Effects of Rubisco kinetics and Rubisco activation state on the temperature dependence of the photosynthetic rate in spinach leaves from contrasting growth temperatures

Authors

  • WATARU YAMORI,

    Corresponding author
    1. Department of Biology, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka 560-0043,
      Wataru Yamori. Fax: +81 6 6850 5808; e-mail: wataru-y@bio.sci.osaka-u.ac.jp
    Search for more papers by this author
  • KENSAKU SUZUKI,

    1. Department of Biology and Environmental Sciences, National Agricultural Research Center for Tohoku Region, Shimo-Kuriyagawa, Morioka, Iwate 020-0198, Japan and
    Search for more papers by this author
  • KO NOGUCHI,

    1. Department of Biology, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka 560-0043,
    Search for more papers by this author
  • MASATO NAKAI,

    1. Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan
    Search for more papers by this author
  • ICHIRO TERASHIMA

    1. Department of Biology, Graduate School of Science, Osaka University, 1-1 Machikaneyama-cho, Toyonaka, Osaka 560-0043,
    Search for more papers by this author

Wataru Yamori. Fax: +81 6 6850 5808; e-mail: wataru-y@bio.sci.osaka-u.ac.jp

ABSTRACT

Recently, several studies reported that the optimum temperature for the initial slope [IS(Ci)] of the light-saturated photosynthetic rate (A) versus intercellular CO2 concentration (Ci) curve changed, depending on the growth temperature. However, few studies compare IS(Ci) with ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) properties. Here, we assessed Rubisco activation state and in vitro Rubisco kinetics, the main determinants of IS(Ci), in spinach leaves grown at 30/25 [high temperature (HT)] and 15/10 °C [low temperature (LT)]. We measured Rubisco activation state and A at a CO2 concentration of 360 µL L−1 (A360) at various temperatures. In both HT and LT leaves, the Rubisco activation state decreased with increasing temperatures above the optimum temperatures for A360, while the activation state remained high at lower temperatures. To compare Rubisco characteristics, temperature dependences of the maximum rate of ribulose 1,5-bisphosphate (RuBP) carboxylation (Vcmax), specificity factor (Sc/o) and thermal stability were examined. We also examined Vcmax and thermal stability in the leaves that were transferred from HT to LT conditions and were subsequently kept under LT conditions for 2 weeks (HL). Rubisco purified from HT, LT and HL leaves are called HT, LT and HL Rubisco, respectively. Thermal stabilities of LT and HL Rubisco were similar and lower than that of HT Rubisco. Both Vcmax and Sc/o in LT Rubisco were higher than those of HT Rubisco at low temperatures, while these were lower at high temperatures. Vcmax in HL Rubisco were similar to those of LT Rubisco at low temperatures, and to those of HT Rubisco at high temperatures. The predicted photosynthetic rates, taking account of the Rubisco kinetics and the Rubisco activation state, agreed well with A360 in both HT and LT leaves. This study suggests that photosynthetic performance is largely determined by the Rubisco kinetics at low temperature and by Rubisco Kinetics and the Rubisco activation state at high temperature.

INTRODUCTION

Most plants show considerable capacity to adjust their photosynthetic characteristics to their growth temperatures. Such plastic adjustments allow the plants to perform efficient photosynthesis at their respective growth temperatures (Berry & Björkman 1980; Yamori, Noguchi & Terashima 2005).

When compared on a leaf area basis, plants grown at low temperatures tend to have greater amounts of ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) and other enzymes for photosynthetic carbon metabolism than those grown at higher temperatures (Badger, Björkman & Armond 1982; Holaday et al. 1992; Hurry et al. 1995; Strand et al. 1999). Hence, the capacity of the enzymatic reactions at low temperatures increases. However, in such plants, photosynthetic performance at high temperatures is usually reduced. Thus, the temperature response curve shifts towards the lower temperature.

At moderately high temperatures, the low Rubisco activation state would be the primary cause of the decrease in the photosynthetic rate (Crafts-Brandner & Salvucci 2000, 2002; Sharkey et al. 2001; Salvucci & Crafts-Brandner 2004a,b; Haldimann & Feller 2004, 2005; Kim & Portis 2005). Because the Rubisco activation state is regulated by Rubisco activase, a nuclear-encoded stromal enzyme (for recent reviews, see Portis 2001, 2003; Spreitzer & Salvucci 2002), the decrease in Rubisco activation state at moderately high temperatures can be attributed to suppressed Rubisco activase activity (Salvucci & Crafts-Brandner 2004a,b). When the net photosynthetic rate and chlorophyll (Chl) fluorescence were examined at the same time with Antarctic hair grass (Deschampsia antarctica, Salvucci & Crafts-Brandner 2004b), creosote bush (Larrea tridentate, Salvucci & Crafts-Brandner 2004b), cotton (Gossypium hirsutum, Crafts-Brandner & Law 2000), oak (Quercus pubescens, Haldimann & Feller 2004), pea (Pisum sativum L., Haldimann & Feller 2005), spinach (Spinacia oleracea, Weis 1981) and strawberry tree (Arbutus unedo, Bilger, Schreiber & Lange 1987), the electron transport capacities were unaffected at moderately high temperatures where CO2 fixation rates decreased. The insensitivity of the electron transport capacity to moderately high temperatures was also supported by the measurements of metabolite levels with cotton and wheat (G. hirsutum and Triticum aestivum, Kobza & Edwards 1987; Law & Crafts-Brandner 1999; Crafts-Brandner & Law 2000). These results support the view that the suppression of Rubisco activase activity limits the photosynthetic rates at moderately high temperatures.

