Evolutionary implications of C₂ photosynthesis: how complex biochemical trade-offs may limit C₄ evolution

Abstract

The C₂ carbon-concentrating mechanism increases net CO₂ assimilation by shuttling photorespiratory CO₂ in the form of glycine from mesophyll to bundle sheath cells, where CO₂ concentrates and can be re-assimilated. This glycine shuttle also releases NH₃ and serine into the bundle sheath, and modelling studies suggest that this influx of NH₃ may cause a nitrogen imbalance between the two cell types that selects for the C₄ carbon-concentrating mechanism. Here we provide an alternative hypothesis outlining mechanisms by which bundle sheath NH₃ and serine play vital roles to not only influence the status of C₂ plants along the C₃ to C₄ evolutionary trajectory, but to also convey stress tolerance to these unique plants. Our hypothesis explains how an optimized bundle sheath nitrogen hub interacts with sulfur and carbon metabolism to mitigate the effects of high photorespiratory conditions. While C₂ photosynthesis is typically cited for its intermediary role in C₄ photosynthesis evolution, our alternative hypothesis provides a mechanism to explain why some C₂ lineages have not made this transition. We propose that stress resilience, coupled with open flux tricarboxylic acid and photorespiration pathways, conveys an advantage to C₂ plants in fluctuating environments.

An alternative theory for bundle sheath ammonia tolerance in C₂species suggests that C₂to C₄evolutionary transitions may depend on bundle sheath nitrogen hub optimization.

Introduction

C₂ photosynthesis is a carbon-concentrating mechanism (CCM) that recovers and reassimilates CO₂ released from photorespiration to attenuate C losses, such as those seen in C₃ species under similar environmental conditions (Schlüter and Weber, 2016). Unlike the C₄ photosynthesis CCM, that dramatically reduces rates of photorespiration, the C₂ photosynthetic pathway relies on it and, in doing so, achieves a potentially ideal relationship with photorespiration. It benefits from the possible positive aspects of photorespiration (e.g. Timm and Bauwe, 2013; Busch et al., 2018; Eisenhut et al., 2019), while minimizing its negative impacts on net C assimilation efficiency under warm and dry conditions (Sage et al., 2013). The success of C₂ photosynthesis hinges upon the confinement of photorespiratory glycine decarboxylation to an inner leaf compartment, the bundle sheath (BS) (Fig. 1). C₂ plants assimilate CO₂ via the Calvin–Benson–Bassham (CBB) cycle and initiate the photorespiratory cycle in the mesophyll. However, because C₂ plants uniquely confine functional glycine decarboxylase complex (GDC) enzymes to the BS, photorespiratory glycine that is produced in the mesophyll must diffuse into the BS to be decarboxylated, which releases and concentrates CO₂ for effective reassimilation into a BS CBB cycle, liberates ammonia (NH₃) and synthesizes serine in this inner compartment (Fig. 1 and discussed in detail below; Rawsthorne et al., 1988). We propose that this phenotype conveys four key benefits in C₂ compared with C₃ plants, namely an expanded ecological niche into warmer and drier climates, enhanced net C assimilation due to better C reclamation and concentration, improved stress tolerance, and reduced nutrient losses that result from the C dilution effect experienced under elevating atmospheric CO₂ concentrations. Despite these potential benefits, relatively little is known about this rare C₂ pathway compared with other types of photosynthetic metabolism. To shed light on these potential benefits conveyed by C₂ physiology, we discuss what is currently known and, where unknown, propose hypotheses that warrant further investigation.

Fig. 1.

Carbon and nitrogen pathways in C₃, C₂, and C₄ species. (A; C₃) and (B; C₂) show photorespiration as the principal source of serine in C₃ and C₂ species, respectively. Mesophyll cells will carry out the bulk of central metabolism in C₃ species (A), with minimal input from bundle sheath cells as a result of their fewer organelles (C₁ metabolism not shown). (C) shows an absence of photorespiration and depicts the non-phosphorylated glycerate pathway (Gly. Pat.) in the cytosol as a viable serine source for C₄ species, although the phosphorylated pathway in the chloroplast may also play a role (not shown) (C is adapted from Igamberdiev and Kleczkowski, 2018; Jobe et al., 2019). Integration with sulfur metabolism (S Metab.) is also shown. Orange blocks represent fundamental enzymes. Blue blocks in between mesophyll and bundle sheath cells represent plasmodesmata. Calvin–Benson–Bassham cycle (CBBC), tricarboxcylic acid cycle (or pathway) (TCA); 2-phosphoglycolate (2-PG); 2-oxoglutarate (2-OG); oxaloacetate (OAA); phosphoenolpyruvate (PEP). Enzymes: glycine decarboxylase complex (GDC), serine hydroxymethyltransferase (SHMT), glutamine synthetase/glutamine:2-oxoglutarate aminotransferase (GS/GOGAT), phosphoenolpyruvate carboxylase (PEPC), malate dehydrogenase (MDH), NADP-malic enzyme (NADP-ME), pyruvate-orthophosphate dikinase (PPDK), glutamine synthetase 1 (GS1), amd nitrate reductase (NR).

