitation in electron transport capacity rather than a consequence of a direct effect of heat on the integrity of Rubisco activase.
Temperature dependence of CO 2 assimilation rate shows a considerable variation with growth temperature (Berry and Bjö rkman, 1980;Hikosaka et al., 2006;Sage and Kubien, 2007). Plants grown at low temperature generally exhibit higher CO 2 assimilation rates at low temperatures compared with plants grown at high temperature, but they exhibit lower rates at high temperature. Furthermore, both the temperature response of Rubisco activation state and the limiting step of CO 2 assimilation rate (a Rubisco versus RuBP regeneration limitation) have been shown to differ depending on growth temperature (Hikosaka et al., 1999;Onoda et al., 2005;Yamori et al., 2005Yamori et al., , 2006aYamori et al., , 2006bYamori et al., , 2008. This suggests that the regulation of Rubisco activation state could also differ in plants grown at different growth temperatures. Here, we analyzed the effects of Rubisco activase content on Rubisco activation state and CO 2 assimilation rate at leaf temperatures ranging from 15°C to 40°C in tobacco grown under two different temperature regimes (day/night temperatures of 20°C/15°C or 30°C/25°C). We used wild-type and transgenic tobacco with a range of reductions in Rubisco activase content to examine the dependence of Rubisco activation on Rubisco activase content over the range of leaf temperatures (Mate et al., 1993(Mate et al., , 1996.

Temperature Acclimation of Photosynthesis in the Wild Type
Growth temperature had a large effect on leaf properties. Leaf mass per area and contents of Rubisco,  3.73 6 0.14a 3.81 6 0.13 a 3.72 6 0.03 a 3.68 6 0.02 a 3.54 6 0.12 a 3.75 6 0.19 a cytochrome f, and chlorophyll were greater in wildtype plants grown at 20°C compared with 30°C. Contents of Rubisco and cytochrome f were 25.7% and 27.7% greater in 20°C-grown plants than in 30°Cgrown plants, respectively (Table I). Thus, the cytochrome f/Rubisco ratio did not change with growth temperature. On the other hand, Rubisco activase contents were similar irrespective of growth temperatures, such that the activase/Rubisco ratio was greater in 30°C-grown plants. Assuming molecular masses of 42 kD for the Rubisco activase monomer and 550 kD for Rubisco, the activase/Rubisco ratio (mol mol 21 ) averaged 0.91 and 1.32 in 20°C-and 30°Cgrown plants, respectively. Although the activase/ Rubisco ratio was greater in 30°C-grown plants than in 20°C-grown plants, Rubisco activation was slightly greater in 20°C-compared with 30°C-grown wild-type plants (Tables I and II; Fig. 1). The temperature response of CO 2 assimilation rate at 380 mL L 21 (A 380 ) and Rubisco activation state was different depending on growth temperature (Fig. 1). The optimum temperature of A 380 was higher in 30°Cgrown plants than in 20°C-grown plants. A 380 measured at 20°C in 20°C-grown plants was similar to A 380 measured at 30°C in 30°C-grown plants. This is known as the temperature homeostasis of photosynthesis and has been observed in other species (Hikosaka et al., 2006;Yamori et al., 2008). Rubisco activation state below 25°C was high and constant irrespective of growth temperatures, whereas above 30°C, Rubisco activation state decreased both in 20°C-grown plants and 30°C-grown plants. The extent of reduction in Rubisco activation state at high temperature was slightly greater in 20°C-grown plants than in 30°C-grown plants. Rubisco activation state at 40°C decreased by 17.1% 6 0.8% in 20°Cgrown plants and by 13.5% 6 1.1% in 30°C-grown plants (Table II).
Relationships between Rubisco Activase Content, CO 2 Assimilation Rate, and Rubisco Activation State Figure 2 shows the relationship between A 380 and Rubisco activation state and Rubisco activase content at 15°C, 25°C, and 40°C (for all measured leaf temperatures, see Supplemental Figs. S1 and S2; for the dependence of A 380 and Rubisco activation state on activase/Rubisco ratio, see Supplemental Fig. S3). Progeny of several primary transformants were used to generate the observed variation of activase content. At 15°C and 25°C, an 80% reduction in Rubisco activase content was required before A 380 or Rubisco activation state was decreased in both 20°C-and 30°Cgrown plants, whereas at 40°C, A 380 and Rubisco activation state decreased slightly, with a more modest decrease in Rubisco activase content, particularly in 20°C-grown plants. In both 20°C-and 30°C-grown  CO 2 assimilation rate at 1,500 mmol photons m 22 s 21 and 380 mL L 21 CO 2 concentration (A 380 ), the Rubisco activation state, the NADP-MDH activation state, and in vivo catalytic turnover rate of Rubisco were analyzed. The catalytic turnover rate of Rubisco was calculated from gross CO 2 assimilation rates (A 380 + dark respiration) and Rubisco carbamylated site contents. Dark respiration was measured after a 10-h dark period. Different letters show significant differences (Tukey-Kramer multiple comparison test; P , 0.05).   plants, A 380 and Rubisco activation state were drastically decreased when Rubisco activase content was less than 20% of wild-type levels, irrespective of leaf temperatures (Fig. 2).

