Suppression of chloroplast triose phosphate isomerase evokes inorganic phosphate-limited photosynthesis in rice

Abstract

The availability of inorganic phosphate (P_i) for ATP synthesis is thought to limit photosynthesis at elevated [CO₂] when P_i regeneration via sucrose or starch synthesis is limited. We report here another mechanism for the occurrence of P_i-limited photosynthesis caused by insufficient capacity of chloroplast triose phosphate isomerase (cpTPI). In cpTPI-antisense transgenic rice (Oryza sativa) plants with 55%–86% reductions in cpTPI content, CO₂ sensitivity of the rate of CO₂ assimilation (A) decreased and even reversed at elevated [CO₂]. The pool sizes of the Calvin–Benson cycle metabolites from pentose phosphates to 3-phosphoglycerate increased at elevated [CO₂], whereas those of ATP decreased. These phenomena are similar to the typical symptoms of P_i-limited photosynthesis, suggesting sufficient capacity of cpTPI is necessary to prevent the occurrence of P_i-limited photosynthesis and that cpTPI content moderately affects photosynthetic capacity at elevated [CO₂]. As there tended to be slight variations in the amounts of total leaf-N depending on the genotypes, relationships between A and the amounts of cpTPI were examined after these parameters were expressed per unit amount of total leaf-N (A/N and cpTPI/N, respectively). A/N at elevated [CO₂] decreased linearly as cpTPI/N decreased before A/N sharply decreased, owing to further decreases in cpTPI/N. Within this linear range, decreases in cpTPI/N by 80% led to decreases up to 27% in A/N at elevated [CO₂]. Thus, cpTPI function is crucial for photosynthesis at elevated [CO₂].

Suppression of chloroplast triose phosphate isomerase suppresses photosynthesis at elevated [CO2] with the typical symptoms of photosynthesis limited by Pi availability for ATP synthesis.

Introduction

As improving photosynthesis is thought to be an effective means for improving plant productivity (Zhu et al., 2010; Evans, 2013; Simkin et al., 2019; Yoon et al., 2020), the molecular mechanisms that limit photosynthetic capacity need to be revealed and have been studied extensively. CO₂ assimilation is carried out by the Calvin–Benson cycle, which consists of 11 enzymes and produces carbohydrates from atmospheric CO₂ using NADPH and ATP supplied by photochemical reactions (Calvin, 1962; Heldt and Piechulla, 2011). The rate of CO₂ assimilation (A) is also closely related to the production of sucrose or starch, which are end products of photosynthesis, or the transient reservoir of photosynthates (Sharkey, 1985a; Stitt et al., 2010; McClain and Sharkey, 2019). According to the widely accepted C₃ photosynthetic model of Farquhar et al. (1980) and von Caemmerer (2000), the maximal A at low [CO₂] is thought to be determined by the enzymatic capacity of Rubisco, whereas that at elevated [CO₂] is thought to be determined by the capacity for regeneration of the Rubisco substrate, ribulose 1,5-bisphosphate (RuBP). In addition, Sharkey (1985a) proposed that the availability of inorganic phosphate (P_i) for ATP synthesis, which depends on sucrose or starch synthesis capacity, also limits photosynthesis. P_i-limited photosynthesis is reported to occur at extremely elevated [CO₂], combination of elevated [CO₂] and low [O₂], low temperatures, or strong sink limitation under high irradiance (Sharkey, 1985a, 1985b; Sharkey et al., 1986a, 1986b; Sage et al., 1988; Sage and Kubien, 2007; Busch and Sage, 2017; Fabre et al., 2019; McClain and Sharkey, 2019). Global gross primary productivity under present-day and future conditions is estimated to decrease when Pi-limited photosynthesis is included in the simulation (Lombardozzi et al., 2018), suggesting the importance of this photosynthetic limitation.

The mechanisms underlying the occurrence of P_i-limited photosynthesis are proposed as follows. When sucrose or starch synthesis cannot keep up with an increase in A, the amounts of phosphorylated intermediates leading to sucrose or starch synthesis and those of general phosphorylated compounds including the Calvin–Benson cycle metabolites increase, leading to the depletion of P_i and the limitation of ATP synthesis and photosynthesis. As A is limited by the availability of P_i, the sensitivity of A to [CO₂] and [O₂] owing to the enzymatic properties of Rubisco is lost (Sharkey, 1985a, 1985b; Sharkey et al., 1986a). Such symptoms in A and even reverse sensitivity of A to [CO₂] have been observed (Sharkey, 1985a, 1985b; Sharkey et al., 1986a, 1986b; Sage et al., 1988; Sage and Kubien, 2007; Busch and Sage, 2017; Fabre et al., 2019; McClain and Sharkey, 2019). The metabolic changes mentioned above were observed experimentally (Sharkey et al., 1986a, 1986b; Sage et al., 1988). Simultaneously, decreases in the activation state of Rubisco were observed (Sharkey et al. 1986a, 1986b; Sage et al. 1988). As expected from this theory, suppression of cytosolic fructose-1,6-bisphosphatase activity in mutants or transgenic plants led to the occurrence of P_i-limited photosynthesis and concomitant metabolic changes (Sharkey et al., 1988; Zrenner et al., 1996; Strand et al., 2000). In addition, in Arabidopsis (Arabidopsis thaliana) knockout mutants of the triose phosphate/phosphate translocator lost the CO₂ sensitivity of A at elevated [CO₂], suggesting the occurrence of P_i-limited photosynthesis (Schneider et al., 2002; Walters et al., 2004). It is noted that P_i-limited photosynthesis is less related to the level of P nutrition. As the buffering action of vacuoles for P_i is strong, changes in [P_i] in plastids are relatively small when plants are grown under low or high P nutrition levels (McClain and Sharkey, 2019). Although the occurrence of P_i-limited photosynthesis is closely related to the metabolic changes in the Calvin–Benson cycle, the symptoms of P_i-limited photosynthesis have rarely been reported when the capacities of the Calvin–Benson cycle enzymes were genetically modified. For instance, suppression of the Calvin–Benson enzymes after Rubisco, such as glyceraldehyde-3-phosphate dehydrogenase (GAPDH), aldolase, transketolase, and sedoheptulose-1,7-bisphosphatase (SBPase), led to RuBP-regeneration limited photosynthesis (Price et al., 1995; Haake et al., 1998; Ruuska et al., 2000; Harrison et al., 2001; Henkes et al., 2001).

