Revisiting RuBisCO

Abstract

Since the discovery of its role in the CO₂ fixation reaction in photosynthesis, RuBisCO has been one of the most extensively researched enzymes in the fields of biochemistry, molecular biology, and molecular genetics as well as conventional plant physiology, agricultural chemistry, and crop science. In addition, the RuBisCO and RuBisCO-like genes of more than 2000 organisms have been sequenced during the past 20 years. During the course of those studies, the origin of the RuBisCO gene began to be discussed. Recent studies have reported that the RuBisCO gene emerged in methanogenic bacteria long before photosynthetic organisms appeared. The origin of similar early genes might have allowed this gene to overcome changes in global environments during ancient and recent eras and to participate in the fixation of 200 GT of CO₂ annually. In this review, I focus on several points that have not been discussed at length in the literature thus far.

Graphical abstract

Proposed reductive hexulose-phosphate pathway and related metabolic processes in Archaea.

Key words:

• RuBisCO
• RuBisCO-like protein
• photosynthesis
• evolution
• methanogenic archaea

The CO₂ fixation reaction in photosynthesis was first described in a paper published in 1950 (Fig. 1).¹⁾ Afterward, general enzymatic chemistry studies were conducted, but the enzymatic properties revealed therein did not explain the gas exchange of individual plant leaves.²⁾ Since 1970, the oxygenase reaction of ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO) and its activation by CO₂ in the CO₂ fixation reaction have been clarified.^3–5) Based on these important discoveries, a biochemical model of plant photosynthesis gas exchange was constructed in 1980,⁶⁾ and the photosynthetic CO₂ fixation reaction advanced to become an integrated science discussed in fields such as biochemistry, plant physiology, crop science and ecology. In addition, RuBisCO has been studied in molecular biology for the past two decades and has been studied extensively as a field intimately related to food production and environmental management. In particular, during the past 10 years, the number of papers on RuBisCO published annually has increased, exceeding 250 papers in 2014 (Fig. 1).

Fig. 1. The number of RuBisCO-related articles published in each year after the discovery of the photosynthetic CO₂ fixation reaction in 1950.

Notes: The PubMed database was searched with the query “ribulos or rubisco.” Note that the number for the year 2017 was the number of papers searched on May 5 2017.

Since the whole picture of photosynthetic carbon metabolism was elucidated in 1970, researchers around the world have attempted to remedy important and unavoidable problems to enhance plant productivity. Given this opportunity to summarize my thinking on how to increase the CO₂ fixation ability of RuBisCO, I would like to suggest a future direction to which my research strategies have yet to venture.

I. On measures to enhance plant productivity

There are 26,000 to 35,000 genes in plant genomes, and the greatest productivity of plants is obtained when they function maximally by successfully detecting surrounding information. With this production level taken as the maximum production potential, the yield of plants under field conditions reaches only approximately 20–25% of their production potential because they are subjected to various kinds of environmental stresses.⁷⁾ Naturally, it is important to know the types of environmental stress to which plants are subjected and the response mechanisms of these plants to those types of stress.

In general, living organisms, including plants, do not always employ the best genes in their evolutionary processes. After differentiation to form unique molecular lines, organisms cannot exchange genes with species of other lineages due to reproductive isolation in mating. As described below with respect to RuBisCO, the genomic sequence of many organisms is being decoded by advances in genome science, and comprehensive consideration with physiological and biochemical knowledge⁸⁾ demonstrates that, in this reproductive isolation, plant photosynthesis has adapted to the long-term changes in the global environment from a carbonated atmosphere to an oxygenated one. Given this historical background, I believe there are two strategies to enhancing photosynthetic function. One involves directly strengthening the functions of rate-limiting metabolic steps in the photosynthetic carbon reduction cycle and increasing the catalytic activity of those steps to improve productivity potential.⁹⁾ The other strategy, which involves improving the intracellular environment so that catalytic potentials of overall rate-determining enzymes can be maximized, has been little discussed.

II. Molecular basis of plant function enhancement

II.i. CO₂ fixation

The fact that photosynthesis is uniquely restricted by the inconvenient properties of the RuBisCO enzyme is still widely accepted globally.⁹⁾ In 1970, it was found that RuBisCO also catalyzes the oxygenase reaction in addition to the carboxylase reaction of ribulose bisphosphate (RuBP).^3,10) In plants, this reaction combined with the CO₂-releasing phenomenon in photorespiration causes photosynthesis efficiency to decrease by 30–50%.¹¹⁾ In addition, the reaction rate of RuBisCO is approximately 0.1–1% of that of ordinary enzymes, and its low affinity to substrate CO₂ also causes plant photosynthesis to progress only at a slow rate. Currently, information about the structure–function relationship of RuBisCO in various organisms in nature is important for the successful protein engineering of the enzyme.¹²⁾