20°C-grown plants
We classified plants into three groups with respect to Rubisco activase contents in 20°C-and 30°C-grown plants (Table I; (Table I). NADP-malate dehydrogenase (MDH) activation state, which is indicative of the redox status in the chloroplast (Scheibe and Stitt, 1988), was greater at 25°C compared with 40°C but was also similar between wild-type and antisense lines, irrespective of growth temperatures (Table II). Thus, there were no apparent differences between the leaf properties of wild-type and antisense lines, other than their Rubisco activase contents, that could confound interpretation of their temperature response of CO 2 assimilation and Rubisco activation state.
When comparing the three groups of plants, it is apparent that the temperature dependence of Rubisco activation is not very different, although there was a slightly greater decline in Rubisco activation state from 25°C to 40°C in plants with intermediate and low activase levels ( Fig. 3; Table II). The in vivo catalytic turnover rate of Rubisco, which was estimated from gross photosynthetic rate (A 380 + dark respiration) and the carbamylated active site content of Rubisco, increased as expected with temperature and was similar between the wild type and plants with intermediate Rubisco activase levels (Fig. 3). This   demonstrates that the differences in Rubisco activation state quantitatively accounted for differences in observed changes in CO 2 assimilation rate. In plants with low activase content, in vivo turnover rates were reduced, suggesting inhibitor binding to Rubisco catalytic sites (He et al., 1997).

Limiting
Step of the CO 2 Assimilation Rate CO 2 response curves of CO 2 assimilation rate (A) and chloroplast electron transport rates estimated from chlorophyll fluorescence (J f ) were measured to determine under what conditions A 380 was limited by Rubisco or RuBP regeneration (Fig. 4 for plants grown at 30°C/25°C; Supplemental Fig. S4 for plants grown at 20°C/15°C). Chloroplast CO 2 concentration (C c ) was calculated from intercellular CO 2 (C i ) as described in "Materials and Methods." A-C c responses were then fitted to the C 3 photosynthesis model (Farquhar et al., 1980). In the wild type and plants with intermediate activase content, the curves showed a transition from Rubisco-limited A (A c ) at lower CO 2 concentrations to RuBP regeneration-limited A (A r ) at higher CO 2 concentrations at all leaf temperatures irrespective of growth temperature (Fig. 4, A, B, D, and E). In plants with low activase content, CO 2 assimilation rate was always limited by Rubisco (Fig. 4, C and F).
The C c measured under C a = 380 mmol mol 21 was compared with the chloroplast CO 2 concentration at which the transition from Rubisco to RuBP regeneration limitation occurs (C trans ). C trans was estimated from two methods. First, the A-C c curve was analyzed for C trans (A-C c curve), based on the C 3 photosynthesis model (see Eq. 3 below; Fig. 4, A-C). Second, J f -C c curves were analyzed to obtain the chloroplast CO 2 concentration above which electron transport rate (J f ) is constant [C trans (J f -C c curve); Fig. 4, D-F]. C trans values estimated from these two independent methods were similar and increased with temperature (Fig. 5). The analysis showed that A 380 was limited by Rubisco above 25°C in 20°C-grown wild-type plants (Fig. 5A) and above 30°C in 30°C-grown wild-type plants (Fig.  5B). On the other hand, A 380 was limited by RuBP regeneration below 20°C in 20°C-grown wild-type plants (Fig. 5A) and below 25°C in 30°C-grown wildtype plants (Fig. 5B). The difference between the two growth temperatures was primarily due to differences in C c measured under C a = 380 mmol mol 21 rather than to differences in the relationship between C trans and temperature (Fig. 5).