There are some Calvin–Benson cycle enzymes whose control over A have not been examined. One example is chloroplast triose phosphate isomerase (cpTPI), which catalyzes the interconversion of glyceraldehyde 3-phosphate (GAP) and dihydroxyacetone phosphate (DHAP), whereas the cytosolic isoform (cytTPI) is responsible for the same reaction in glycolysis (Bukowiecki and Anderson, 1974; Kurzok and Feierabend, 1984; Pichersky and Gottlieb, 1984). Generally, TPI is thought to be a fast and efficient enzyme (Knowles and Albery, 1977; Blacklow et al., 1988). The capacity of cpTPI was estimated to be very high and thought to be in large excess for photosynthesis (Fridlyand, 1992; Harris and Königer, 1997; Fridlyand et al., 1999). In contrast, there are some indications that in vivo cpTPI activity affects the metabolism related to the Calvin–Benson cycle. cpTPI was found to be inhibited by 2-phosphoglycolate (2-PG), the product of Rubisco oxygenation reaction, in Arabidopsis mutants and transformants in which 2-PG over-accumulated owing to dysfunctions in the photorespiratory metabolism (Flügel et al., 2017; Li et al., 2019). Concomitantly, the pool sizes of most of the Calvin–Benson cycle metabolites decreased (Flügel et al., 2017). The inhibitory effects of 2-PG were suggested to lead to the disequilibrium of cpTPI reaction in wild-type (WT) Arabidopsis plants, as the GAP:DHAP ratio measured was higher than expected from the equilibrium state (Li et al., 2019). The higher concentration of GAP is thought to be favorable for photosynthesis, as much more energy is provided to drive the aldolase reaction and help synthesis of fructose 1,6-bisphosphate (FBP) (Sharkey, 2019). The contribution of cpTPI to primary metabolism and plant development has only been examined in a cptpi-knockdown Arabidopsis mutant (Chen and Thelen, 2010). This mutant showed an amount of cpTPI decreased by 75% and over-accumulation of some by-products of triacylglycerol catabolism, including DHAP and methylglyoxyal, during the early stage of plant development. These phenomena led to the inhibition of the transition from heterotrophic growth to photoautotrophic growth, and hence, severely stunted growth. Probably, because of such growth characteristics, the photosynthetic characteristics of this mutant are unknown.

In this study, we aim to investigate whether the capacity of cpTPI affects photosynthetic capacity. For this purpose, transgenic plants with antisense suppression of cpTPI were generated in rice (Oryza sativa). As triacylglycerol is not the major form of carbon storage in rice seeds, the negative effects of its catabolism are expected to be avoided. cpTPI-antisense rice plants grew photoautotrophically without severe abnormalities but showed the typical symptoms of P_i-limited photosynthesis. To obtain further evidence for the occurrence of P_i-limited photosynthesis, metabolic changes related to the Calvin–Benson cycle were examined. In addition, the control of cpTPI over A was evaluated from relationships between A measured at different [CO₂] and the amounts of cpTPI. These parameters were expressed per unit amount of total leaf-N, as there tended to be some variations in the amounts of total leaf-N depending on the genotypes. It is shown that insufficient capacity of cpTPI evokes P_i-limited photosynthesis and reverse CO₂ sensitivity of A at elevated [CO₂]. These results suggest that the functioning of cpTPI is crucial for photosynthesis at elevated [CO₂].

Results

Generation of cpTPI-antisense transgenic rice plants

In rice, one gene (Os09g0535000) is annotated as cpTPI (Zhang et al., 2010; Izawa et al., 2011; Hamamoto et al., 2012). To suppress gene expression of cpTPI in leaves, antisense RNA was expressed under the control of the promoter of rice Rubisco small subunit 2 (RBCS2) (Kyozuka et al., 1993). Its promoter activity is expected to be weak in tissues other than leaves, according to the transcript levels of RBCS2 in various rice tissues (Suzuki et al., 2009a; Sato et al., 2013). Three transgenic plant lines (TPIAS4, TPIAS7, and TPIAS14) were selected for analysis. Plants were grown under a strong light intensity (photosynthetic photon flux density [PPFD] of 1,000-μmol quanta m⁻² s⁻¹ at a maximum) and a normal [CO₂] of 40 Pa throughout the experiments. The amounts of cpTPI, Rubisco, and cytochrome f (Cyt f) subunit of the cytochrome b₆/f complex, chlorophyll, and total leaf-N were determined in mature leaves of T₂ progenies of these transgenic plants (Figure 1). WT plants and null segregants obtained from transgenic plants were used as control plants. The cpTPI protein was detected by immunoblotting using a peptide antibody designed to avoid cross-reaction to cytTPI. Single bands were detected at ∼26 kDa (Supplemental Figure S1), which is similar to the value for the cpTPI monomer in spinach (Spinacia oleracea) and rye (Secale cereale) (27 kDa; Pichersky and Gottlieb, 1984; Kurzok and Feierabend, 1984). When the transit peptide predicted using ChloroP (Emanuelsson et al., 1999) was excluded from the full-length amino acid sequence for rice cpTPI, the molecular weight was calculated to be 27.4 kDa. The amount of cpTPI protein decreased to 14%, 22%, and 45% of those in WT plants in lines TPIAS4, TPIAS7, and TPIAS14, respectively (Figure 1A). The specificity of the antibody for cpTPI was tested by activity staining of TPI after crude extracts were separated by native-polyacrylamide gel electrophoresis (PAGE), followed by immunoblotting (Supplemental Figure S2A). Two bands were detected using activity staining. Only a band with high mobility was detected by immunoblotting. These results are in accordance with the previous observations by Dorion et al. (2005) and Chen and Thelen (2010) that the bands of TPI activity with high and low mobility correspond to cpTPI and cytTPI, respectively. Thus, the peptide antibody for cpTPI was unlikely to cross-react with cytTPI. When all genotypes were subjected to activity staining, trends in differences in the band intensities of cpTPI resembled those in the amounts of cpTPI (Supplemental Figure S2B; Figure 1A), whereas the band intensities of cytTPI did not differ markedly among the genotypes (Supplemental Figure S2B). These results also suggest that the amounts of cpTPI specifically decreased.