Diversity and distribution of C₂ photosynthesis

C₂ photosynthesis has been characterized in >50 species of herbaceous and semi-woody shrubs (but not trees), originating from 20 plant lineages (4 monocot and 16 eudicot) representing 11 plant families (Sage et al., 2018; Lundgren, 2020). Fifteen of these 20 C₂ lineages also contain species using C₄ photosynthesis, while five C₂ lineages lack close C₄ relatives.

With far less than 1% of plant species currently identified as using C₂ photosynthesis, it is comparatively much rarer than C₃ and C₄ types that are used by >95% and ~3% of global plant species, respectively. Despite so few species using this rare physiology, C₂ plants are remarkably geographically and ecologically widespread, inhabiting diverse environments across every continent except Antarctica (Lundgren and Christin, 2017). The biogeography of C₂ species reveals an overall preference for regions with high light, heat, and possible drought (i.e. environments that typically induce high rates of photorespiration). However, C₂ species will also successfully establish in particularly wet, cold, and shady environments (e.g. Powel, 1978; Christin et al., 2011; Lundgren et al., 2015; Khoshravesh et al., 2016; Lundgren and Christin, 2017). Interestingly, C₂ species from lineages that lack C₄ relatives are among the most geographically widespread and are more likely to inhabit cooler, wetter environments with richer soils than C₂ lineages with C₄ relatives (Lundgren and Christin, 2017). Overall, the broad ecological distribution of C₂ plants points to a suite of adaptations particular to the C₂ phenotype, which might differ between lineages that did or did not also evolve C₄ photosynthesis.

Current hypotheses suggest that the complex trade-offs between C and N in the C₂ phenotype convey strong selection pressure for a C₄ photosynthetic pathway to evolve (Monson and Rawsthorne, 2000; Mallmann et al., 2014; Bräutigam and Gowik, 2016; Schlüter and Weber, 2016). Indeed, the C₂ phenotype is often associated with its role in facilitating the evolution of the complex C₄ photosynthetic state, having been repeatedly identified in C₃–C₄ intermediate species (i.e. evolutionary intermediate species with phenotypes that fall between that of typical C₃ and C₄ plants) (Schlüter et al., 2017; Sage, 2021). This is not the case for all C₂ lineages, however, as C₂ photosynthesis has been identified as existing for very long periods of time (e.g. >10 million years; Christin et al., 2011) in those lineages where C₄ photosynthesis never emerged (Lundgren and Christin, 2017), suggesting that it can be a stable evolutionary state and not always an inevitable path to C₄ photosynthesis (Edwards, 2019; Lundgren, 2020, Lyu et al., 2022). This notion of a stable-state C₂ phenotype is supported by constraint-based modelling, which identifies the C₂ phenotype as an optimal metabolic solution under a defined set of resource limitations (Blätke and Bräutigam, 2019). However, it remains unclear why some C₂ lineages transition to adopt C₄ physiology while others do not, and to what extent their broad ecological tolerance plays a role.

C₂ plants have an inherent physiological plasticity that may allow them to inhabit such large ecological ranges. Because the glycine shuttle only activates under environments that promote photorespiration, C₂ plants will be physiologically similar to C₃ plants under non-photorespiratory conditions (e.g. cool, wet environments): that is, they will both engage the CBB cycle in the mesophyll cells. However, under warm, arid, and high light conditions, CO₂ concentrations within the leaf wane, which increases the oxygenation to carboxylation ratio of Rubisco, and consequently rates of photorespiration and the glycine shuttle in C₂ plants (Schulze et al., 2016). Therefore, any negative effects of photorespiration on plant growth and yield that result from C losses should be less pronounced in C₂ compared with C₃ plants. Thus, in theory, C₂ species could thrive largely on a mesophyll CBB across a range of environments, supported by a C₂ pathway only when challenging conditions arise. Despite these theoretical advantages of physiological plasticity and optimized utilization of photorespiration, the relative rarity of C₂ species suggests that there might be as yet unidentified significant costs to this phenotype or a gross underestimation of the ubiquity of C₂ physiology. Indeed, if C₂ plants do encounter higher photorespiratory flux compared with C₃ plants—which remains unconfirmed in the literature—then they may be more common in hot, dry environments than previously thought. More work is needed to identify whether the C₂ phenotype exists across a broader range of species than is currently known.