Recently, however, it has been reported that the decrease in the photosynthetic rate at moderately high temperatures may be caused by the limitation of the ribulose 1,5-bisphosphate (RuBP) regeneration capacity rather than by the low Rubisco activation state per se in leaves of Pima cotton (Gossypium barbadense) (Schrader et al. 2004; Wise et al. 2004). Cen & Sage (2005) showed parallel decreases in the Rubisco activation state and in the electron transport rate with the decrease in the net photosynthetic rate at moderately high temperatures in sweet potato (Ipomoea batatas), and suggested that the decrease in the Rubisco activation state at moderately high temperatures is regulated by the electron transport capacity. Therefore, the decrease in the Rubisco activation state may be a regulatory response to the limitation of one of the processes contributing to RuBP regeneration, including the damage to thylakoid reactions, as Sharkey (2005) pointed out. Although the primary mechanisms underlying the Rubisco deactivation at moderately high temperatures are still unclear, apparently, the Rubisco activation state eventually decreases at moderately high temperatures.

Temperature dependence of the RuBP carboxylation rate has been considered to vary little across different species and growth conditions, because the kinetic parameters of Rubisco are thought to be conservative (Badger et al. 1982; Brooks & Farquhar 1985; von Caemmerer 2000). According to the model of C3 photosynthesis, the initial slope [IS(Ci)] of the A − Ci curve, expressing the dependence of the light-saturated photosynthetic rate (A) on the intercellular CO2 concentration (Ci), is an index of RuBP carboxylation activity (Farquhar, von Caemmerer & Berry 1980; von Caemmerer & Farquhar 1981). In recent studies, it has been shown that the optimum temperatures for IS(Ci) differ, depending on the growth temperatures (Hikosaka, Murakami & Hirose 1999; Bunce 2000; Yamori et al. 2005). For example, the optimum temperature for IS(Ci) in spinach leaves changed within only 2 weeks after the transfer to a lower growth temperature (Yamori et al. 2005). Although the temperature dependence of IS(Ci) is affected by the Rubisco activation state, these observations would indicate that the parameters of Rubisco kinetics differ depending on the growth temperature and contribute to the changes in the temperature dependence of the photosynthetic rate.

Huner & Macdowall (1979) showed that Rubisco purified from the low-temperature (LT) leaves (LT Rubisco) of Puma rye (Secale cereale L. cv. Puma) grown at 4/2 °C (day/night) had a higher apparent affinity for CO2 at lower temperatures than Rubisco purified from the high-temperature (HT) leaves (HT Rubisco) of plants grown at 25/20 °C. In contrast, HT Rubisco had a higher apparent affinity for CO2 at higher temperatures than LT Rubisco. However, in a later study, Huner (1986) found no significant effect of the growth temperature on the in vivo carboxylation efficiency in Puma rye leaves and concluded that the differences in the in vitro kinetics between HT and LT Rubisco were not important in the photosynthetic acclimation to the growth temperature. Because such studies have not been made for other species, it is still unclear whether Rubisco kinetics changes depending on the growth temperature. If the kinetics changes, the extent and importance in the photosynthetic acclimation should be quantified.

In this study, we examined whether temperature dependences of Rubisco activation state and in vitro Rubisco kinetics, two main factors potentially responsible for the changes in IS(Ci), changed depending on the growth temperature. We used spinach plants grown at 30/25 (HT) and at 15/10 °C (LT). In addition, we determined the effect of transfer HT plants to LT conditions. We measured net photosynthetic rate and Rubisco activation state in these leaves. With Rubisco purified from these leaves, we measured kinetic parameters such as the maximum rate of RuBP carboxylation (Vcmax) and the specificity factor (Sc/o), as well as the thermal stability. The effects of the differences in these parameters on the photosynthetic performance of spinach leaves were quantitatively evaluated.

MATERIALS AND METHODS

Plant growth conditions

Spinach (Spinacia oleracea L. cv. Torai) plants were grown in vermiculite, as described in Yamori et al. (2005). The day/night lengths were 8 and 16 h, respectively. Photosynthetically active photon flux density (PPFD) in the daytime was 230 µmol photons m−2 s−1. The day/night air temperatures were either 30/25 or 15/10 °C. These are referred to as the HT and LT conditions, respectively. The leaves grown at HT and LT are called HT and LT leaves, respectively. Likewise, Rubisco purified from HT and LT leaves are called HT and LT Rubisco, respectively. Some plants grown at HT were transferred to LT when the seventh leaves were just fully expanded; they were kept at LT for another two weeks. Their seventh leaves and their Rubisco are called HL leaves and HL Rubisco, respectively. The plants were watered once a week and fertilized with 200 mL of the Hoagland’s nutrient solution once a week.

Net photosynthetic rate and Rubisco activation state

Photosynthetic rates were measured as described previously (Yamori et al. 2005). Photosynthetic rates of intact leaves were measured at an ambient CO2 concentration (Ca) of 360 µL L−1 under the saturating light of 1500 µmol m−2 s−1 with a portable gas exchange system (LI-6400; Li-Cor, Lincoln, NE, USA). The leaf temperature was increased at 1 °C min−1 until the desired leaf temperature was attained (Law & Crafts-Brandner 1999; Crafts-Brandner & Salvucci 2002). The photosynthetic rates were measured at 45 min after the attainment of the temperature. Measurements were repeated three times using different plants for each temperature.