Rubisco Activity Limits CO 2 Assimilation Rate at High But Not at Low Temperature in Tobacco
There has been controversy in the literature regarding what causes the decline in CO 2 assimilation rate at high temperature and the role that Rubisco activase plays in the regulation of CO 2 assimilation under these conditions. Our measurements of CO 2 response curves of CO 2 assimilation rate and chloroplast electron transport rates estimated from chlorophyll fluorescence (J f ) provided two independent means to assess the limitations and showed that in tobacco, A 380 was limited by RuBP regeneration only at low temperature, whereas at high temperature A 380 was limited by Rubisco (Fig. 5). Thus, decreases in Rubisco activation state reduce the potential of CO 2 assimilation rates at high temperature. For some species, Rubisco limitation has been observed at high temperature (Crafts-Brandner and Salvucci, 2000;Crafts-Brandner, 2004a, 2004c;Yamori et al., 2006aYamori et al., , 2006bYamori et al., , 2008; however, for other species, RuBP regeneration has been reported as the major limitation at high temperature Cen and Sage, 2005;Makino and Sage, 2007). There is considerable species variation in the limiting step of CO 2 assimilation rate (Yamori et al., 2010). Since A 380 at low temperature was limited by RuBP regeneration in 20°C-and 30°C-grown plants (Fig. 5), Rubisco activation state did not affect A 380 . Cen and Sage (2005) suggested that Rubisco activation state decreased at low temperature due to a regulatory feedback reflect-ing limitations in the RuBP regeneration capacity. However, some studies found little evidence for decreases in Rubisco activation state at low temperatures in spinach (Yamori et al., 2006b), rice (Oryza sativa; Makino and Sage, 2007), and tobacco (Kubien and Sage, 2008;this study). Therefore, at lower temperatures, Rubisco activation state is maintained at high levels in many plant species and does not limit CO 2 assimilation rate.
Tobacco plants acclimated to low growth temperature with an increase in leaf thickness and several photosynthetic components (Table I), but there was no change in the balance between Rubisco capacity and electron transport capacity, and this was also reflected by the fact that the C trans , the chloroplast CO 2 at which the transition from Rubisco to RuBP regeneration limitation occurs, was independent of growth temperature (Fig. 5). Interestingly, Rubisco activase contents did not vary with growth temperature. Thus, the activase/Rubisco ratio was greater in 30°C-grown plants than in 20°C-grown plants, but Rubisco activation was nevertheless slightly less in 30°C-grown plants (Table I; Fig. 1). The observed variation in Rubisco activation state with temperature was small (approximately 15%) and was also less than what has been observed for some other species (Salvucci and Crafts-Brandner, 2004b;Cen and Sage, 2005;Kim and Portis, 2005;Yamori et al., 2005). The fact that CO 2 assimilation rate in tobacco is Rubisco limited at high temperature makes it an ideal experimental system to examine the relationship between Rubisco activase content, Rubisco activation state, and CO 2 assimilation rate.