Carbon assimilation

In both C₃ and C₂ plants, C is assimilated via the CBB cycle in mesophyll cells when Rubisco catalyses the addition of CO₂ to ribulose 1,5-bisphosphate (RuBP) (Fig. 1A, B). Rubisco will also catalyse a reaction of O₂ with RuBP, which creates toxic by-products that need to be detoxified through photorespiration (Fig. 1A, B). In C₃ plants, the entire photorespiratory pathway occurs within a single cell type, usually the mesophyll. The C₂ CCM, however, functions by spreading photorespiration across mesophyll and BS cells. Specifically, the GDC enzyme is exclusively localized to the BS of C₂ plants, such that any glycine that is produced during photorespiration builds up in the mesophyll, creating a concentration gradient from which glycine diffuses into the BS (Monson and Rawsthorne, 2000). Once in the BS, GDC activity—coupled with serine hydroxymethyltransferase (SHMT) using tetrahydrofolate/5,10-methylene tetrahydrofolate as a cofactor—catalyses the release of CO₂, along with NH₃ and serine (Fig. 1B) (Sage et al., 2012). This effective CCM approximately triples intraplastidial CO₂ concentrations compared to closely related C₃ species (Keerberg et al., 2014). Indeed, by running a mesophyll C₃ photosynthetic cycle in parallel with the dual-cell glycine shuttle, C₂ plants can recycle CO₂ lost from photorespiration more effectively than C₃ plants (Sage et al., 2012). Some C₃ species, such as rice, have developed effective single-celled strategies to reclaim photorespiratory CO₂ release in climates where rates of photorespiration are high (Sage and Sage, 2009). However, despite minimizing CO₂ losses, these C₃ species fail to achieve both the elevated photosynthetic rate and low CO₂ compensation points that C₂ plants do, which tend to be intermediate between those measured in C₃ and C₄ species (Vogan et al., 2007).

Successful establishment of the C₂ CCM relies on adaptations to leaf anatomy that underpin the biochemical changes essential for metabolic remodelling. This is evident in the higher leaf vein densities that have evolved in some lineages as a prerequisite for C₂ pump establishment, which not only enables efficient metabolite exchange between mesophyll and BS cells, but also maintains hydraulic flow under elevated evapotranspiration (Sage et al., 2011). Efficient decarboxylation of glycine in the C₂ BS also requires both an increase and realignment of BS chloroplast, mitochondria, and peroxisome organelles compared with those in mesophyll cells to accommodate the enhanced influx (Khoshravesh et al., 2016). The proliferation of BS mitochondria specifically not only facilitates CO₂ release from glycine decarboxylation, but may also improve biosynthetic capacity, for example through a non-cyclic, dual branched tricarboxylic acid (TCA) pathway, as discussed in detail below (Fig. 2B) (Tcherkez et al., 2009; Sweetlove et al., 2010). It is worth noting that because C₃ plants have functional GDC in both mesophyll and BS cells, they could, in theory, also translocate glycine between the two cell types. This seems unlikely, however, as it would be more parsimonious for the glycine to pass through the abundant GDC/SHMT in the mesophyll, rather than unnecessarily translocating to the BS to be decarboxylated. Furthermore, mesophyll glycine levels would likely not concentrate high enough to permit diffusion into the BS. There are also many fewer organelles within C₃ BS cells, which would limit decarboxylation capacity within this cell type should any glyine reach the BS.

Fig. 2.

The tricarboxylic acid (TCA) flux modes. The cyclic mode (A) and the non-cyclic dual branched pathway (B). Our C₂ mechanism hypothesis suggests that the traditional cyclic mode may operate in the dark similarly to C₃ species (some light operation will still occur), while a non-cyclic pathway may be operational in the light under photorespiratory conditions within C₂ species (malate and citrate valves not shown).

While the literature often highlights the negative impacts of photorespiration (e.g. Walker et al., 2016; Fernie and Bauwe, 2020), it is also an essential and beneficial metabolic pathway (Timm and Bauwe, 2013; Busch et al., 2018; Eisenhut et al., 2019). It has even been suggested that photorespiration may not convey CO₂ losses but instead show regulatory atmospheric advantages (Tolbert, 1994). Of course, any effects produced by photorespiration will be flux dependent. For example, at a 25–30% O₂ fixation rate, flux through the C₂ cycle can be as high as one-third the flux through Rubisco (Sharkey, 1988). As a consequence, the flux into NH₃ and through C₁ metabolism and serine can also be up to one-third the flux through Rubisco, making these among the highest metabolic fluxes in the plant, depending on species type and environmental conditions (Busch, 2020). Therefore, plants using C₃ or C₂ photosynthesis, but not C₄ photosynthesis, will have appreciable flux through photorespiration and thus into NH₃ and serine and through C₁ metabolism (Mouillon et al., 1999; Bauwe et al., 2010; Schulze et al., 2016; Jardine et al., 2017).