Photosynthetic RuBisCO is roughly classified into two forms based on amino acid sequence and protein structure.^12,13) Form I adopts the L₈S₈ structure, which is composed of eight large subunits (L) and eight small subunits (S); the green-like form I RuBisCO exists in plants, eukaryote green algae, cyanobacteria and photosynthetic bacteria, and the red-like form I RuBisCO exists in red algae and some bacteria. Form II adopts the L_2n structure, which is composed only of L subunits, and exists in photosynthetic bacteria, chemical bacteria and dinoflagellates. The enzymatic properties of RuBisCO vary greatly depending on the evolutionary lineage of the species. The k^C_cat of the carboxylase reaction of higher plant RuBisCO, which includes the green-like RuBisCOs of form I, ranges from 2.9 to 4.2 catalytic site⁻¹ s⁻¹; for green algae enzymes, this value ranges from 2.3 to 5.4 site⁻¹ s⁻¹, and for cyanobacterial enzymes 2.6 to 13.2 site⁻¹ s⁻¹. However, red-like RuBisCOs have a k^C_cat ranging from 1.3 to 3.4 site⁻¹ s⁻¹, and in photosynthetic bacteria, red-like RuBisCOs have a value ranging from 2 to 3.2 site⁻¹ s⁻¹. Additionally, the k^C_cat is 3.5 to 7.3 for the photosynthetic bacterial enzyme, which is form II. In particular, the CO₂/O₂ specificity factor (or relative specificity, S_rel), which is an important characterization among the kinetic parameters of RuBisCO, is (k^C_cat/K^C_m)/(k^O_cat/K^O_m), where k^C_cat and k^O_cat are the maximum carboxylase reaction and oxygenase reaction rates, respectively, and K^C_m and K^O_m are the km values for CO₂ and O₂, respectively.¹⁴⁾ RuBisCO with a high S_rel has a strong ability to discriminate in favor of CO₂ in a reaction system containing both CO₂ and O₂, meaning that it is specialized in the carboxylase reaction compared with the oxygenase reaction. The S_rel of higher plant RuBisCO, which is the green-like form I RuBisCO, ranges from 70 to 94; this range is lower in the green algal form I (54 to 83) and further lower in the cyanobacterial RuBisCO (38 to 56).^12–14) Additionally, the S_rel of photosynthetic bacteria ranges from 26 to 53. In contrast, the S_rel of the red-like RuBisCO of photosynthetic bacteria ranges from 105 to 238; in particular, the RuBisCO of red algae shows higher values than does green-like RuBisCO. RuBisCO from the primitive red alga Galdieria partita shows the highest S_rel among RuBisCOs that have been found thus far.¹⁵⁾ Note that increasing affinity for CO₂ is an important factor in an increase in S_rel. However, as a result, the CO₂ fixation rate decreases. Since RuBisCO does not have binding sites for the gaseous substrates,¹³⁾ it has been proposed that the observed higher S_rel values are due to the characteristics of the enzyme that bind the carboxylase reaction intermediate to the catalytic site much more tightly than the oxygenase reaction intermediate. Therefore, there is a so-called tradeoff between affinity for CO₂ and reaction rate (Fig. 2).⁸⁾ This trade-off is why the S_rel value has a positive relationship with the reactivity of the carboxylase reaction but has the opposite relationship with the reactivity of the oxygenase reaction.

Fig. 2. Natural variation in the kinetic parameters of RuBisCO measured in vitro at 25 °C.⁸⁾

Notes: The organisms from which RuBisCOs were isolated included the following: (1) Rhodospirillum rubrum, (2) Chromatium vinosum, (3) Synechococcus PCC 6301, (4) Synechococcus PCC 7002, (5) Chlamydomonas reinhardtii, (6) Sorghum bicolor, (7) Zea mays, (8) Amaranthus hybridus, (9) Spinach oleracea, (10) Nicotiana tabacum, (11) Triticum aestivum, (12) Phaeodactylum tricornutum, (13) Griffithsia monilis, and (14) Galdieria sulphuraria. Dashed lines are linear regressions through all the data except for cyanobacterial outliers (3) and (4) both in (a) and (b), which were excluded. (Reprinted with permission from the publisher).

Thus, the fact that the S_rel of RuBisCO differs depending on the species is believed to be indicative of the adaptation of the properties of RuBisCO to the environment in which each organism evolved on Earth. Indeed, in photosynthetic organisms such as bacteria that have RuBisCOs showing lower S_rel than plant RuBisCO, many grow in anaerobic environments; in aerobic environments, these lower-S_rel RuBisCOs are subjected to the inhibition of carboxylase activity by oxygen. Algae attained an ability to induce the CO₂ concentrating mechanism (CCM) to increase intracellular CO₂ concentrations to inhibit the oxygenase reaction.¹⁶⁾ However, in higher plants, it appears that molecular evolution increased the S_rel of RuBisCO to fix CO₂ even on land where O₂ is present in greater amounts than CO₂.¹²⁾ Furthermore, in C₄ plants, by evolving a metabolic CCM—the so-called C₄ pathway—the CO₂ concentration around RuBisCO is elevated.

In fact, we can simulate how the S_rel of RuBisCO affects plant photosynthesis using the photosynthetic model of Farquhar.¹⁷⁾ In this model, there are two phases in the dependency of CO₂ fixation rate on the intercellular CO₂ concentration. One is the RuBisCO activity-limiting phase under low CO₂ concentrations; as the CO₂ concentration increases, the dependency shifts to the RuBP regeneration rate-determining phase. Using this model, a simulation of the CO₂ fixation rate in plant leaves with high-S_rel RuBisCO is shown in Fig. 3.¹²⁾ Compared with the wild type, when the S_rel of RuBisCO increases by a factor of 2, the CO₂ fixation rate is expected to increase at all CO₂ concentrations, and the CO₂ compensation point decreases from 50 to 30 μbar. In particular, the increase in S_rel either by increasing k^C_cat or by decreasing K^C_m causes a dramatic increase in the CO₂ fixation rate under low CO₂ concentrations. As such, in a plant using RuBisCO that has such a high S_rel, the CO₂ concentration necessary for achieving the same CO₂ fixation rate as wild type is lower, and the water use efficiency in photosynthesis is higher. These simulation results suggest that increasing the S_rel of RuBisCO is an important point for improving plant photosynthetic CO₂ fixation rate.^8,12)