Immediately after the measurements of the photosynthetic rates, the leaf was rapidly frozen in liquid nitrogen and stored at −80 °C until used for the Rubisco activation state assay. The carboxylation rate of Rubisco was measured according to Lilley & Walker (1974) and Sawada et al. (1990) with slight modifications. Each frozen leaf (1.0 cm2) was rapidly homogenized in a chilled mortar and pestle with 0.892 mL of the extraction medium. The medium contained 100 mm 2-[4-(2-Hydroxyethyl)-1-piperazinyl]-ethanesulfonic acid (HEPES)–KOH (pH 7.8), 10 mm MgCl2, 5 mm dithiothreitol (DTT), and 1 mm ethylenediaminetetraacetic acid (EDTA). The homogenate was centrifuged at 16 000 g for 30 s at 4 °C, and the supernatant was used for the Rubisco assay. The initial activity of Rubisco was assayed at 25 °C for 1 min in the assay medium containing 100 mm bicine–KOH (pH 8.2), 20 mm MgCl2, 20 mm NaHCO3, 5 mm ATP, 5 mm creatine phosphate, 63 mm NADH, 10 units mL−1 of creatine kinase, 10 units mL−1 of 3-phosphoglyceric phosphokinase (PGK) and 25 units mL−1 of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (all enzymes were purchased from Sigma, St. Louis, MO, USA). Neither of the activity of PGK nor that of GAPDH limited Vcmax at any measurement temperatures (data not shown). These procedures, from the onset of the extraction to that of the initial activity assay, required about 1 min. Another portion of the extract was used for the determination of the total Rubisco activity. Rubisco was activated for 20 min at 4 °C in an activation medium that contained 375 mm HEPES–KOH (pH 7.8), 50 mm MgCl2, and 50 mm NaHCO3, before the measurement of the total activity with the same medium as was used for the determination of the initial activity. After establishing a steady baseline, the carboxylation of Rubisco was started by the addition of 0.6 mm RuBP. The reaction was measured as the decrease in absorbance at 340 nm as a result of the oxidation of NADH (Lilley & Walker 1974; Usuda 1985). The activation state was estimated as the ratio of the initial to total Rubisco activities. Protein extracted from leaves was quantified by the method of Bradford (1976) using a protein-assay kit (Bio-Rad protein assay; Bio-Rad, Tokyo, Japan).

Purification of Rubisco

Fresh spinach leaves (c. 0.5 g) were cut into pieces with a razor in an ice-cold blending buffer containing 50 mm HEPES–KOH (pH 7.8), 0.33 m sorbitol, 2 mm EDTA, 1 mm 2-mercaptoethanol, 1 mm MgCl2, 1 mm MnCl2 and 25 mm Na ascorbate, for approximately 5 min. The filtrate through an 11 mm filter (Millipore, MA, USA) was centrifuged at 15 000 g for 5 min at 4 °C. The precipitated chloroplasts were washed twice with the buffer, and subsequently suspended in a hypotonic buffer (10 mm HEPES–KOH, pH 7.5) on ice. The suspension was centrifuged at 15 000 g for 5 min. The supernatant, the stroma fraction, was used for purification of Rubisco.

Fractionation of Rubisco by the ion-exchange chromatography using a Mono-Q column (Amersham Pharmacia Biotech, Buckinghamshire, UK) was conducted according to Salvucci, Portis & Ogren (1986). The stroma fraction was eluted with the KCl linear gradient (0.15–1.00 m). The flow speed was 100 µL min−1, and the fraction volume was 100 µL tube−1. Fractions showing maximum absorbance at 280 nm were collected as purified Rubisco fractions, desalted and concentrated with Microcon YM-100 (Millipore). Rubisco prepared from lysed chloroplasts by the ion-exchange chromatography was highly pure as examined by sodium dodecyl sulphate–polyacrylamide gel electrophoreseis (SDS–PAGE) (data not shown).

Assay of the maximum rate of RuBP carboxylation rate of the purified Rubisco (Vcmax)

The purified Rubisco was activated for 20 min at 4 °C in an activation medium that contained 100 mm bicine–KOH (pH 8.2), 20 mm MgCl2, and 20 mm NaHCO3. The Vcmax was then assayed for 1 min in the assay medium (for ingredients, see Net Photosynthetic Rate and Rubisco Activation State).

Thermal stability of the purified Rubisco

The thermal stability of the purified Rubisco was examined according to Chen, Hong & Spreitzer (1993) with slight modifications. The purified Rubisco (5 µg) was incubated in 1.0 mL of the assay buffer (for ingredients, see Net Photosynthetic Rate and Rubisco Activation State), at a given treatment temperature ranging from 30 to 70 °C for 10 min. After incubation for 10 min, the sample was cooled on ice for 5 min. Then, the reaction was started at 25 °C by the addition of 0.6 mm RuBP.

Sc/o of Rubisco and Γ*

The Sc/o of the purified Rubisco was determined according to Suzuki, Mamedov & Ikawa (1999). The enzyme reaction was performed at temperatures ranging from 5 to 45 °C. Sc/o values were calculated as follows (Kane et al. 1994; Uemura et al. 1996):

image(1)

where α is the ratio between the solubilities of CO2 and O2 at each temperature in the aqueous phase (Hodgman, Weast & Selby 1958; Jordan & Ogren 1984) 3-phosphoglycerate ([PGA]) and phosphoglycolate ([PGO]) are the final concentrations of PGA and PGO after the reaction, and [O2] and [CO2] are the concentrations of O2 and CO2 in the gas phase, respectively.

The Γ* value, the CO2 compensation point in the absence of day respiration, was calculated from Sc/o by the following equation (Laing, Ogren & Hageman 1974):

image(2)

where Vcmax and Vomax are the maximum carboxylase and oxygenase activities, Kc and Ko are the Michaelis constants for CO2 and O2, respectively, and O is the concentration of O2 in the solution in equilibrium with 21% O2. By definition (Vcmax/Kc)/(Vomax/Ko) is the Sc/o.

Models

Model A: C3 photosynthesis model

Rubisco catalyzes both the carboxylation and oxygenation of RuBP (Farquhar et al. 1980). The net leaf photosynthetic rate measured as CO2 fixation (A, µmol m−2 s−1) can be expressed as

A = Vc − 0.5Vo − Rd(3)

where Vc and Vo (µmol m−2 s−1) are the rates of carboxylation and oxygenation, respectively, and Rd (µmol m−2 s−1) is the day respiration rate (Farquhar et al. 1980).