Figure 1.

Photosynthetic components in transgenic plants. Amounts of (A) cpTPI; (B) Rubisco; (C) Cyt f; (D) chlorophyll; and (E) total leaf-N in the uppermost and fully expanded leaves in WT plants, null plants, and transgenic plants with antisense suppression of cpTPI (TPIAS4, TPIAS7, and TPIAS14) in rice. Data are presented as the means ± standard deviation (SD) (n = 4). ^† and ^*** denotes a significant difference in comparison with the values in WT plants by the Dunnett’s test at P < 0.1 and P < 0.001, respectively.

In these lines, the amounts of Rubisco, Cyt f, and total leaf-N tended to be slightly greater than in WT plants in most cases (Figure 1, B, C, and E). In contrast, the amounts of chlorophyll in these lines were almost identical to that in WT plants (Figure 1D). The characteristics of null plants were similar to those of WT plants (Figure 1, A–E). These results indicate that transgenic rice plants with decreased amounts of cpTPI were successfully generated. The results of immunoblotting for the determination of Cyt f are presented in Supplemental Figure S3.

In contrast to the cptpi-knockdown Arabidopsis mutant (Chen and Thelen, 2010), severe abnormalities in transgenic rice plant growth were not observed. The representative appearance of cpTPI-antisense plants on the 61st day after sowing is presented in Supplemental Figure S4A. Biomass production in lines TPIAS4 and TPIAS7 was ∼70% of that in WT plants, whereas those in line TPIAS14 and null plants did not differ considerably from that in WT plants (Supplemental Figure S4B). The main reason for growth abnormality in the Arabidopsis mutant was the metabolic disorder related to triacylglycerol catabolism during the early stage of plant development (Chen and Thelen, 2010). Such effects were likely to be avoided in rice, in which the major form of seed carbon storage is starch. The use of the RBCS2 promoter for antisense suppression could be another reason as its activity is expected to be high only in leaf blades.

Transcript levels of the genes for cpTPI, cytTPI, Rubisco, and Cyt f were determined along with transcript levels of the transgene, that is, the antisense construct for cpTPI (Figure 2). Surprisingly, transcript levels of cpTPI were substantially higher in cpTPI-antisense plants than in other genotypes (Figure 2A). This result was confirmed when almost the entire region of the full-length cDNA for cpTPI was amplified by reverse transcription-polymerase chain reaction (RT-PCR)  (Supplemental Figure S5). Similar trends were observed when the expression levels of transgene were examined by amplifying the conjunction region of the RBCS2 promoter and cpTPI antisense cDNA (Figure 2B). In both cases, these transcript levels were the highest in line TPIAS4, in which the amount of cpTPI protein was most severely suppressed (Figure 1A). Although gene expression of cpTPI might have been suppressed by co-suppression, the reason for the substantial increase in transcript levels of cpTPI is unknown. In rice, two genes (Os01g0147900 and Os01g0841600) are annotated as cytTPI (Izawa et al., 2011; Hamamoto et al., 2012). As transcript levels in various tissues, including leaves, were much higher in Os01g0147900 than in Os01g0841600 (Izawa et al., 2011; Sato et al., 2013), Os01g0147900 was selected as the dominant cytTPI gene. Transcript levels of cytTPI were similar among the genotypes (Figure 2C). There were also no significant differences in the transcript levels of the genes for Rubisco large subunit (RBCL), RBCS, and Cyt f (PETA) (Figure 2, D–F). These results indicate that antisense suppression of cpTPI did not affect the gene expression of cytTPI, Rubisco, and Cyt f. The characteristics of null plants were similar to those of WT plants (Figure 2, A–F). Levels of internal standard 18 S rRNA were almost identical among the genotypes (Figure 2G), indicating the validity of the analyses.

Figure 2.

Transcript levels of photosynthesis-related genes and transgene in transgenic plants. Transcript levels of (A) cpTPI; (B) transgene for antisense suppression of cpTPI; (C) cytTPI; (D) RBCL; (E) RBCS; and (F) PETA in young and expanding leaves in WT plants, null plants, and transgenic plants with antisense suppression of cpTPI (TPIAS4, TPIAS7, and TPIAS14) in rice. (G) Levels of 18 S rRNA were also presented as controls. Data are presented as the means ± sd (n = 3). ^** and ^*** denote a significant difference in comparison with the values in WT plants by the Dunnett’s test at P < 0.01 and P < 0.001, respectively.

Photosynthetic characteristics in cpTPI-antisense plants

Photosynthetic characteristics in cpTPI-antisense plants were examined by measuring A under a PPFD of 1,500-μmol quanta m⁻² s⁻¹ and various [CO₂] (Figure 3). On a leaf area basis, there were no large differences in A among the genotypes at intercellular CO₂ partial pressures (pCi) of 13.5 and 20 Pa (Figure 3A). In line TPIAS4, A was 86% of that in WT plants at an ambient CO₂ partial pressure (pCa) of 40 Pa and tended to decrease as [CO₂] increased. At pCa ≥ 60 Pa, A in line TPIAS4 was 61%–64% of that in WT plants. In line TPIAS7, A was similar to that in WT plants at pCa = 40 Pa and tended to slightly decrease as [CO₂] increased. At pCa ≥ 60 Pa, A in line TPIAS7 was 80%–89% of that in WT plants. There were no large differences in A between WT plants and line TPIAS14 or null plants at pCa ≥ 40 Pa.

Figure 3.