Fig. 3. Effect of the properties of RuBisCO on photosynthetic CO₂ assimilation rates.¹²⁾

Notes: Photosynthetic CO₂ assimilation rate (A) calculated in accordance with the methods of Farquhar et al.⁶⁾ The CO₂ assimilation rate in the RuBisCO activity-limited phase at low CO₂ pressure is expressed as follows: A = [RuBisCO]·k^C_cat·(p_c·s_c – 0.5·[O₂]/S_rel)/{p_c·s_c + K^C_m (1 + [O₂]/K^O_m)} – Rd, where [RuBisCO] is the mole number of active sites of RuBisCO per unit leaf area, p_c is the intercellular CO₂ partial pressure, s_c is the CO₂ solubility in water (0.0334 M bar⁻¹), [O₂] is the O₂ concentration in the chloroplast (252 μM), and Rd is the daytime respiration rate (1 μmol m⁻²). The CO₂ assimilation rate in the RuBP regeneration-limited phase at high CO₂ pressure was obtained using the following equation: A = J·(p_c·s_c – 0.5·[O₂]/ S_rel)/4·(p_c·s_c + [O₂]/ S_rel) – Rd, where J is the photosynthetic electron flux of 120 μmol m⁻² s⁻¹. (a) The kinetic parameters of wild-type tobacco RuBisCO used in the simulation were as follows: 3.4 s⁻¹ site⁻¹ for k^C_cat, 10.7 μM for K^C_m, 0.87 s⁻¹ site⁻¹ for k^O_cat, and 295 μM for K^O_m (solid line). For simulating doubled S_rel values, the kinetic parameters were changed as follows: either k^C_cat (closed squares) or K^O_m (opened circles) was doubled, or either K^C_m (closed circles) or k^O_cat (opened squares) was halved. (b) CO₂ assimilation rates of red algal RuBisCO from Griffithsia monilis (closed triangles). This figure was adapted from reference 17 with permission from the publisher. The kinetic parameters were as follows: 2.6 s⁻¹ site⁻¹ for k^C_cat, 2.6 μM for K^C_m, 710 μM for K^O_m, and 167 for S_rel. In the case of wild-type tobacco RuBisCO (solid line), the data are the same as those in (a). In all calculations, the mole number of the active sites of RuBisCO per unit leaf area was set to 20 μmol m⁻².

The carboxylase reaction of RuBisCO proceeds sequentially as follows. Substrate RuBP binds to the activated RuBisCO reaction center, and substrate CO₂ is added to the C₂ of enediol-type RuBP, which is generated by the enzymatic withdrawal of the C3 proton of RuBP, after which 3-keto-2-carboxyarabinitol bisphosphate is produced. In the case of the oxygenase reaction, 3-keto-2-peroxyarabinitol bisphosphate is generated.¹³⁾ Because there are no binding sites for these gaseous substrates in the RuBisCO protein, the addition rate of the gaseous substrates should depend simply on the concentrations of the substrates, and the reported large difference in S_rel among RuBisCOs of different origins cannot be explained by the affinity of the enzyme to the gaseous substrates.^8,12) These major differences depend on the process by which reaction products are formed after the gaseous substrate binds.

II.ii. Activation of RuBisCO

For all RuBisCO enzymes to show activity, a CO₂ molecule binds to the ε-amino group of Lys-201 to form a carbamyl anion. This CO₂ molecule is involved only in the activation and does not become the substrate. Mg²⁺ is further coordinated to the carbamyl anion.^4,5,18) Thereafter, an enzymatically active RuBisCO Lys-201-NH-COO^–-Mg²⁺ (ECM) ternary complex is formed, and the complex is stabilized by the carboxylate groups of Asp-203 and Glu-204. ^19,20) Mg²⁺ and the activator CO₂ play important roles, such as withdrawing H⁺ from the C₃ of the substrate RuBP and stabilizing the carboxylate of the CO₂-fixing reaction intermediate.²⁰⁾ To activate RuBisCO in vitro, 20 mM Mg²⁺ and 0.1 to 0.2 mM CO₂ are both necessary, but the stromal CO₂ concentration during photosynthesis can be as low as 5 to 7 μM,²¹⁾ and the reported high RuBisCO activation levels cannot be explained without one or more effectors.

Ogren’s group at the University of Illinois explained this gap eloquently. RuBisCO activase, which belongs to the AAA + family, is involved in the activation of RuBisCO and the removal of endogenous inhibitors that bind strongly to RuBisCO.^22,23) Interestingly, in the presence of RuBisCO activase, even in higher plants under saturated light in normal atmosphere, RuBisCO is 80–90% activated. Increasing amounts of RuBisCO by sufficient use of nitrogen fertilizer and overexpression of the S subunit gene of RuBisCO in plant leaves decreases this activation ratio.^24,25) In contrast, if the amount of RuBisCO is forcibly reduced by the antisense method, the activation ratio of RuBisCO increases.²⁶⁾ Moreover, increasing the amount of activase in rice decreases the content of RuBisCO and decreases the CO₂ fixation activity.²⁷⁾ The mechanism by which such trade-offs involving RuBisCO occur in these cases has not yet been described.