To highlight that A is a function of the chloroplastic CO2 concentration (Cc), we use Pc(Cc) instead of A. A under the RuBP carboxylation-limited conditions is expressed as

image(4)

When Rubisco activation state is taken into account, Pc(Cc) is expressed as

image(5)

where R* is the Rubisco activation state. Because the IS(Cc) of A versus Cc curve is identical to dPc(Cc)/dCc at Cc = Γ*, Vcmax is expressed as

image(6)

We measured temperature dependences of Vcmax and Sc/o, and calculated Γ* from Sc/o. Therefore, Kc(1 + O/Ko) values can be derived from Eqn 6, if IS(Cc) is known. Using the initial slope [IS(Ci)] of A versus Ci, obtained for spinach leaves grown under the conditions identical to the present experiment (Yamori et al. 2005), IS(Cc) was calculated as

image(7)

where gi is the internal conductance, the conductance for CO2 diffusion from the substomatal cavities to the carboxylation sites. Using gi, Ci and Cc, Pc(Cc) is expressed as

Pc(Cc) = gi × (Ci − Cc)(8)

In most of these calculations, we assumed that gi was constant over the temperature range and 0.2 mol m−2 s−1 both in HT and LT leaves, on the basis of the gi value reported for spinach leaves (Delfine et al. 1998, 1999). In some calculations, we took account of the temperature dependence of gi according to Bernacchi et al. (2002).

Because the photosynthetic rates at an ambient CO2 concentration (Ca) of 360 µL L−1 under the saturating light of 1500 µmol m−2 s−1 were limited mostly by RuBP carboxylation rate in HT and LT leaves of spinach for the temperature range from 9 to 39 °C (Yamori et al. 2005), we regarded Pc(Cc) as the in vivo RuBP carboxylation rate in this study.

Model B: effects of Rubisco kinetics and Rubisco activation state on the temperature dependence of the photosynthetic rate

We compared the measured photosynthetic rates and the predicted photosynthetic rates in HT and LT leaves, respectively (Fig. 6). The predicted photosynthetic rates in HT and LT leaves were calculated for two conditions: (A) differences in the Rubisco kinetic parameters were taken into account, assuming that Rubisco is fully activated, and (B) differences in both the Rubisco kinetic parameters and the Rubisco activation state were taken into account. In these analyses, we assumed that (1) Ci was constant at 300 µL L−1 in HT and LT leaves, because Ci was almost constant, 280–300 µL L−1 both in HT and LT leaves, over the temperature range employed in this study; (2) gi was 0.2 mol m−2 s−1 for the overall temperature range; and (3) Rd was half the respective dark respiration rates that were reported for HT and LT leaves (Yamori et al. 2005). We first calculated the temperature dependence of Kc(1 + O/Ko) values, using Eqn 6. Vcmax (µmol m−2 s−1) was calculated from the maximum carboxylation rate of the purified Rubisco (nmol CO2 mg−1 Rubisco s−1), and the Rubisco contents on the leaf area basis in HT and LT leaves of spinach (1.03 g m−2 and 2.00 g m−2, respectively) (Yamori et al. 2005). IS(Cc) were determined from gi and IS(Ci) obtained for spinach leaves (Yamori et al. 2005), using Eqn 7. Using Γ* and R* obtained in this study, we calculated the temperature dependence of Kc(1 + O/Ko).

Figure 6.

Effects of ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) kinetics and Rubisco activation state on the temperature dependence of the photosynthetic rate. We calculated Pc(Cc) as the in vivo ribulose 1,5-bisphosphate (RuBP) carboxylation rate at an ambient CO2 concentration (Ca) of 360 µL L−1 under the saturating light, as described in models A and B in Materials and Methods. The broken line shows the photosynthetic rates predicted for leaves that were assumed to have fully activated Rubisco in high-temperature (HT) and low-temperature (LT) leaves (condition A). The solid line shows the predicted photosynthetic rates taking account of the Rubisco activation states (condition B) in (a) HT (open triangle) and (b) LT leaves (filled triangle), respectively. These data were fitted by cubic curves. The values of the measured photosynthetic rates were equal to those presented in Fig. 1 (HT, open circle; LT, filled square).

Next, we calculated the absolute values of Vcmax (µmol  m−2 s−1) to match the Rubisco contents between the photosynthetic rates measured in this study and those estimated from the carboxylation rate of the purified Rubisco (nmol CO2 mg−1 Rubisco s−1). Using the photosynthetic rates measured at 15 and 30 °C in this study (Fig. 1), Cc was obtained using Eqn 8 for the respective temperatures. Vcmax on the leaf area basis was estimated from Eqn 5, using Cc thus obtained. Then, the Vcmax estimated for 15 and 30 °C were averaged for the HT and LT leaves, respectively. At either temperature, we measured the photosynthetic rates in three leaves. Therefore, the Vcmax on the leaf area basis was the mean values of six leaves for HT and LT, respectively. The Vcmax thus estimated at 15 °C were 14.8 µmol m−2 s−1 and 21.1 µmol m−2 s−1 for HT and LT leaves, respectively.

Figure 1.

Temperature dependences of net photosynthetic rate (a) and ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) activation state (b). The net photosynthetic rates were measured at an ambient CO2 concentration (Ca) of 360 µL L−1 under the saturating light of 1500 µmol m−2 s−1 (HT, open circle and broken line; LT, filled square and solid line). The data were fitted by cubic curves (HT: y = 0.0003x3 − 0.0647x2 + 2.7487x − 17.471, R2 = 0.97; LT: y = 0.0012x3 − 0.1194x2 + 3.3098x − 13.666, R2 = 0.99). The Rubisco activation state was determined as the ratio of initial activity to total activity (HT, open triangle and broken line; LT, filled triangle and solid line). Data represent means ± SE, n = 3. HT, leaves grown at high temperature; LT, leaves grown at low temperature.