Photosynthetic characteristics in transgenic plants. The rate of CO₂ assimilation (A) in the uppermost and fully expanded leaves were measured with an LI-6400XT at a PPFD of 1,500 µmol quanta m⁻² s⁻¹, a leaf temperature of 25°C, and different CO₂ partial pressures (intercellular CO₂ partial pressure of 13.5 and 20 Pa and ambient CO₂ partial pressure of 40, 60, 80, and 120 Pa) in WT plants, null plants, and transgenic plants with antisense suppression of cpTPI (TPIAS4, TPIAS7, and TPIAS14) in rice. A was presented per unit of leaf area (A) and per unit amount of total leaf-N (A/N) (B). White circles, white diamonds, gray upward-pointing triangles, gray downward-pointing triangles, and gray squares indicate WT plants, null plants, TPIAS4, TPIAS7, and TPIAS14, respectively. Data are presented as the means ± sd (n = 4). ^{†, *, **,} and ^*** denote a significant difference in comparison with the values of A or A/N in WT plants by the Dunnett’s test at P < 0.1, P < 0.05, P < 0.01, and P < 0.001, respectively.

A is usually affected by the amount of total leaf-N. The amount of total leaf-N tended to be slightly higher in cpTPI-antisense plants (Figure 1E), possibly reflecting the decreased biomass production, at least in the case of TPIAS4 and TPIAS7 (Supplemental Figure S4B). Therefore, A per unit amount of total leaf-N (A/N) was calculated to cancel the effects of total leaf-N (Figure 3B). Although overall trends were unchanged, differences between line TPIAS14 and WT plants were significant when evaluated by A/N, as A/N in line TPIAS14 decreased to 87%–88% of that in WT plants at pCa ≥ 60 Pa. These results clearly indicate that A/N at elevated [CO₂] decreased as the amount of cpTPI decreased.

Relationships between A/N and the amount of cpTPI per unit amount of total leaf-N (cpTPI/N), instead of these parameters on a leaf area basis, were examined (Figure 4). Formulae of the regression lines and their Pearson correlation coefficients and P-values are presented in Supplemental Table S1. Significant correlations (P < 0.05) were not observed at pCi of 13.5 and 20 Pa (Figure 4, A and B). A similar trend was observed at a pCa of 40 Pa among the genotypes except TPIAS4 (Figure 4C), in which further decreases in cpTPI/N led to sharp decreases in A. In contrast, at pCa ≥ 60 Pa, strong, positive, and linear correlations were observed between these parameters among the genotypes except TPIAS4 with R > 0.85, and P < 0.001 (Figure 4, D–F; Supplemental Table S1). Decreases in cpTPI/N by 80% led to decreases in A/N at pCa = 60, 80, and 120 Pa by 19%, 22%, and 27%, respectively. In addition, the slopes of the regression lines tended to increase as [CO₂] increased (Supplemental Table S1).

Figure 4.

Relationships between A per unit amount of total leaf-N (A/N) and the amount of cpTPI per unit amount of total leaf-N (cpTPI/N). Data are taken from Figures  1, A and E, and 3B. Relationships observed at (A) intercellular CO₂ partial pressure (pCi) of 13.5 Pa; (B) pCi of 20 Pa; (C) ambient CO₂ partial pressure (pCa) of 40 Pa; (D) pCa of 60 Pa; (E) pCa of 80 Pa; and (F) pCa of 120 Pa. Symbols are the same as in Figure  3. In (A) and (B), solid lines represent the regression line obtained using all genotypes. In each part of (C) to (F), one solid line represents the regression line obtained using WT plants, null plants, TPIAS14, and TPIAS7, whereas another solid line represents the regression line obtained using TPIAS7 and TPIAS4. Formulae and Pearson correlation coefficients for these regression lines are presented in Supplemental Table S1.

Changes in the Calvin–Benson cycle and its related metabolism in cpTPI-antisense plants

As A under elevated [CO₂] decreased in cpTPI-antisense plants, the pool sizes of the Calvin–Benson cycle and photorespiration metabolites were determined at pCa = 80 Pa in lines TPIAS4 and TPIAS7 and WT plants (Figure 5). After A was equilibrated, leaves were immediately frozen using the rapid-kill system (Badger et al., 1984; Suzuki et al., 2012), followed by capillary electrophoresis-time-of-flight mass spectrometry (CE-TOFMS) analysis. The pool sizes were presented as relative values to those of WT plants on a leaf area basis. In lines TPIAS4 and TPIAS7, the pool sizes of both GAP and DHAP decreased to ∼60%–70% of those in WT plants (Figure 5, C and D). Decreases of a similar magnitude were observed for sedoheptulose 7-phosphate (S7P) (Figure 5H). Similar trends were observed for fructose 6-phosphate (F6P) (Figure 5F). Simultaneously, the pool sizes of the Calvin–Benson cycle metabolites from pentose phosphate to 3-PGA increased to 125%–188% of those in WT plants (Figure 5, A, B, I, and J). As for the metabolites involved in photorespiration, the pool sizes of glycolate, glyoxylate, and Gly tended to decrease (Figure 5, K–M), whereas those of Ser and glycerate tended to increase (Figure 5, N and O). The pool sizes of the metabolites involved in sucrose or starch synthesis tended to generally decrease in lines TPIAS4 and TPIAS7 (Table 1). The pool sizes of ATP also decreased by ∼15%, whereas significant differences were not observed for ADP and AMP (Table 2). In contrast, drastic increases were observed in the pool size of NADPH in these lines, especially in line TPIAS4 (Table 2). In addition, the pool sizes of many other metabolites were affected by the suppression of cpTPI, as presented in Supplemental Table S2.

Figure 5.

Pool sizes of the metabolites involved in the Calvin–Benson cycle and photorespiration in transgenic plants. WT rice plants and transgenic rice plants with antisense suppression of cpTPI (TPIAS4 and TPIAS7) were used as samples. The rate of CO₂ assimilation (A) in the uppermost and fully expanded leaves were monitored using a rapid-kill chamber (Badger et al., 1984; Suzuki et al., 2012) connected to an LI-6400XT at a photosynthetic photon flux density of 1,500-µmol quanta m⁻² s⁻¹, a leaf temperature of 25°C, and ambient CO₂ partial pressure of 80 Pa. After A was in a steady-state, leaves were immediately frozen with liquid N₂, followed by CE-TOFMS analysis. Pool sizes of (A) RuBP; (B) 3-phosphoglycerate (3-PGA); (C) GAP; (D) DHAP; (E) FBP; (F) F6P; (G) erythrose 4-phosphate (E4P); (H) S7P; (I) ribose 5-phosphate (R5P); (J) sum of xylulose 5-phosphate and ribulose 5-phosphate (X5P + Ru5P); (K) glycolate; (L) glyoxylate; (M) Gly; (N) Ser; (O) glycerate; (P) Gln; (Q) Glu; and (R) 2-oxoglutarate (2-OG). White bars, light gray bars, and dark gray bars indicate WT plants and lines TPIAS4 and TPIAS7, respectively. Data are presented as values relative to those in WT plants on leaf area basis (means ± sd, n = 3). ^†, ^*, and ^** denote a significant difference in comparison with the values in WT plants by the Dunnett’s test at P < 0.1, P < 0.05, and P < 0.01, respectively.