III. Enhancement of photosynthetic function by promoting the activation of RuBisCO

III.i. Details of the activation mechanism

The activation reaction of RuBisCO can be described as follows. The RuBisCO protomer binds activator CO₂ on the ε-amino group of Lys-201 to make a carbamate ion. Mg²⁺ ion binds to the carbamate ion to form the ECM complex, where E is the RuBisCO protomer, C is the activator CO₂, and M is Mg²⁺. The carbamate ion formation is very slow, and the binding of Mg²⁺ proceeds very quickly.⁵⁾ When 44 μM RuBisCO protomer was activated with 112 μM CO₂ at 10 °C, the activation reached an equilibrium in 10 min in activation experiments with RuBisCO.⁵⁾ The rate constant of CO₂ binding to the amino group of the model compound monoethanolamine (MEA) is 6.1 × 10³ M⁻¹ s⁻¹ at 30 °C.²⁸⁾ Calculated using the rate constant for MEA carbamylation, the equilibrium would have been reached in 3.2 s with the concentrations of the RuBisCO protomer and CO₂ used by Lorimer et al.⁵⁾ The primary reaction rate at which CO₂ is liberated from MEA carbamate is 29.8 s⁻¹, and its half-life is 0.023 s, while the half-life at which activated RuBisCO is inactivated under inactivation conditions is 20 s (Fig. 4). The difference between the carbamylation rate of MEA and the ε-amino group of Lys-201 cannot be explained by the difference in reaction temperatures used by these groups.^25,28) Interesting research results related to this discrepancy have been reported by Vater et al.²⁹⁾ In their analysis using p-toluidinonaphthalene sulfonate, the fluorescence emission from the sulfonate reflects the binding of the probe to the hydrophobic regions of RuBisCO. These authors found that an increase in fluorescence emission reflects a conformational change of the RuBisCO protein by the binding of either CO₂ or Mg²⁺ or both to RuBisCO; this conformational change may be important in RuBisCO activation, as discussed later.

Fig. 4. Effector-dependent suppression of the deactivation of RuBisCO.

Notes: This figure is a recasting of Fig. 4 of reference 41 and includes additional calculations as explained in the text. (a) The enzyme (approximately 7 mg mL⁻¹) was first activated in Bicine buffer (pH 8.2) containing 20 mM MgCl₂ and 5 mM NaHCO₃. At time zero, 5 μL (35 μg of enzyme) of the solution was added to 400 μL of Bicine buffer (pH 7.8) containing 1 mM EDTA and 0.6 mM NaH¹⁴CO₃ at 10 °C (line 1). Where appropriate, this mixture contained 0.1 mM 6PG (line 2) or 0.2 mM NADPH (line 3). At the indicated times, 10 ml of 20 mM RuBP was introduced to measure the remaining activity at that time for 30 s. A_t is the activity at time t, and A_i is the initial activity. (b) This panel replotted the reaction courses of lines 1 and 3 for the initial 20 s. Line 4 is the course of the chemical reaction of CO₂ released from MEA carbamate as calculated by the following equation: A_t/A_i = e^–kt, where k is the rate constant of the liberation of CO₂ from MAE carbamate and where t is the time in seconds. A_t is the concentration of MEA carbamate remaining at time t, and A_i is the initial concetration. See the text for details. (Reprinted with permission from the publisher).

III.ii. RuBisCO activation and RuBisCO activase

Discoveries in numerous pioneering studies of the structure-function relationship of RuBisCO have clarified the formation of an ECM:Glu-203:Asp-204 triangle contained in the periphery of Lys-201, and the tightly fixed ECM allows RuBisCO to function and produce a CO₂ fixation reaction.²⁰⁾ It is also important that RuBisCO activase is deeply involved in this process. Recent reports from the Max Planck Research Institute have posed important challenges to the mechanism of this activation.^30,31)

For some time now, it has been known that the concentration of RuBisCO activase in chloroplasts is considerably lower than that of RuBisCO.³²⁾ When RuBisCO activase exerts its function on RuBisCO, the activase functions by forming a hexamer.^30,31) The RuBisCO holoenzyme (L₈S₈) content is 3 μmol m⁻², and the hexameric activase (A₆) content is 0.57 μmol m⁻².³³⁾ Mate et al.³⁴⁾ determined these values to be 2.5 and 0.15 μmol m⁻², respectively. In antisense experiments for the activase, the gas exchange rate was not affected until the activase amount was reduced to 1/5 of the wild-type level;^33–35) therefore, the amount of the activase required for 1 μmol m⁻² of the RuBisCO holoenzyme ranges from 0.04 to 0.06 μmol activase hexamer m⁻² for photosynthesis to proceed at normal speed.

In the in vitro activation of RuBisCO from the ER form, in which the RuBisCO holoenzyme concentration and activase hexamer concentration are almost the same, nearly full activity was attained.³⁶⁾ The reason for the difference in the required concentrations of the activase in vivo and in vitro is an important research target.