For condition (A), we calculated the temperature dependence of Pc(Cc) in HT and LT leaves that have the temperature dependences of Rubisco kinetics (Vcmax, Γ* and Kc(1 + O/Ko)) and Rd in HT and LT leaves, using Eqns 4 and 8. For condition (B), we calculated the temperature dependence of Pc(Cc) in HT and LT leaves that have the temperature dependences of Rubisco kinetics (Vcmax, Γ* and Kc(1 + O/Ko)), Rd and Rubisco activation state in HT and LT leaves, using Eqns 5 and 8.

RESULTS

Temperature dependences of net photosynthetic rate and Rubisco activation state

The net photosynthetic rates at a Ca of 360 µL L−1 under the saturating light peaked at their respective growth temperatures both in HT and LT leaves (Fig. 1a) and declined markedly with the increase in measuring temperature, as has been reported (Yamori et al. 2005). The Rubisco activation states started to decrease at the same temperatures that the photosynthetic rates started to decrease (Fig. 1b). The extents of the decreases in the photosynthetic rates and those in the Rubisco activation states were similar at temperatures just above the respective growth temperatures, but the extents of the decreases in the photosynthetic rates were greater than those of the Rubisco activation states at much higher temperatures.

Although the net photosynthetic rates declined at temperatures below the growth temperatures both in HT and LT leaves, the Rubisco activation states remained high (Fig. 1).

Thermal stability of Rubisco

There was a clear difference in the thermal stabilities between HT and LT Rubisco (Fig. 2). HT Rubisco was stable even at 62.5 °C. However, LT Rubisco decreased its activity above 60 °C. The activity of LT Rubisco after the treatment at 65 °C was less than that of HT Rubisco by 50%. The thermal stability of HL Rubisco was similar to that of LT Rubisco.

Figure 2.

Thermal stability of ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco). Purified Rubisco (5 µg/mL) was treated at each temperature for 10 min. Then, the samples were cooled on ice, and the Rubisco activity was assayed at 25 °C. The activities were normalized against the level of activity measured after the 30 °C treatment. Data represent means ± SE, n = 3. Different letters show significant differences (Tukey–Kramer’s multiple comparison test, P < 0.05). HT Rubisco (open circle), Rubisco purified from high-temperature (HT) leaves; LT Rubisco (filled square), Rubisco purified from low-temperature (LT) leaves; HL Rubisco, Rubisco purified from leaves of plants, in which the seventh leaves were fully expanded at the transfer from 30/25 to 15/10 °C conditions. HL, leaves that were transferred from HT to LT conditions and were subsequently kept under LT conditions for 2 weeks.

Temperature dependence of the in vitro maximum rate of RuBP carboxylation of Rubisco (Vcmax)

The in vitro maximum rate of RuBP carboxylation of Rubisco (Vcmax) both in HT and LT Rubisco progressively increased with the increase in the measurement temperature up to 45 °C (Fig. 3a). Vcmax was higher in LT Rubisco than in HT Rubisco below 10 °C, while Vcmax was higher in HT Rubisco than in LT Rubisco above 40 °C. The temperature dependence of Vcmax in HL Rubisco was interesting. At low temperature, Vcmax in HL Rubisco was almost identical to that in LT Rubisco, while, at high temperature, the value was almost identical to that in HT Rubisco (Fig. 3a).

Figure 3.

(a) Temperature dependence of the in vitro maximum rate of ribulose 1,5-bisphosphate (RuBP) carboxylation (Vcmax). After activation for 20 min at 4 °C, the activities of purified Rubisco (20 µg/mL) were determined at the indicated temperatures. The data of Vcmax in HT and LT Rubisco were fitted by cubic curves (HT: broken line, y = 0.0007x3 − 0.0028x2 + 0.7832x − 2.5945, R2 = 1.00; LT: solid line, y = −0.0005x3 + 0.0726x2 − 0.5454x + 4.0938, R2 = 1.00). Data represent means ± SE, n = 3. Different letters show significant differences (Tukey–Kramer’s multiple comparison test, P < 0.05). (b) Arrhenius plots for the temperature response of Vcmax. Data represent the means in Fig. 3a. HT Rubisco (open circle), Rubisco purified from high-temperature (HT) leaves; LT Rubisco (filled square), Rubisco purified from low-temperature (LT) leaves; HL Rubisco, Rubisco purified from leaves of plants, in which the seventh leaves were fully expanded at the transfer from 30/25 to 15/10 °C conditions. HL, leaves that were transferred from HT to LT conditions and were subsequently kept under LT conditions for 2 weeks.

The Arrhenius plot of Vcmax values are shown in Fig. 3b. Above 15 °C, little difference in the temperature dependence of Vcmax was detected between HT and LT Rubisco. However, breaks in the Arrhenius plot were observed at 15 °C. The break in HT Rubisco was much sharper than that in LT Rubisco, indicating that the activation energy below 15 °C was greater in HT Rubisco than in LT Rubisco (data not shown). The temperature dependence of Vcmax in HL Rubisco was very similar to that of LT Rubisco below 15 °C (Fig. 3b).

Temperature dependence of the specificity factor (Sc/o) and Γ*

The specificity factor (Sc/o) of Rubisco measured at 25 °C was approximately 95 both in HT and LT Rubisco (Fig. 4), and consistent with the value of spinach Rubisco (Spinacia oleracea L. cv. Jiro-Maru) reported by Uemura et al. (1996). In addition, temperature dependences of Sc/o generally agreed with those reported by Jordan & Ogren (1981). The temperature dependences of Sc/o were clearly different, depending on the growth temperature. Sc/o was markedly higher in LT Rubisco than in HT Rubisco below 10 °C, while it was higher in HT Rubisco than in LT Rubisco above 35 °C.

Figure 4.