Table 1.

Relative amounts of the metabolites related to starch or sucrose synthesis in transgenic plants

Metabolite	(Relative amount m−2)
	WT	TPIAS4	TPIAS7
G6P	1.00 ± 0.17	0.67 ± 0.06*	0.67 ± 0.12*
G1P	1.00 ± 0.19	0.63 ± 0.11*	0.82 ± 0.11
ADPGlc + GDPFrc	1.00 ± 0.06	0.85 ± 0.08	0.83 ± 0.18
UDPGlc + UDPGal	1.00 ± 0.05	0.83 ± 0.09	0.80 ± 0.18
Suc6P	1.00 ± 0.10	0.68 ± 0.04**	0.77 ± 0.10*

These metabolites were determined simultaneously with those presented in Figure 5 in WT plants and transgenic plants with antisense suppression of cpTPI (TPIAS4 and TPIAS7) in rice. G6P, G1P, ADPGlc + GDPFrc, UDPGlc + UDPGal, and Suc6P represent G6P, glucose 1-phosphate, the sum of ADP-glucose and GDP-fructose, the sum of UDPG and UDP-galactose, and sucrose 6-phosphate, respectively. Data are presented as the relative values to those in WT plants on a leaf area basis (means  ± sd, n = 3). ^* and ^** denote a significant difference in comparison with the values in WT plants by the Dunnett’s test at P < 0.05 and P < 0.01, respectively.

Table 2.

Relative amounts of adenine nucleotides, NADPH, and NADP⁺ in transgenic plants

Metabolite	(Relative amount m−2)
	WT	TPIAS4	TPIAS7
ATP	1.00 ± 0.12	0.83 ± 0.06*	0.85 ± 0.06†
ADP	1.00 ± 0.07	0.86 ± 0.09	1.00 ± 0.04
AMP	1.00 ± 0.16	1.08 ± 0.20	0.96 ± 0.42
NADPH	1.00 ± 0.13	4.72 ± 0.41*	2.60 ± 1.29†
NADP+	1.00 ± 0.07	0.63 ± 0.19*	0.98 ± 0.08

These metabolites were determined simultaneously with those presented in Figure 5 in WT plants and transgenic plants with antisense suppression of cpTPI (TPIAS4 and TPIAS7) in rice. Data are presented as the relative values to those in WT plants on a leaf area basis (means ± sd, n = 3). ^† and ^* denote a significant difference in comparison with the values in WT plants by the Dunnett’s test at P < 0.1, and P < 0.05, respectively.

Discussion

Suppression of cpTPI evokes the typical symptoms of P_i-limited photosynthesis

Not only did the CO₂ sensitivity of A decrease at pCa ≥ 60 Pa in line TPIAS14, but the reverse CO₂ sensitivity of A was also observed in lines TPIAS4 and TPIAS7 (Figure 3, A and B). These symptoms are typical for P_i-limited photosynthesis (Sharkey, 1985a, 1985b; Sharkey et al. 1986a, 1986b; Sage et al. 1988; Sage and Kubien, 2007; Busch and Sage, 2017; Fabre et al., 2019; McClain and Sharkey, 2019). P_i-limited photosynthesis has been reported to induce metabolic changes, including increases in the pool size of F6P, glucose 6-phosphate (G6P), UDP-glucose (UDPG), RuBP, 3-PGA, and triose-phosphates, and decreases in the pool size of ATP and the ATP/ADP ratio (Sharkey et al. 1986a, 1988; Sage et al. 1988; Zrenner et al., 1996; Strand et al., 2000). Such changes resembled those observed in isolated chloroplasts photosynthesizing under low [P_i], in which a decrease in the pool size of triose-phosphates was also observed (Heldt et al., 1978). The metabolic changes in lines TPIAS4 and TPIAS7 were similar to these changes, as the pool sizes of the Calvin–Benson cycle metabolites from pentose phosphates to 3-PGA tended to increase, whereas the pool size of ATP decreased (Figure 5, A, B, I, and J; Table 2). These results strongly suggest that suppression of cpTPI evokes P_i-limited photosynthesis even at moderately elevated [CO₂] and normal [O₂], and that sufficient capacity of cpTPI is necessary to prevent the occurrence of P_i-limited photosynthesis.

In previous studies on the suppression of Calvin–Benson enzymes after Rubisco, such as GAPDH, aldolase, transketolase, and SBPase, RuBP-regeneration limited photosynthesis was thought to occur (Price et al., 1995; Haake et al., 1998; Ruuska et al., 2000; Harrison et al., 2001; Henkes et al., 2001). In contrast to this study, the pool sizes of RuBP and 3-PGA were observed to decrease in most of these studies and the suppression studies of phosphoribulokinase (Paul et al., 1995, 2000; Price et al., 1995; Haake et al., 1998; Ruuska et al., 2000; Henkes et al., 2001). This difference in the Calvin–Benson cycle metabolism might be one of the causes for the difference in the symptoms in photosynthesis. Interestingly, the CO₂ sensitivity of A tended to decrease in transgenic rice plants with overproduction of Rubisco (Makino and Sage, 2007), in which over-accumulation of phosphorylated compounds including 3-PGA and S7P was observed (Suzuki et al., 2012).