III.iii. Activation in the presence of sugar phosphates and reduced nicotinamide adenine dinucleotide phosphate

Several sugar phosphates have an interesting influence on the activation of RuBisCO. Glucose 6-phosphate and ribulose 5-phosphate are considered negative effectors and have no effect on the activation, but 6-phosphogluconate (6PG), fructose 1,5-bisphosphate (FBP), reduced nicotinamide adenine dinucleotide phosphate (NADPH) and sodium phosphate promote the activation of RuBisCO and are considered positive effectors.^37–39) The RuBisCO substrate RuBP itself also contributes to the prolonged binding of the activator CO₂ to Lys-201.⁴⁰⁾ In a detailed analysis with 6PG,⁴¹⁾ the K_i based on competitive inhibition kinetics against RuBP was 8.5 μM, while the dissociation constant in the equilibrium binding analysis was 37 μM. Plotting of the concentration of 6PG toward the activation level of RuBisCO produced a saturation curve, and the obtained half-saturation concentration for activation was 50 μM (Fig. 5).

Fig. 5. Double reciprocal plots between the preincubation concentration of the effectors on the activation of RuBisCO activity.

Notes: This figure is a recasting of Fig. 1 of reference 41 under permission from the publisher. The enzyme (0.1 mg mL⁻¹) was prepared in 100 mM Bicine (pH 7.8), 1 mM MgCl₂, and 1 mM NaHCO₃. Aliquots (10 μL) were added to tapered 1-ml reaction vials, and 1 μL of the appropriate effector was added to bring the final concentration to that indicated in the figure. The enzyme was then incubated for 30 min at 25 °C; the activity during this time was assayed by the addition of 300 μLl of reaction mixture [l00 mM Bicine (pH 7.8), 20 mM MgC12.0 mM NaH¹⁴CO₃ (59,000 dpm nmol⁻¹), and 0.3 mM RuBP] equilibrated to 25 °C, and the assays were then run for 20 s.

The km of RuBisCO for RuBP is 20 μM.⁴¹⁾ This implies that 6PG merely binds to the RuBP binding site competitively with respect to RuBP but adopts a binding mode that differs from that of the competitive inhibition mode in the absence of RuBP. NADPH is likewise a competitive inhibitor against RuBP,⁴¹⁾ but there is no difference between the K_i and the half-saturation concentration in the activation of NADPH unlike 6PG. The binding mode of NADPH may be different from that of RuBP, as discussed below. This point was also consistent with the results from the structural analysis of the complexes of rice RuBisCO and its positive effectors. In rice RuBisCO crystallized in the presence of 40 mM 6PG or 40 mM NADPH as well as atmospheric CO₂ and 40 mM MgCl₂, the active site structures were markedly different from those deduced from the above kinetic analyses.⁴²⁾ Sugar phosphate was bound to the catalytic site, where the ε-amino group of Lys-201 was carbamylated, and further Mg²⁺ was bound to the carbamate anion. The position of the carbamyl-group oxygen of Lys-201 stabilizing this Mg²⁺ and the positions of the Asp-203 and Glu-204 carboxyl group oxygens in the 6PG-binding enzyme form closely overlapped with those in the NADPH-binding form. The structures of the rice RuBisCO complexed with 6PG and NADPH are very similar to the catalytic center structure of spinach RuBisCO in the activated state. One small difference between these two enzymes is that the carboxyl groups of 6PG and the oxygen of the hydroxyl group on C2 are directly coordinated to Mg²⁺ in the former, but in the NADPH-binding enzyme, the adenylate phosphate oxygen and the oxygen of adenylate ribose C3 stabilize Mg²⁺ via two water molecules (Fig. 6). However, in this state, where two phosphate groups of RuBP are more than 9.4 Å apart, RuBisCO displays open-state conformations. In a closed-state conformation, where two phosphate groups are 9.1 Å apart or less, as seen in the quaternary complex between ECM and the transition state analog 2-carboxyarabinitol bisphosphate, the N-loop, 60s loop, loop 6 and C-loop of L are in ordered states. However, in the NADPH-binding form, these loops are in disordered states. In the 6PG-binding form, only the 60s loop is ordered.

Fig. 6. Structures of the catalytic sites of (a) 2-CABP, (b) 6PG-, and (c) NADPH-bound RuBisCOs with electron density maps of the ligands.⁴²⁾

Notes: The three ligands and amino acid residues coordinated to Mg²⁺ are depicted as stick models, where C atoms are yellow, N atoms are blue, O atoms are red, and P atoms are magenta. The red spheres denote Mg²⁺, and the blue spheres denote water molecules (W). The blue broken lines denote hydrogen bonds. The σA-weighted Fo−Fc electron density maps (contoured at 3.5 σ in blue and 6.0 σ in red for NADPH and at 4.0 σ in blue for both 6PG and 2CABP) were calculated after the omission of the ligand moiety from each model. The figures are viewed from the same angle as in (a). (Reprinted with permission from the publisher).

It is interesting to analyze the functions regarding the activation of 6PG and NADPH from deactivation kinetics (Fig. 4). When activated RuBisCO is diluted with buffer devoid of CO₂ or with the ethylenediaminetetraacetate (EDTA)-containing buffer, deactivation of RuBisCO proceeds with a long half-time of approximately 20 s.⁴¹⁾ This may be because the conformational change induced by the binding of activator CO₂ and Mg²⁺ to RuBisCO stabilizes this activated conformation even in the presence of EDTA. Then, the addition of 6PG or NADPH to this RuBisCO deactivation buffer containing EDTA caused the deactivation half-time to increase from 20 s to 1800 and 300 s, respectively. In other words, these positive effectors maintained RuBisCO in the activated conformation for a longer period, even in the presence of EDTA.