Temperature dependence of the specificity factor (Sc/o). The values of the Sc/o were determined with 100 µg/mL purified ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) and gas phases of 0.05% CO2, 99.95% O2. The data were fitted by cubic curves (HT: broken line, y = − 0.0019x3 + 0.1597x2− 5.5915x + 164.54, R2 = 1.00; LT: solid line, y = − 0.0022x3 + 0.175x2− 6.4275x + 178.59, R2 = 0.99). Data represent means ± SE, n = 3–6. * represents a significant difference (< 0.05), according to Student’s t-test. HT Rubisco (open circle), Rubisco purified from high-temperature (HT) leaves; LT Rubisco (filled square), Rubisco purified from low-temperature (LT) leaves.

Γ* was calculated from Eqn 2, using Sc/o. The Γ* values obtained in this study were compared with those in other studies using spinach (Jordan & Ogren 1984; Brooks & Farquhar 1985) and tobacco (Bernacchi et al. 2001) (Fig. 5). Γ* in Brooks & Farquhar (1985) and Bernacchi et al. (2001) were determined by the gas exchange technique according to the procedure of Laisk (1977) in which Γ* values were determined as the CO2 concentration where the photosynthetic CO2 uptake equals the photorespiratory CO2 efflux at several PPFDs. However, Jordan & Ogren (1984) determined Γ* from the measurements of kinetic constants of purified Rubisco. The temperature dependences of Γ* obtained in this study showed similar tendencies in the previous studies. The temperature dependences of Γ* in this study, however, were clearly different, depending on the growth temperature. Γ* in HT Rubisco was markedly lower than that in LT Rubisco at temperatures above 40 °C.

Figure 5.

Temperature dependences of the CO2 compensation point in the absence of day respiration (Γ*) estimated from the this study and obtained from the previous studies. The Γ* values in this study were calculated from averages of the specificity factor (Sc/o) values in Fig. 4, using Eqn 2. These data were fitted by cubic curves (HT: broken line, y = 0.0012x3 − 0.0613x2 + 1.8469x + 14.348, R2 = 1.00; LT: solid line, y = 0.0021x3− 0.1083x2 + 2.5821x + 9.8365, R2 = 1.00). The values of the other studies are for spinach (open triangle, Jordan & Ogren 1984; shaded and filled triangle, Brooks & Farquhar 1985) and tobacco (filled diamond, Bernacchi et al. 2001). HT ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) (open circle), Rubisco purified from high-temperature (HT) leaves; LT Rubisco (filled square), Rubisco purified from low-temperature (LT) leaves.

Effects of Rubisco kinetics and Rubisco activation state on the temperature dependence of the photosynthetic rate

We examined the effects of the differences in in vitro Rubisco kinetics and Rubisco activation state on the temperature dependences of the in vivo RuBP carboxylation rates [Pc(Cc)] at a Ca of 360 µL L−1 under the saturating light (Fig. 6). The measured photosynthetic rates were the same values as those in Fig. 1a. The predicted photosynthetic rates were calculated (condition A) assuming that Rubisco is fully activated, (condition B) taking account of the Rubisco activation state, as described in models A and B in Materials and Methods (Fig. 6). At temperatures below the growth temperatures, the Rubisco activation states were assumed to be the same values that were measured at the lowest temperatures in HT and LT leaves, respectively (HT, 15 °C; LT, 10 °C). However, at temperatures above the growth temperatures, the temperature dependences of the Rubisco activation state for HT and LT leaves were fitted by quadratic curves, respectively. The predicted photosynthetic rates under condition (A) closely matched the measured photosynthetic rates at low temperatures both in HT and LT leaves (Fig. 6). The extents of the decreases in the photosynthetic rates at low temperatures in the HT leaf were greater than those in the LT leaf.

However, at moderately high temperatures, the predicted photosynthetic rates under condition (A) were much higher than the measured photosynthetic rates both in HT and LT leaves, although the extents of the decreases in the predicted photosynthetic rates in the LT leaf were greater than those in the HT leaf. Because the Rubisco activation state decreased with increasing temperatures above the growth temperatures (Fig. 1b), the photosynthetic rates predicted for condition (B) decreased more steeply than those predicted for condition (A) at moderately high temperatures, and agreed well with the measured photosynthetic rates both in HT and LT leaves (Fig. 6).

The optimum temperature of the predicted photosynthetic rates under condition (B) in the HT leaf was higher than that in the LT leaf by approximately 5 °C, when the temperature dependences of the predicted photosynthetic rate were fitted by the cubic curves (Fig. 6). The temperature dependences of the predicted photosynthetic rates under condition (B) closely matched those of the measured photosynthetic rates over the entire temperature range.

We analyzed relationships between the predicted photosynthetic rates under condition (B) and the measured photosynthetic rates at temperatures above the respective optimum temperatures in HT and LT leaves (30 and 25 °C for HT and LT leaves). We pooled the relevant data and analyzed their relationships. The predicted photosynthetic rates (y) were strongly dependent on the measured photosynthetic rates (x) (y = 1.12x, R2 = 0.89). Next, we calculated the photosynthetic rates for condition (B) assuming that gi was 0.2 mol m−2 s−1 at 25 °C and showed that the temperature dependence had a sharp peak at 35.0–37.5 °C as reported in Bernacchi et al. (2002). The results showed a trend similar to Fig. 6 (y = 1.10x, R2 = 0.82).

DISCUSSION

Effects of Rubisco kinetics and Rubisco activation state on the temperature dependence of the photosynthetic rate

When the photosynthetic rates at a Ca of 360 µL L−1 under the saturating light were predicted for leaves that were assumed to have fully activated Rubisco (condition A), the resultant temperature responses matched the measured photosynthetic rates at temperatures below the optimum temperatures for both HT and LT leaves (Fig. 6). At low temperatures, the differences in the temperature dependence of the photosynthetic rate were not due to the differences in the Rubisco activation state (Fig. 1b), but mainly due to the differences in the Rubisco kinetics (Figs 3a, 4 & 5). Plants grown at low temperatures have higher amounts of enzymes of the photosynthetic carbon reduction cycle, including Rubisco (Holaday et al. 1992; Hurry et al. 1995; Strand et al. 1999; Yamori et al. 2005). Generally, large amounts of the enzymes would be needed to compensate for decreased activities of these enzymes at low temperatures. In this study, we suggest that not only the increases in the amount of photosynthetic enzymes, but also the changes in Rubisco kinetics contribute to the increases in the photosynthetic capacity at low temperatures.