Regulation of Rubisco Activation State and CO 2 Assimilation Rate at High and Low Temperatures
The dependence of Rubisco activation state on Rubisco activase content at the various leaf temperatures ( Fig. 2; Supplemental Fig. S2) is very similar to previous observations in tobacco made at 25°C (Mate et al., 1996) and in Flaveria bidentis, a C 4 dicot (Hendrickson et al., 2008). That is, when Rubisco activase is present at levels less than 20% of the wild type, both Rubisco activation and CO 2 assimilation rate are severely reduced, but a 50% reduction in Rubisco activase levels has only minor effects on Rubisco activation, in particular at low temperature. Since there was a slight increase in the dependence on activase content above 35°C in the low-temperaturegrown plants, the role of Rubisco activase may be more important in 20°C-grown plants compared with 30°Cgrown plants.
It has been proposed that Rubisco activation state decreases at high temperature, because the activity of Rubisco activase is insufficient to keep pace with the faster rates of Rubisco inactivation at high temperatures (Crafts-Brandner and Salvucci, 2000;Crafts-Brandner, 2004a, 2004c;Kim and Portis, 2006), but the fact that the temperature dependence of Figure 5. Temperature dependence of chloroplast CO 2 concentration (C trans ) at which the transition from Rubisco (A c ) to RuBP regeneration (A r ) limitation occurs (black symbols) and chloroplast CO 2 concentration (C c ) for CO 2 assimilation rate measured at C a = 380 mmol mol 21 (white symbols) in 20°C-grown wild-type (WT; A) and 30°C-grown wild-type (B) plants. C trans was analyzed from two methods. First, A-C c curve was analyzed for C trans (A-C c curve; black circles), based on the C 3 photosynthesis model (Eq. 3). Second, J f -C c curve was analyzed to obtain the chloroplast CO 2 concentration at which electron transport rate (J f ) remains constant [C trans (J f -C c curve); black triangles]. C c for A 380 less than the C trans indicates that CO 2 assimilation is limited by A r , whereas C c for A 380 above C trans indicates that CO 2 assimilation is limited by A c (gray areas). Data represent means 6 SE; n = 4.
Rubisco activation state is not strongly linked to Rubisco activase content argues against this hypothesis (Figs. 2 and 3). The mode of action of Rubisco activase is to remove tight binding inhibitors from Rubisco uncarbamylated and carbamylated catalytic sites (Portis, 2003), and it has been shown that the rate of inhibitor formation increases with temperature (Salvucci and Crafts-Brandner, 2004a;Kim and Portis, 2006). In our experiments, the in vivo turnover rates (estimated from gross photosynthetic rate [A 380 + dark respiration] and the carbamylated active site content of Rubisco) did not differ between wild-type plants and plants with intermediate activase levels (Fig. 3). Thus, there was no evidence that CO 2 assimilation rate was reduced at high temperature due to inhibitor buildup at carbamylated sites. On the other hand, in plants with the low Rubisco activase content, in vivo turnover rate was reduced, particularly at high temperature ( Fig. 3; Table II), similar to what had previously been observed by He et al. (1997).
There appears to be no compelling link between Rubisco activase content and the temperature dependence of Rubisco activation state, and this raises the question of what regulates Rubisco activation state at high temperature. Several plant species express two activase isoforms, the longer (a) and shorter (b) forms (Spreitzer and Salvucci, 2002). Since the a-form is subjected to redox regulation via thioredoxin, reductions in ATP/ADP levels and the redox potential of the chloroplast could cause a down-regulation of activase activity and, in turn, a reduction in the activation state of Rubisco (Zhang and Portis, 1999;Spreitzer and Salvucci, 2002;Zhang et al., 2002). However, several plant species (e.g. tobacco, tomato [Solanum lycopersicum], and maize [Zea mays]) express only the nonredox-regulated b-form of activase (Salvucci et al., 1987;Qian and Rodermel, 1993). Nevertheless, in transgenic tobacco, which reduced amounts of cytochrome b/f complex, Rubisco activation state at 25°C changed with the redox status in the chloroplast stroma but not with the ATP/ADP ratio Ruuska et al., 2000). Thus, it appears that Rubisco activase activity in tobacco may also be regulated by the thioredoxin system by an as-yet-unknown mechanism. This is supported by our results that the NADP-MDH activation state was decreased at high temperature as well as Rubisco activation state (Table  II), as was also reported by Schrader et al. (2004). Moreover, Kim and Portis (2006) showed that low stromal Mg 2+ concentration reduced Rubisco activation state at high temperature. Moderate heat stress can have very large effects on thylakoid reactions in tobacco and induce increases in proton conductance and ion movement (Zhang et al., 2009). Thus, if in vivo Mg 2+ concentration is reduced at higher temperatures, it is also possible that a decrease in the Rubisco activation state could be due to a change in Mg 2+ concentration. These factors could also be the cause for the deactivation of Rubisco at high temperature. It should be noted that, even when A 380 at high temper-ature was limited by Rubisco and not by RuBP regeneration (Fig. 5), the levels of various regulatory metabolites (e.g. ATP and NADPH) and ionic concentration in stroma (e.g. Mg 2+ ) can be changed by the photosynthetic electron transport and membrane leakiness. We conclude that the decrease in Rubisco activation state at high temperature is not solely influenced by Rubisco activase activity.

CONCLUSION
At low temperature, Rubisco activation state was maintained at high levels and did not limit CO 2 assimilation rate in both 20°C-and 30°C-grown plants.
On the other hand, Rubisco activation state decreased at high temperature and limited CO 2 assimilation rate both in 20°C-and 30°C-grown plants. However, the temperature response of Rubisco activation was not strongly dependent on Rubisco activase content, suggesting that other processes also modulate Rubisco activation. Our results indicate that a selective enhancement of Rubisco activase capacity could partly enhance photosynthesis at high temperature, but this enhancement would generally be small.