Another point to discuss is that the concentration of NADPH in chloroplasts is much lower than the concentration of the RuBisCO protomer.⁴³⁾ As 6PG is synthesized only at night,⁴⁴⁾ this will be discussed later in detail.

III.iv. RuBisCO: a conformation-memorizing enzyme?

As discussed above regarding references 5 and 28, the carbamate formation of the ε-amino group of Lys-201 for the activation of RuBisCO requires more than 10 min in the presence of 44 μM RuBisCO protomer, 112 μM CO₂, and 20 mM MgCl₂, while the chemical reaction of 44 μM MEA and 112 μM CO₂ reached the equilibrium in 3.2 s; meanwhile, CO₂ liberation from carbamylated MEA proceeded 870 times faster than the decarbamylation reaction of the activated enzyme. Such slow activation/deactivation empirically measured for RuBisCO can be reasonably explained if we assume that the conformational change of the RuBisCO protein is involved in these processes as discussed above. The detailed analysis of RuBisCO activation/deactivation by Badger and Lorimer⁴¹⁾ showed that when transferred to EDTA-containing buffer, RuBisCO activated in the presence of NADPH was decarbamylated 13,000 times slower than the chemical decarbamylation (Fig. 4(b)). When RuBisCO activated by NADPH was placed in a reaction medium containing saturated concentrations of RuBP and CO₂ in the presence of EDTA, RuBisCO showed full activity without any delay, indicating that the substrate RuBP quickly removed NADPH from the catalytic site, but surprisingly, exogenously added EDTA could not deprive Mg²⁺ or of course activator CO₂ from the ECM complex. In other words, activated RuBisCO must have the capability to maintain or memorize its activated conformation for some period of time. It has been thought that the reported concentration of NADPH (0.25–0.5 mM)⁴³⁾ was too low compared with the concentration of the RuBisCO protomer in chloroplasts (4 mM). FBP and phosphate are reported to be present at approximately 5.4 and 12 mM, respectively, in chloroplasts^45,46), and may similarly maintain the activated state of RuBisCO as does NADPH.^37–39)

The minimum required amount of RuBisCO activase hexamer required for 1 μmol m⁻² of the RuBisCO holoenzyme in chloroplasts ranges from approximately 0.04 to 0.06 μmol m⁻² RuBisCO activase, as discussed above. Activase is necessary not only for activating RuBisCO but also for removing various sugar phosphate inhibitors from the RuBisCO active center. It is conceivable that the conformation-memorizing characteristics of the activated state of RuBisCO play an important role to exert this function even with lower concentrations of RuBisCO activase. Considering these points, I propose that the activation state may be maintained for a while in vivo, aided by various positive effectors and by the conformation-memorizing character of activated-form RuBisCO.

III.v. Accelerating photosynthetic functions by optimizing the functional environment of RuBisCO in chloroplasts

To clarify the rate-determining step and the rate-limiting factor of the Calvin cycle that affect photosynthetic ability and carbon partitioning, a number of transformants with suppressed expression of various enzymes have been produced by the antisense method.⁴⁷⁾ Moderate suppression of glyceraldehyde 3-phosphate dehydrogenase,⁴⁸⁾ phosphoribulokinase,⁴⁹⁾ aldolase,⁵⁰⁾ and fructose 1,6-bisphosphatase (FBPase)⁵¹⁾ did not cause photosynthesis to decrease. These results indicate that many enzymes of the Calvin cycle are present in sufficient quantities to maintain photosynthetic capacity at the protein level. On the other hand, suppression of RuBisCO expression is directly linked to a decrease in photosynthetic ability under light saturation, but under unsaturated light conditions, RuBisCO is not the only rate-limiting factor for photosynthetic performance.^52,53)

In antisense plants with sedoheptulose 1,7-bisphosphate (SBPase),⁵⁴⁾ the photosynthetic ability decreases proportionally to a decrease in SBPase activity. In lines in which the SBPase activity decreased to 29% of that of wild type, the photosynthetic activity decreased to 36% of that of wild type. Based on these reports, it is clear that the reduction in RuBP regeneration ability in the Calvin cycle affects carbon assimilation and growth, and in particular, the change in SBPase level has a significant effect on those Calvin cycle processes.