The photosynthetic rates are markedly affected by Vcmax and Kc(1 + O/Ko) (Eqn 5). Generally, the temperature dependence of Kc(1 + O/Ko) mostly depends on Kc, because the temperature dependence of Ko is small (Jordan & Ogren 1984). In the present study, Kc(1 + O/Ko) were calculated from Eqn 6 in HT and LT leaves. The values of Kc(1 + O/Ko) in LT Rubisco were greater than those in HT Rubisco over the temperature range that were examined (data not shown). The Vcmax in LT Rubisco were greater at low temperatures than those in HT Rubisco, while the Vcmax in HT Rubisco were greater at high temperatures than those in LT Rubisco (Fig. 3a). The photosynthetic rates predicted for the LT leaf that were assumed to have fully activated Rubisco were greater than those of the HT leaf at low temperatures (Fig. 6). These results indicate that the effects of Vcmax on the photosynthetic rates at low temperatures were greater than those of Kc. Generally, there is a trade-off relationship between Vcmax and Kc (von Caemmerer & Quick 2000). The greater Vcmax was associated with a decreased affinity for CO2 (increased Kc), when comparison was made between wheat and rice (Makino, Mae & Ohira 1985, 1988) and between C3 and C4 plants (Yeoh, Badger & Watson 1980, 1981; Seemann, Badger & Berry 1984; Sage 2002). If the CO2 concentration in aqueous phase is sufficiently high, and Kc is sufficiently low at low temperature, it is thought that the changes in Vcmax are more important than those in Kc for the increase in the photosynthetic rate, at least, at low temperatures.

At moderately high temperatures, the predicted photosynthetic rates under condition (A) deviated from the measured photosynthetic rates in both HT and LT leaves (Fig. 6), although the extent of the decreases in the predicted photosynthetic rates in the LT leaf were greater than those in the HT leaf. However, the predicted photosynthetic rates taking account of the Rubisco activation state (condition B) closely matched the measured photosynthetic rates at moderately high temperatures, and properly reproduced the shift in the optimum temperature of the photosynthetic rates (Fig. 6). This suggests that photosynthetic performance at moderately high temperatures was mainly determined by the Rubisco activation state and Rubisco kinetics, irrespective of the growth temperatures. Recently, it is thought that the decrease in the Rubisco activation state at moderately high temperatures can be attributed to (1) the suppressed Rubisco activase activity (Crafts-Brandner & Salvucci 2000, 2002; Salvucci & Crafts-Brandner 2004a,b; Haldimann & Feller 2004, 2005; Kim & Portis 2005) or (2) the regulatory response to the limitation of thylakoid reactions (Schrader et al. 2004; Wise et al. 2004; Cen & Sage 2005; Sharkey 2005). Possibly, these mechanisms would differ among plant species and/or growth conditions, because the plants used in these studies were all different (see Introduction). In our previous study, comparison of temperature dependencies between IS(Ci) and the maximum electron transport rate (Jmax) led us to conclude that the photosynthetic rate at a Ca of 360 µL L−1 under saturating light is limited by the RuBP carboxylation in spinach leaves (Yamori et al. 2005). Moreover, we showed that, at moderately high temperatures where Rubisco activase liability would substantially reduce the photosynthetic rates (Salvucci & Crafts-Brandner 2004a), IS(Ci) exhibited considerable reductions in spinach leaves (Yamori et al. 2005). Thus, the decreases in the photosynthetic rate at moderately high temperatures in spinach leaves were probably attributed to the decreases in the Rubisco activation state. Clearly, Rubisco activase plays an essential role in the process of regulating the Rubisco activation state, and the activation state eventually becomes limiting or co-limiting with thylakoid reactions for the photosynthetic rates at moderately high temperatures. Further studies are necessary to clarify the primary mechanisms of the Rubisco deactivation at moderately high temperatures.

There are small inconsistencies between the predicted photosynthetic rates under condition (B) and the measured photosynthetic rates both in HT and LT leaves. In this study, Rubisco contents were estimated from the measured photosynthetic rate (Fig. 1a), using Eqns 5 and 8, as described in model B in Materials and Methods. Some differences between the predicted photosynthetic rates and the measured photosynthetic rates were possible, because the leaves used for the measurements were different for each of the measurement temperatures. The small inconsistencies could also be caused by the assumption of the constant gi. Although gi is thought to be an important limiting factor of photosynthesis (for a review, see Evans & Loreto 2000; von Caemmerer 2000), there is only one report for its temperature dependence. Bernacchi et al. (2002) reported that gi had the optimum temperature at 35.0–37.5 °C in tobacco leaves grown at 25/18 °C (day/night). There are some indications concerning the effects of the growth temperature on gi. Makino, Nakano & Mae (1994) suggested that Cc was lower in rice leaves grown at lower temperatures from the analyses of the gas exchange and Rubisco activity. However, with simultaneous measurements of gas exchange and Chl fluorescence, Hikosaka & Hirose (2001) argued that the Cc/Ci ratio was not different, depending on the growth temperature in leaves of Nerium oleander. Clearly, more intensive studies on the temperature dependence of gi are needed.

These results strongly suggest that the changes in the temperature dependence of the RuBP carboxylation rate depending on the growth temperature revealed by gas exchange studies (Hikosaka et al. 1999; Bunce 2000; Yamori et al. 2005) are at least partly due to the changes in the Rubisco kinetics (Vcmax, Sc/o and Γ*) and the Rubisco activation state.