Plant Materials and Growth Conditions
Tobacco (Nicotiana tabacum 'Wisconsin 38') plants and the progeny of several primary transformants of anti-activase tobacco (lines A36, A37, A41, and A49) were grown in controlled environmental growth cabinets (Mate et al., 1993(Mate et al., , 1996He et al., 1997). Plants were grown at irradiance of 200 mmol m 22 s 21 with a photoperiod of 12 h and CO 2 concentration of 3,000 mmol mol 21 . The day/night air temperatures were either 20°C/15°C or 30°C/25°C, and the relative humidity was 70%. Plants were grown in 5-L pots in garden mix containing approximately 2 g L 21 slow-release fertilizer (Osmocote; 15:4.8:10.8:1.2 nitrogen:phosphorus:potassium:magnesium and trace elements boron, copper, iron, manganese, molybdenum, and zinc; Scotts Australia) and watered daily. The low irradiance and high CO 2 were selected to minimize the differences in the growth rate of plants and the capacity of CO 2 assimilation at the growth condition.

Gas-Exchange and Fluorescence Measurements
CO 2 gas exchange of leaves was measured with a portable gas-exchange system (LI-6400, Li-COR). The whole portable gas-exchange system was enclosed in a temperature-controlled cabinet (Yamori et al., 2005). The CO 2 assimilation rate (A) versus intercellular CO 2 concentration (C i ) was measured at high light intensity of 1,500 mmol photons m 22 s 21 under several measurement temperatures. Then, the chloroplast CO 2 concentration (C c ) was estimated with an assumption that the mesophyll conductance to CO 2 diffusion ( g m ) at 25°C was 0.234 mol m 22 s 21 in 20°C-grown plants and 0.202 mol m 22 s 21 in 30°C-grown plants, based on the relationship between g m and A at 25°C (g m = 0.012 3 A; Evans and von Caemmerer, 1996). We used the temperature dependence of g m in tobacco (Bernacchi et al., 2002). Measurements of temperature response were initiated at 15°C, and leaf temperature was subsequently increased in 5°C intervals to 40°C. The leaf was allowed to equilibrate for at least 15 min before data were recorded at each measurement leaf temperature. The vapor pressure deficit was kept under 3.0 kPa even at the highest leaf temperature of 40°C.
Chlorophyll a fluorescence was also determined simultaneously with gas exchange during the temperature response measurements by an integrated fluorescence chamber head (LI-6400, and LI-6400-40 leaf chamber fluorometer; LI-COR). After measurements of the quantum yield of PSII (FPSII; Genty et al., 1989), the rate of linear electron transport (J f ) was determined as J f = FPSII 3 f 3 I 3 a leaf , where f is the fraction of absorbed light reaching PSII (assumed 0.5 for C 3 plants; Ogren and Evans, 1993), I is incident photon flux density, and a leaf is leaf absorptance. Leaf absorptance to the red and blue light-emitting diode light source of the LI-6400 was measured with an integrating sphere and spectroradiometer (LI-1800; LI-COR): for the wild type (30°C), 0.89 6 0.01; for antisense lines (30°C), 0.89 6 0.02; for the wild type (20°C), 0.90 6 0.01; for antisense lines (20°C), 0.89 6 0.01.

Analyses of the Limiting
Step of the Photosynthetic Rate at 380 mL L 21 A-C c curves were fitted with the C 3 photosynthesis model (Farquhar et al., 1980). When CO 2 assimilation is limited by the capacity of Rubisco to consume RuBP, the CO 2 assimilation rate (A c ) is expressed as: where V cmax (mmol m 22 s 21 ) is the maximum rate of Rubisco carboxylation on the leaf area basis, K c (mmol mol 21 ) and K o (mmol mol 21 ) are the K m values for CO 2 and O 2 , respectively, C c (mmol mol 21 ) and O (mmol mol 21 ) are chloroplastic CO 2 and O 2 concentrations, respectively, G* (mmol mol 21 ) is the CO 2 compensation point in the absence of day respiration, and R d (mmol m 22 s 21 ) is the day respiration rate. When CO 2 assimilation is limited by the RuBP regeneration rate, the CO 2 assimilation rate (A r ) is expressed as: where J g (mmol m 22 s 21 ) is the rate of electron transport. Temperature dependencies of Rubisco kinetics were obtained in tobacco from Bernacchi et al. (2002). Fitting was performed with the software Kaleidagraph (Synergy Software), and V cmax and J g were estimated from CO 2 assimilation rate at low CO 2 concentration and at high CO 2 concentration, respectively. The chloroplast CO 2 concentration at which the transition from Rubisco to RuBP regeneration limitation occurs (C trans ) was determined as: (von Caemmerer and Farquhar, 1981;. Moreover, the J f -C c curve was also analyzed to obtain the chloroplast CO 2 concentration at which electron transport rate (J f ) remains constant [C trans (J f -C c curve)], since the C trans (J f -C c curve) indicates the transition point from Rubisco to RuBP regeneration limitation (von Caemmerer and Farquhar, 1981). We analyzed which limits the CO 2 assimilation rate at 380 mmol mol 21 CO 2 concentration (A 380 ). The C c measured under C a = 380 mmol mol 21 was compared with C trans . If C c for A 380 is less than C trans , it indicates a limitation by RuBP regeneration, whereas if C c for A 380 is greater than C trans , it indicates a limitation by Rubisco. At C trans , a colimitation by RuBP regeneration and Rubisco exists.