Information from such studies has encouraged attempts to overexpress the genes of these enzymes. Cyanobacteria have a unique Calvin cycle enzyme not present in higher plants.⁵⁵⁾ Since this enzyme catalyzes the reactions of both FBPase and SBPase with one protein, it was named fructose 1,6-/sedoheptulose 1,7-bisphosphatase (FBP/SBPase). In actuality, when the activity of both of these enzymes in chloroplasts was increased by introducing the cyanobacterial FBP/SBPase gene into tobacco chloroplasts via Agrobacterium⁵⁶⁾ or chloroplast DNA⁵⁷⁾ directly, the photosynthetic CO₂ fixation activity of individual leaves under a light intensity of 400 μmol photons m⁻² s⁻¹ increased by 30 to 50%, and the total biomass of the plant increased by 50 to 80%. Little effect of the introduced FBP/SBPase on the photosynthetic activity was observed at 200 μmol m⁻² s⁻¹ or less, and the promotion of photosynthesis was pronounced at 700 μmol m⁻² s⁻¹ or more. In transformed plants with introduced FBP/SBPase, the RuBP concentration almost doubled, and the activation ratio of RuBisCO increased by 30%. I believe that this is one of the causes of increased photosynthetic activity in the transformant. The activity of RuBisCO in chloroplasts is controlled by RuBisCO activase as mentioned above. A high concentration of RuBP is necessary for the function of RuBisCO activase,⁵⁸⁾ and the phenotype of this transformed plant seems to be due to the activation of the RuBP regeneration system by introduction of FBP/SBPase. Much higher concentrations of RuBP also further increase the activity of RuBisCO.⁵⁹⁾ The K_m value of RuBisCO for RuBP is approximately 20 μM in the catalytic reaction, and the maximum activity is obtained between 0.5 and 1 mM. We found that when the RuBP concentration was 2 mM or higher, the activity further increased by 30%. In actuality, in biochemical analyses by equilibrium ligand binding and conformational change with p-toluidinylnaphthalene sulfonate, RuBisCO had one activity-regulating site per protomer that binds various sugar phosphates in addition to the catalytic site.^60,61) Also, reported stromal concentrations of RuBP have sometimes exceeded 6 mM.^45,62)

A similar photosynthesis-promoting effect has also been observed when the NAD kinase-2 gene was introduced into chloroplasts of rice via Agrobacterium. The introduction of NAD kinase-2 caused the concentration of NADPH to increase to levels 1.5 times higher than those in the non-transformant and the concentration of RuBP to increase 2.7-fold.⁶³⁾ It seems that both of these increase the activation of RuBisCO as discussed in this review, although the authors of this paper⁶³⁾ presented a different way of thinking.

IV. A challenge regarding the area that we have not been able step into thus far: RuBisCO is still plagued by its origin

Even today, when the nucleotide sequences of the genomes of more than 4000 species have been completely decoded, research fields that make good use of these large data remain to be developed. For example, the photosystem II protein D1 was an enzyme that appeared when photosynthetic organisms attained the oxygen-generating function, but the gene located at the root of the phylogenetic tree of the D1 gene has not been found in non-photosynthetic organisms.⁶⁴⁾ An ancient oxygen-generating photosynthetic organism must have appeared using such a gene.

In the case of the RuBisCO rbcL gene, until approximately 1990, the phylogenetic root of the rbcL gene was considered to be the cyanobacterial rbcL, similarly to the D1 gene, and the tree only determined the position of evolution among photosynthetic organisms. When the existence of archaea was proposed and the genome of archaea began to be deciphered, an authentic archaean rbcL gene, the gene of an enzyme that can fix CO₂ on its acceptor RuBP, was found in 1999.^65,66) In 2007 the metabolic function of this gene became clear.⁶⁷⁾ Further decoding of archaebacterial and eubacterial genomes revealed a great number of genes that encode proteins lacking some residues essential for RuBisCO reactions but that were very similar to RuBisCO in other ways. The phylogenetic tree of these rbcL-related genes revealed the occurrence of an enormous and diverse rbcL-related gene family different from forms I to III of the rbcL families. These genes constituted form IV, which was distinct from the form I to III families. The protein of the form IV family was named RuBisCO-like protein (RLP). The functional relationship between forms I to III and form IV became the next intriguing question.⁶⁸⁾

Bacillus subtilis RLP (bsRLP), which belongs to the form IV family, has only 11 residues among the 19 catalytic residues of RuBisCO, and its amino acid sequence shares only 18.3% homology with that of a form II RuBisCO of the photosynthetic bacterium Rhodospirillum rubrum. We found that bsRLP catalyzed the enolization reaction between C1 and C2 of 2,3-diketo-5-methylthiopentyl-1-phosphate, which is structurally very similar to RuBP.⁶⁹⁾ This finding revealed the overall functions of the members of the enormous RuBisCO rbcL gene families that had been clarified in previous genome decoding studies.⁷⁰⁾ However, the origin of the rbcL gene families, including bsRLP, remains to be clarified.

Tabita et al.⁷¹⁾ reported an interesting point by creating an evolutionary phylogenetic tree of rbcL family genes using four software packages for the evolutionary phylogenetic tree generation. The form III rbcL gene consisted of two subfamily forms, III-1 and III-2, but there was a clade consisting of other form III rbcL genes in the vicinity of the III-1 and III-2 groups. The newly found group had the gene for the eukaryotic phosphoribulokinase (PRK).⁷²⁾ In our most recent findings, photosynthetic PRK genes were classified into two groups: one from photosynthetic bacteria and the other from plant-type oxygenic phototrophs, such as plants, algae, and cyanobacteria.⁷³⁾ PRK homologs from methanogenic archaea including Methanospirillum hungatei showed approximately 30% amino acid sequence identity with PRKs in photosynthetic organisms and formed a clade clearly distinct from those of bacterial and plant-type PRKs in a phylogenetic tree. The enzymatic properties of the archaeal PRKs were different from those of photosynthetic bacteria and plant-type oxygenic phototrophs. The archaeal form III consisted of two subfamilies: III-a and III-b. All methanogenic archaea included in the III-a and III-b forms have the RuBisCO gene and compose the primitive Calvin pathway, also known as “the reductive hexulose-phosphate (RHP) pathway.”⁷³⁾ Because this pathway is the main pathway for methane synthesis, formaldehyde (FA), which would be provided from the RHP pathway, is likely to have also played an important role (Fig. 7).