Temperature effects in Rubisco activation state

We clearly showed that the decreasing patterns of the Rubisco activation state at moderately high temperatures were different between HT and LT leaves (Fig. 1b). Although Haldimann & Feller (2005) showed that responses of the Rubisco activation state to moderately high temperatures were similar in pea plants grown at 25 and 35 °C, those of the photosynthetic rate were also similar. Therefore, it is difficult to discuss the temperature acclimation of Rubisco activase on the basis of their study. Salvucci & Crafts-Brandner (2004b) reported that the optimum temperature and thermal stability of Rubisco activase in creosote bush native to warm regions were higher than those of Antarctic hair grass native to cold regions and that this correlated with the superior photosynthetic performance in creosote bush at moderately high temperatures. This suggests that Rubisco activase could acclimate to the growth temperature, and the optimum temperature for Rubisco activase would differ depending on the growth temperatures. Rubisco activase in spinach consists of two isoforms generated by the alternative splicing of a single pre-mRNA (Werneke, Chatfield & Ogren 1989). These two isoforms of activase in spinach markedly differed in thermal stability both in vitro and in vivo (Crafts-Brandner, van de Loo & Salvucci 1997; Rokka, Zhang & Aro 2001). It was also reported that the kinetics of ATP binding differed between the two forms of Rubisco activase in spinach (Shen, Orozco & Ogren 1991; Crafts-Brandner et al. 1997). Therefore, the differential expression of two forms of Rubisco activase could provide a mechanism of acclimation of Rubisco activase to the growth conditions.

Temperature acclimation of Rubisco kinetic properties

The thermal stability of Rubisco (Fig. 2) and the temperature dependences of Rubisco kinetic parameters such as Vcmax, Sc/o and Γ* (Figs 3a, 4 & 5) were different, depending on the growth temperature. LT Rubisco performed more efficiently at low temperature than HT Rubisco, while HT Rubisco performed more efficiently at high temperature than LT Rubisco. The temperature dependence of Vcmax in HL Rubisco was intermediate between those of HT and LT Rubisco (Fig. 3a), while the thermal stability of HL Rubisco was almost similar to that of LT Rubisco (Fig. 4). By using two-dimensional gel electrophoresis, we were able to detect the electrophoretic differences between HT and LT Rubisco. Moreover, we observed gradual changes in the electrophoretic pattern with the duration at LT during leaf development (unpublished results). Thus, it is fair to say that Rubisco changes depending on the growth temperature and/or in response to changes in the growth temperature. Moreover, it is highly likely that Rubisco kinetic parameters change with the structural changes of Rubisco, and that such relationship underlies the acclimation of Rubisco properties to the growth temperature. In this study, HL Rubisco purified from leaves that experienced the low growth temperature for 2 weeks showed marked acclimation (Figs 2 & 3). Simpson (1978) and Simpson, Cooke & Davies (1981), using [3H]acetic anhydride and tritiated water, have shown that Rubisco from maize leaves is degraded under conditions in which the level of the enzyme remains constant. Moreover, Simpson et al. (1981) reported that the half lives of Rubisco in 13-day-old second leaves of maize under continuous light and 14 h light/10 h darkness were approximately 7 and 6 d, respectively. These would support our results of the temperature acclimation of Rubisco within only 2 weeks.

Huner (1985) showed that HT Rubisco from Puma rye leaves formed hexagonal crystals, whereas LT Rubisco formed cubic crystals. However, we do not know detailed differences of crystal structures between HT and LT Rubisco. Rubisco is a hexadecameric enzyme consisting of eight small subunits (SSU) of 14 kDa and eight large subunits (LSU) of 53 kDa (Knight, Andersson & Branden 1990). LSU, which carries the catalytic site (Lorimer 1981), is encoded in a single chloroplast gene (Coen et al. 1977). In contrast, SSU is encoded in the nuclear genome in the form of a multigene family consisting of 2 (Chlamydomonas) to 12 (wheat) members (Dean, Pichersky & Dunsmuir 1989). Moreover, a number of co- and post-translational modifications are known to occur in both LSU and SSU during the expression, import and assembly of the protein (Houtz et al. 1992; Houtz & Portis 2003). In spinach leaves expanded at the different temperatures, expression of isozyme(s) of SSU and/or some post-translational modifications of the enzyme may be different, which were probably reflected from the above-mentioned difference in their electrophoretic properties.

CONCLUSION

This study showed that Rubisco kinetic properties acclimated to the growth temperature. At low temperatures, the changes in Rubisco kinetics depending on the growth temperature contributed to the increases in the photosynthetic rates. At moderately high temperatures, the photosynthetic rates were mainly determined by Rubisco kinetics and the Rubisco activation state, and the activation state also differed depending on the growth temperature. These changes, depending on the growth temperature, were responsible, at least partly, for the differences in the temperature dependence of the photosynthetic rates. Because the photosynthetic rates are mostly limited by the RuBP carboxylation rates under a normal CO2 condition (360 µL L−1), it is important to change the temperature dependences of the Rubisco kinetics and the Rubisco activation state for efficient photosynthesis at the growth temperature.

ACKNOWLEDGMENTS

We are grateful to the late Professor S. Sawada (Hirosaki University, Japan) for technical advice on measurements of Rubisco activity. We are also grateful to Dr K. Hikosaka (Tohoku University, Japan), Dr M.E. Salvucci (Western Cotton Research Laboratory, Phoenix, AZ, USA), Professor A. Yokota, Dr K. Akashi (Nara Institute of Science and Technology, Japan) and Professor T. Hase (Osaka University, Japan) for valuable advice, and to Dr T. Yabe (Osaka University, Japan) for instruction on the Rubisco purification technique. This work was supported by a grant from Japan Society for the Promotion of Science (JSPS) for young research fellows to W.Y.

Ancillary