Possible mechanism for the occurrence of P_i-limited photosynthesis via suppression of cpTPI

Typical P_i-limited photosynthesis is thought to occur owing to a limitation in the use of triose phosphates for sucrose or starch synthesis and the regeneration of P_i via these processes (Sharkey, 1985a, 1985b; Sharkey et al., 1986a). This study suggests another mechanism for P_i-limited photosynthesis caused by a limitation of the Calvin–Benson cycle metabolism at the step of cpTPI. The mechanisms for the occurrence of P_i-limited photosynthesis have been investigated from changes in the fluxes and pool sizes of the metabolites related to Calvin–Benson cycle and sucrose or starch synthesis (Sharkey et al., 1986a; Sage et al., 1988). How P_i-limited photosynthesis was evoked via suppression of cpTPI is discussed in relation to changes in the pool sizes of the metabolites when A reached the steady state (Figure 5).

Suppression of cpTPI would initially suppress the conversion of GAP to DHAP, leading to reduced synthesis of sedoheptulose 1,7-bisphosphate (SBP) and FBP owing to reduced availability of DHAP and a limitation of the Calvin–Benson cycle metabolism. These events were likely to lead to the decrease in the pool size of S7P at the steady-state conditions (Figure 5H). The pool size of F6P also tended to slightly decrease (Figure 5F), although its interpretation is complicated as F6P is present in both in the chloroplast and cytosol (Gerhartd et al., 1987). Similar events were thought to occur in isolated chloroplasts photosynthesizing under low [P_i], as the pool size of SBP decreased (Heldt et al., 1978). Decreases in the pool sizes of FBP, F6P, SBP, and S7P were also observed in Arabidopsis mutants and transformants in which cpTPI was inhibited by over-accumulation of 2-PG (Flügel et al., 2017). In turn, C supplied for sucrose or starch synthesis was thought to be insufficient, leading to the decreases in the pool size of the intermediates for sucrose or starch synthesis at the steady-state conditions (Table 1). This event would prevent the regeneration of P_i via sucrose or starch synthesis. Transient increases in the pool size of triose phosphates might simultaneously occur, leading to sequestration of P_i. Then, ATP synthesis was thought to be limited by P_i availability, leading to the decrease in its pool size (Table 2) and the CO₂-insensitivity of A (Figure 3). As observed in the typical P_i-limited photosynthesis (Sharkey et al. 1986a, 1986b; Sage et al. 1988), the activation state of Rubisco probably decreases, leading to the increases in the pool size of pentose phosphates to RuBP (Figure 5, A, I, and J). The occurrence of Rubisco deactivation is predicted from the decreasing trends in the pool sizes of the photorespiratory metabolites, such as glycolate, glyoxylate, and Gly (Figure 5, K–M). These events have previously been observed when Rubisco capacity decreased in RBCS-antisense rice plants (Suzuki et al., 2012). In contrast, the amounts of Rubisco in lines TPIAS4 and TPIAS7 were not lower than that in WT plants (Figure 1B), suggesting the occurrence of Rubisco deactivation. The decreases in ATP availability were likely to decrease the production of triose phosphates and their pool sizes at the steady-state conditions (Figure 5, C and D). These events could also lead to the drastic increase in the pool size of NADPH (Table 2) owing to a decrease in its consumption by the GAPDH reaction. In Arabidopsis mutants and transformants with the cpTPI inhibition by over-accumulated 2-PG, the increases in the pool sizes of pentose phosphates to 3-PGA were not observed, probably because 2-PG also inhibited SBPase (Flügel et al., 2017).

Capacity of cpTPI moderately affects photosynthetic capacity at elevated [CO₂]

A/N and cpTPI/N at pCa ≥ 60 Pa were strongly and linearly correlated among the genotypes except TPIAS4 (Figure 4, D–F). A/N moderately decreased as cpTPI/N decreased within this range of cpTPI/N. The slope of the regression lines tended to increase slightly as the [CO₂] increased (Supplemental Table S1). In contrast, the decrease in the amount of cpTPI did not affect A at low [CO₂] (Figure 3A), where Rubisco is thought to limit A (Farquhar et al., 1980; von Caemmerer 2000). These results suggest that the capacity of cpTPI moderately affect the photosynthetic capacity at elevated [CO₂]. The control of cpTPI over A might be weaker than that of other Calvin–Benson cycle enzymes, such as aldolase, transketolase, and SBPase, as the effects on A were reported to be stronger when the activities of these enzymes were suppressed (Haake et al., 1998; Harrison et al., 2001; Henkes et al., 2001). To further investigate the control of cpTPI over A and to improve photosynthetic capacity, overproduction of cpTPI also need to be conducted. It has been reported that overproduction of SBPase, aldolase, and GAPDH led to an improvement in the photosynthetic capacity at elevated [CO₂] to some extent (Lefebvre et al., 2005; Driever et al., 2017; Simkin et al., 2017; Suzuki et al., 2021a), whereas no effects have been reported for transketolase (Khozaei et al., 2015; Suzuki et al., 2017) and SBPase in rice (Suzuki et al., 2019). If cpTPI has a control over A, co-overproduction of cpTPI and other Calvin–Benson cycle enzymes is also of interest for further improving photosynthetic capacity. So far, the magnitude of improvement in photosynthetic capacity seems to be rather limited, if any, when some Calvin–Benson cycle enzymes are co-overproduced (Simkin et al., 2015, 2017; Suzuki et al., 2017, 2019, 2021b), whereas co-overproduction of Rubisco and Rubisco activase was effective for improving photosynthetic capacity at low to normal [CO₂] and moderately high temperatures (Suganami et al., 2021; see also Makino, 2021).

Conclusions

In this study, suppression of cpTPI is strongly suggested to evoke the limitation of photosynthesis by P_i availability for ATP synthesis at elevated [CO₂], as judged from the responses of A to [CO₂] and the changes in the metabolism related to Calvin–Benson cycle. Enough capacity of cpTPI is suggested to be necessary to avoid P_i-limited photosynthesis at elevated [CO₂]. This is the second mechanism for the occurrence of P_i-limited photosynthesis, as it has been thought to be caused by a limitation in sucrose or starch synthesis. In addition, the photosynthetic capacity at elevated [CO₂] was moderately affected by the capacity of cpTPI. It is of interest and to be studied how overproduction of cpTPI and its co-overproduction with other Calvin–Benson cycle enzymes affect photosynthesis.