Determinations of Total and Carbamylated Rubisco Active Sites
Samples used for the Rubisco activation assay were collected from a leaf equilibrated at steady-state conditions in the gas-exchange chamber. After gas exchange had reached the steady-state rate for at least 30 min at a given leaf temperature, the leaf in the chamber was taken and immediately frozen in liquid N 2 . Rubisco catalytic sites and Rubisco activation state were determined by the stoichiometric binding of [ 14 C]carboxy-arabinitol-P 2 , as described by Ruuska et al. (1998), using a CO 2 free extraction buffer containing 50 mM Bicine-NaOH buffer (pH 8.0), 2 mM MgCl 2 , 5 mM dithiothreitol (DTT), 2 mM EDTA, 1.5% (w/v) polyvinylpyrrolidone, and 1.5% (v/v) protease inhibitor cocktail (Sigma).

Determinations of NADP-MDH Activation State
The activation state of chloroplast NADP-MDH was assayed according to the method of Scheibe and Stitt (1988) and Ruuska et al. (2000) with minor modifications. Frozen leaf tissue (1.0 cm 2 ) was homogenized in ice-cold extraction buffer (bubbled with humidified nitrogen gas) consisting of 50 mM sodium acetate (pH 6.0), 10 mM MgSO 4 , 1 mM EDTA, 4.0 mM DTT, 0.1% (v/v) Triton X-100, 1.5% (w/v) polyvinylpyrrolidone, and 1.5% (v/v) protease inhibitor cocktail. The crude extract was centrifuged at 10,000g for 5 min at 4°C, and the initial activity of NADP-MDH was assayed immediately in assay buffer (bubbled with humidified nitrogen gas) consisting of 100 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM DTT, 0.2 mM NADPH, 2 mM oxaloacetic acid, and 50 mL of the supernatant. The decline in A 340 was monitored. To obtain the activity of the fully reduced enzyme, a portion of the supernatant was incubated in 250 mM Tris-HC1 (pH 9.0) and 125 mM DTT at room temperature for 20 min.
Determinations of Rubisco Activase, Cytochrome f, and Chlorophyll Immediately after the measurements of gas exchange, leaf discs of 0.5 cm 2 were taken, immersed in liquid nitrogen, and stored at 280°C until determinations of Rubisco activase, cytochrome f, and chlorophyll. The frozen leaf sample was ground in liquid nitrogen and homogenized in an extraction buffer containing 50 mM HEPES-KOH buffer (pH 7.8), 5 mM DTT, 10 mM MgCl 2 , 1 mM EDTA, 1.5% (w/v) polyvinylpyrrolidone, 0.1% (v/v) Triton X-100, and 1.5% (v/v) protease inhibitor cocktail. Rubisco activase was quantified by immunoblotting with anti-activase antibody (Mate et al., 1996). The cytochrome f content was quantified from two combination methods, by immunoblotting with anti-cytochrome f antibody (Baroli et al., 2008) and by the hydroquinone-reduced, ferricyanide-oxidized difference spectrum of the thylakoid membranes (Bendall et al., 1971;Yamori et al., 2005). For measurements of the hydroquinone-reduced, ferricyanide-oxidized difference spectrum, the difference spectrum at 554 nm was recorded with a dual-beam spectrophotometer (model 557; Perkin-Elmer). The millimolar extinction coefficient of 20 mM 21 cm 21 was used. Chlorophyll was extracted in 80% (v/v) acetone and determined by the procedure of Porra et al. (1989).

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. A 380 as a function of Rubisco activase contents.
Supplemental Figure S2. Rubisco activation state as a function of Rubisco activase contents.
Supplemental Figure S3. A 380 and Rubisco activation state as a function of the activase/Rubisco ratio.
Supplemental Figure S4. CO 2 response of CO 2 assimilation rate and electron transport rate in plants grown at 20°C/15°C.