Fig. 7. Proposed RHP pathway and related metabolic processes in archaea.

Notes: The RHP pathway (highlighted in yellow) has metabolic links to methanogenesis via methylene-H₄MPT, which may be synthesized with formaldehyde released from the RHP pathway. The successive black arrows show multiple reaction steps. The RHP pathway (red lines and arrows) is superimposed on the Calvin cycle (green lines and arrows), and reaction steps from Ru5P to F6P are common in both cycles. Missing Calvin cycle steps in M. hungatei are indicated by gray dashed lines and arrows. Ru5P, ribulose-5-phosphate; RuBP, ribulose-1,5-bisphosphate; 3-PGA, 3-phosphoglycerate; BPG, 1,3-diphosphoglycerate; GAP, glyceraldehyde-3-phosphate; DHAP, dihydroxyacetone phosphate; FBP, fructose-1,6-bisphosphate; F6P, fructose-6-phosphate; Hu6P, D-arabino-3-hexulose-6-phosphate; FA, formaldehyde; E4P, erythrose-4-phosphate; Xu5P, xylulose-5-phosphate; SBP, sedoheptulose-1,7-bisphosphate; S7P, sedoheptulose-7-phosphate; R5P, ribose-5-phosphate; NMP, nucleoside 50-monophosphate; RiBP, ribose-1,5-bisphosphate; H4MPT, tetrahydromethanopterin; MF, methanofuran; PGK, 3-phosphoglycerate kinase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; FBPase, fructose-1,6-bisphosphatase; TK, transketolase; SBPase, sedoheptulose-1,7-bisphosphatase; RPE, ribulose-5-phosphate 3-epimerase; RPI, ribose-5-phosphate isomerase; AMPpase, AMP phosphorylase; RiBP isomerase, ribose-1,5-bisphosphate isomerase. (The RHP pathway was reprinted with permission from the publisher).

In primeval times, where methane-producing archaea began to synthesize methane, the atmosphere was almost saturated with CO₂ and almost free of oxygen.⁷⁴⁾ Under such circumstances, methanogenic archaea synthesize N-carboxymethanofuran (carbamate) by a chemical reaction⁷⁴⁾ rather than by an enzymatic reaction of CO₂ at a concentration of several tens of millimoles per liter, and methane is formed by a consecutive reduction of N-carboxymethanofuran (carbamate) by hydrogen gas (Fig. 7). As described above, the reaction rate constant of MEA and CO₂ is 6.1 × 10³ M⁻¹ s⁻¹ at 30 °C. Since the CO₂ concentration in water in equilibrium with 100% atmospheric CO₂ is 23 mM, assuming that the intracellular concentration of methanofuran is 1 μM, the formation rate of N-methanofuran carbamate is 1 μmol s⁻¹.⁷⁵⁾ RuBisCO, which appeared in the ancient environment that contained high concentrations of CO₂ without oxygen, may not have required any catalytic residue for the carbamylation of the ε-amino group of Lys-201 and for the chemical addition of gaseous CO₂ on the C2 of the endiolate of RuBP.²⁰⁾ I suppose that all of the extant RuBisCOs perform these two chemical reactions with CO₂ during photosynthesis and that this inconvenience is forcing this enzyme and photosynthetic organisms to pay a large cost.

V. Epilogue

Through the subsequent appearance and exuberance of oxygenic photosynthesis, the concentration of oxygen in the atmosphere has reached a high of 210,000 ppm, and the atmospheric CO₂ concentration has decreased to 0.04% or less. In this review, I have touched on the history of the bold adaptations of RuBisCO to these changes. As a result, RuBisCO has suffered a tradeoff of a drastic decrease in k^C_cat but has overcome this adversity via the algae CCM. Higher plants mainly strengthened the promoters of multi-copy rbcS to develop green leaves as a factory to utilize RuBisCO as the CO₂ fixation enzyme. Perhaps detailed analyses of the evolutionary process of the gene of this enzyme from the ancient origins of RuBisCO to the present form may provide hints toward improving current photosynthesis. I look forward to these discoveries continuing in the future.

Funding

This work was supported by Japan Science and Technology Agency [grant number CREST]

Disclosure statement

No potential conflict of interest was reported by the author.

Acknowledgements

I would like to dedicate this review to late professors emeritus Shozaburo Kitaoka, Akira Wadano, and Yasushi Kai, who were my distinguished co-researchers during the past thirty years.

Notes

This review was written in response to the author’s receipt of the JSBBA Senior Scientist Award in 2014.

Abbreviations: Bs-RLP, Bacillus subtilis RuBisCO-like protein; CCM, CO₂-concentrating mechanism; EDTA, ethylenediaminetetraacetate; FA, formaldehyde; FBP, fructose 1,6-bisphosphate; FBPase, fructose 1,6-bisphosphatase; FBP/SBPase, fructose 1,6-/sedoheptulose 1,7-bisphosphatase; MEA, monoethanolamine; 6PG, 6-phosphogluconate; PRK, phosphoribulokinase; RHP, reductive hexulose-phosphate, RLP, RuBisCO-like protein; RuBP, ribulose 1,5-bisphospahte; S_rel, relative specificity; SBPase, sedoheptulose 1,7-bisphosphatase.