Materials and methods

Generation of transgenic plants

The gene expression of cpTPI (Os09g0535000) was suppressed in rice (O. sativa) by expressing the full-length cDNA of cpTPI in an antisense orientation under the control of the rice RBCS2 promoter (Kyozuka et al., 1993). A binary vector pBIRS (Suzuki et al., 2007) was digested with Hind III and Sac I. The rice RBCS2 promoter was amplified by PCR from pBIRS as a template using KOD-Plus-Neo (TOYOBO, Osaka, Japan). DNA fragments containing the full-length cDNA of cpTPI were amplified by RT-PCR using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific, Yokohama, Japan) and KOD-Plus-Neo (TOYOBO). The primers used are listed in Supplemental Table S3. These DNA fragments were cloned into the digested pBIRS using the In-Fusion HD Cloning Kit (Takara Bio USA, Mountain View, CA, USA) according to the manufacturer’s instructions to generate vectors for antisense suppression of cpTPI. The vector was introduced into rice (“Notohikari”) using the Agrobacterium transformation method (Toki et al., 2006). T₀ progenies of transgenic plants were grown hydroponically in an isolated and temperature-controlled greenhouse, as described by Suzuki et al. (2007). Transgenic lines with a decrease in the amount of cpTPI protein were selected by immunoblotting as described by Yamaoka et al. (2016), with some modifications as described later. Transgenic lines with a substantial decrease in the amount of cpTPI protein were selected.

Screened transgenic lines were allowed to self-fertilize so that T₁ seeds could be collected. T₁ progenies were grown hydroponically in an isolated and temperature-controlled greenhouse, as described by Suzuki et al. (2007). Homozygotes were selected before transplantation using the comparative cycle threshold method (Prior et al., 2006) as described by Ogawa et al. (2012). Cytoplasmic glutathione reductase (Os02t0813500-01) and the region from the 3′-end of the RBCS2 promoter to the 5′-end of the cpTPI fused to the RBCS2 promoter in an antisense orientation were used to detect the endogenous reference gene and transgenes, respectively. The primer pairs and probes used are listed in Supplemental Table S4. The selected T₁ progenies were allowed to self-fertilize to collect the T₂ seeds. Among the T₂ progenies, three lines with a substantial decrease in the amount of cpTPI protein (TPIAS4, TPIAS7, and TPIAS14) and null segregants were used in this study.

Plant culture and sampling

The cpTPI-antisense plants, null plants, and WT rice plants (“Notohikari”) were hydroponically grown in an environmentally controlled growth chamber (NC-411HC, NKsystem, Osaka, Japan). LED light unit and fluorescent tube lights used were described by Suzuki et al. (2021a). Seedlings were grown in tap water for 19 days after sowing. The maximum light intensity during the daytime was a PPFD of 500-µmol quanta m⁻² s⁻¹, with a photoperiod of 14 h. The maximum day and night temperatures were 25°C and 20°C, respectively. [CO₂] and relative humidity were adjusted to 40 Pa and 60%, respectively. Thereafter, one seedling was transplanted to a 1.1 L plastic pot containing nutrient solution as previously described by Makino et al. (1988). The maximum light intensity during the daytime was a PPFD of 1,000-µmol quanta m⁻² s⁻¹, with a photoperiod of 14 h. The maximum day and night temperatures were 27°C and 22°C, respectively. [CO₂] and relative humidity were adjusted to 40 Pa and 60%, respectively. The strength of the nutrient solution was changed according to plant growth as follows: 1/3 strength for 2 weeks, 1/2 strength for 1 week, 2/3 strength for 1 week, and full strength thereafter. On the 42nd day after transplanting, which corresponds to the 61st day after sowing, leaf blades, leaf sheaths, and roots were harvested separately, dried at 70°C for more than a week, and used for the measurement of plant biomass. From the 42nd day after transplanting, the uppermost, fully expanded leaves were collected after measuring A and stored at −80°C until biochemical assays. Leaf area was measured using a scanner and ImageJ software (Schneider et al., 2012). Leaves that had emerged from their sheaths by 50% were also collected for RNA analysis, as described by Suzuki et al. (2007).

Measurements of gas exchange

A was measured using a portable gas exchange system (LI-6400XT, Li-Cor, Lincoln, NE, USA). Plants were placed in an environmentally controlled growth chamber (NC-411HC, NKsystem, Osaka, Japan) operated with a PPFD of 500-µmol quanta m⁻² s⁻¹, a [CO₂] of 40 Pa, and an air temperature of 25°C for at least 30 min. After the plants were taken out from the growth chamber, measurements of A were immediately initiated at a PPFD of 1,500-µmol quanta m⁻² s⁻¹, a pCa of ∼26 Pa, a leaf temperature of 25°C, and a leaf-to-air vapor pressure difference of 1.0–1.2 kPa. After A increased and reached the steady state at pCi = 20 Pa, [CO₂] was changed to obtain pCi = 13.5 Pa, and pCa = 40, 60, 80, and 120 Pa. Gas exchange parameters were calculated according to the equations proposed by von Caemmerer and Farquhar (1981).

Biochemical assays

Frozen leaves were homogenized in a chilled mortar with pestle in 50 mM Na-phosphate (pH 7.5) containing 2-mM iodoacetic acid, 0.8% (v/v) 2-mercaptoethanol, and 5% (v/v) glycerol (Suzuki et al., 2007). An aliquot of the homogenate was used for chlorophyll determination using Arnon’s (1949,) method. Another aliquot was subjected to Kjeldahl digestion, followed by determination of total N using Nessler’s reagent (Makino et al., 1984). A part of the remaining homogenate was centrifuged at 17,900g for 5 min at 4°C. An equal volume of sodium dodecyl sulfate (SDS) sample buffer (150 mM Tris–HCl [pH 8.5] containing 2% [w/v] SDS, 20% [v/v] glycerol, and 5% [v/v] 2-mercaptoethanol) was added to the resulting supernatant. The mixture was boiled for 2 min and used for determination of cpTPI protein by immunoblotting. A sample for Rubisco determination was prepared in a similar manner, except that Triton X-100 was added to a part of the remaining homogenate at a final concentration of 0.1% (v/v). Another part of the remaining homogenate was directly treated with SDS sample buffer and centrifuged at 17,900 g for 5 min at room temperature. The resulting supernatant was used for determination of the amount of the Cyt f subunit of the cytochrome b₆/f complex by immunoblotting. These SDS-treated samples were kept at –30°C until use.

The amounts of cpTPI and Cyt f protein were determined by immunoblotting using TGX FastCast acrylamide kit with 12% (w/v) acrylamide (Bio-Rad, Hercules, CA, USA), a semi-dry blotting apparatus (Trans-Blot Turbo Transfer System; Bio-Rad), a PVDF membrane (Trans-Blot Turbo RTA Transfer Kit, Mini, PVDF; Bio-Rad), the chemiluminescence detection kits described below, and an image analyzer (ImageQuant LAS 4000, Citiva, Tokyo, Japan) as described by Yamaoka et al. (2016) and Wada et al. (2019), respectively, with slight modifications. In the case of cpTPI determination, an aliquot of an SDS-treated sample corresponding to 70 µg of leaf fresh weight was loaded onto a gel, except that the amount of sample loaded corresponded to 140 µg of leaf fresh weight in the case of lines TPIAS4 and TPIAS7. Different volumes of SDS-treated samples prepared from WT plants were loaded to construct a calibration curve. An antiserum against the fragment of the cpTPI protein [residues 149–162 (HVIGEDDQFIGKKA) of the deduced amino acid sequence of Os09t0535000-02] raised in rabbit (Sigma Aldrich Japan, Ishikari, Japan) was used for immunological detection. Signals were detected using a chemiluminescence detection kit (SuperSignal West Dura Extended Duration Substrate; Thermo Fisher Scientific, Tokyo, Japan). In the case of Cyt f determination, an aliquot of SDS-treated sample corresponding to 20 µg of leaf fresh weight was loaded onto a gel. A chemiluminescence detection kit (ECL Start, Citiva) was used. In both cases, quantification of detected signals was performed using ImageJ software (Schneider et al., 2012).

The amount of Rubisco was determined by formamide extraction of Coomassie Brilliant Blue R-250-stained bands corresponding to the large and small subunits of Rubisco separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described by Makino et al. (1985) except that calibration curves were prepared from bovine serum albumin (Pierce™ Bovine Serum Albumin Standard Ampules; Thermo Fisher Scientific).

RNA analysis

Total RNA was extracted as described by Suzuki et al. (2004), with slight modifications (Suzuki et al., 2009b). Semi-quantitative RT-PCR analysis was performed as described by Ogawa et al. (2012) and Yamaoka et al. (2016) with some modifications as described by Suzuki et al. (2021a). For determination of the transcript level of RBCS, a part of the coding region that is highly homologous among the RBCS genes expressed highly in leaf blades, that is, RBCS2, 3, 4, and 5 (Suzuki et al., 2009a), was amplified to obtain transcript levels of total RBCS, using the primer pair reported previously (Suzuki et al., 2007). Transcript levels of the transgene were also determined by semi-quantitative RT-PCR using primers that amplify the conjunction region of the RBCS2 promoter and cpTPI antisense cDNA. To determine the transcript levels of cpTPI and the transgene, calibration curves were prepared with sequentially diluted cDNA prepared from total RNA of line TPIAS4. The primer pairs used are listed in Supplemental Table S5.

Metabolome analysis

A was monitored using a rapid-kill chamber (Badger et al., 1984; Suzuki et al., 2012) connected to LI-6400XT (Li-Cor) under the conditions mentioned earlier, except that measurements were initiated at pCa = 80 Pa. Leaves were immediately frozen with liquid N₂ after A was in a steady state and stored at −80°C until analysis. Samples for CE-TOFMS analysis were prepared as described by Suzuki et al. (2012). Briefly, frozen leaves were extracted with methanol, chloroform, and water. The resultant water phases were filtrated, dried, and dissolved in ultrapure water. The samples obtained were subjected to CE-TOFMS analysis using the Agilent CE-TOFMS system (Agilent, CA, USA) as described by Soga and Heiger (2000), Soga et al. (2002, 2003), and Ooga et al. (2011).

Statistical analysis

Statistical analysis was conducted by analysis of variance followed by Dunnett’s test using R version 3.5.0 (Ihaka and Gentleman, 1996) and the package “multicomp” (version 1.4-8 (Hothorn et al., 2008).

Accession numbers

Sequence data from this article can be found in RAP-DB/GenBank under the following accession numbers: cpTPI, Os09g0535000; cytTPI, Os01g0147900; RBCS2, Os12g0274700; RBCS3, Os12g0291100; RBCS4, Os12g0292400; RBCS5, Os12g0291400; RBCL, OrsajCp033; PETA, OrsajCp041.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Table S1. Formulae and Pearson correlation coefficients (R) for the regression lines presented in Figure 5.

Supplemental Table S2. Changes in the pool sizes of metabolites determined by CE-TOFMS analysis.

Supplemental Table S3. Primer pairs used for the construction of vectors for antisense suppression of cpTPI in rice.

Supplemental Table S4. Primer pairs and probes used for the selection of homozygotes of rice plants transformed with a binary vector for antisense suppression of cpTPI.

Supplemental Table S5. Primer pairs used for semi-quantitative RT-PCR analysis.

Supplemental Figure S1. Results of immunoblotting of cpTPI protein.

Supplemental Figure S2. Results of activity staining of TPI and immunoblotting of cpTPI.

Supplemental Figure S3. Results of immunoblotting of the Cyt f subunit of the cytochrome b₆/f complex.

Supplemental Figure S4. Growth of transgenic plants on the 61st day after sowing.

Supplemental Figure S5. RT-PCR analysis for cpTPI and RBCS.

 

Y.S. conceived and designed the experiments. E.K. and M.S. generated the transgenic plants. Y.S., K.I., D.-K.Y., Y.T.-T., and E.K. performed the experiments. All authors analyzed and discussed the data. Y.S. wrote the manuscript. All authors reviewed and edited the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Yuji Suzuki (ysuzuki@iwate-u.ac.jp).