In summary, the C₂ glycine shuttle transports CO₂, NH₃, and serine into the BS, and necessitates high flux through C₁ metabolism (Fig. 1B). In any angiosperm plant, the BS is the gateway between the leaf and the remainder of the plant. Through the BS, leaves export sucrose to move assimilated C, the amino acids glutamate, glutamine, aspartate, asparagine, and notably serine to distribute assimilated N (Fig. 1B). The C:N ratio in the foliar compartment can be broadly approximated to 10:1 (Hager et al., 2016). If photorespiration runs at 30% of C fixation, each C fixed results in 0.15 ammonia produced which are detoxified using glutamate and glutamine. Since serine and glycine are directly produced in photorespiration and glutamate and glutamine are produced during ammonia refixation, all N that is exported from the leaf to support the remainder of the plant may be supplied via the C₂ glycine shuttle (e.g. as per data reported in Wilkinson and Douglas, 2003). Therefore, one needs to postulate that the stream of metabolites out of GDC/SHMT is split into those returning to the mesophyll, entering the veins to be distributed throughout the plant, and serving as substrates in the BS itself. To understand the evolutionary implications of the C₂ pathway, one consequently needs to consider the possible advantages afforded to C₂ plants beyond mere increased efficiency in C assimilation.

Nitrogen assimilation

It is often assumed that the NH₃ released into the BS from the C₂ glycine shuttle would return to the mesophyll to complete the photorespiratory cycle (Rawsthorne et al., 1988; Mallmann et al., 2014; Bräutigam and Gowik, 2016), similar to the way that amino acids (e.g. alanine) or organic acids (e.g. pyruvate) are used to balance N between the mesophyll and BS cells of C₄ plants. However, a complex coordination of photosynthesis, respiration, and photorespiratory N assimilation, similar to that occurring within the mesophyll cells of C₃ plants (Stitt and Krapp, 1999), could also be present in the BS of stable-state C₂ species, foregoing parts of the N shuttle previously described in C₂ species from lineages with C₄ relatives. (Rawsthorne et al., 1988; Monson and Rawsthorne, 2000). Leaf C and N use efficiencies in C₂ leaves lie largely in between those of C₃ and C₄ species (Vogan and Sage, 2011), which suggests that the C to N balance must provide an advantage to C₂ individuals over C₃ plants under certain environmental conditions, and that NH₃ cycling has concurrently been managed in accordance with C gain. Notably, the further down the evolutionary trajectory towards C₄ metabolism a particular C₂ phenotype is, the greater its C and N use efficiencies are realized (Vogan and Sage, 2011; Sage et al., 2018). Indeed, Sage (2021) proposes that the main evolutionary driver for C₄ photosynthesis may be NH₃ recovery and reassimilation, rather than C acquisition, as has been the dominant thinking in the field for decades. This scenario could present itself when C₃ plants are subjected to frequent periods of high photorespiratory flux, which require adjustment to their N stoichiometry. Consequently, this restructuring of N metabolism could allow the co-evolution of the stable state C₂ species alongside C₄ phenotypes (Lyu et al., 2022), resulting in a ‘super C2’ phenotype in these few lineages that could optimise both their resource use and environmental range.

When CO₂ concentrations increase in the BS, C₂ plants must also adjust their N stoichiometry; one theory on how this could be achieved is by using either side of a dual branched TCA pathway (Fig. 2B). This accommodates for increases to either C or N, by making amino acids if N is higher or organic acids if C is too high. These efforts could improve the N use efficiency (NUE) of C₂ plants through acclimatory transitions that redistribute N allocation throughout the plant, as is found in C₃ plants (Stitt and Krapp, 1999).

The need for photorespiratory glycine to be returned to the C₂ mesophyll should not be ruled out, nor should nitrate sequestration within the vacuole. Yet, evidence from the C₂ salad crop wild rocket/arugula (Diplotaxis tenuifolia), which must undergo rigorous leaf nitrate assessment before sale, suggests that it may be the C₂ mechanism causing the crop to hyperaccumulate nitrate within the leaf, which frequently exceeds safe levels for human consumption (Weightman et al., 2012; Santamaria et al., 2002). This characteristic would suggest a prominent role for nitrate within C₂ physiology, that cannot be said for C₃ salad crops, whose leaf nitrate levels lie well beneath those of wild rocket in the same environment (Signore et al., 2020). Given that wild rocket lacks C₄ relatives, this species is a likely candidate for the ‘super C2’ metabolism proposed above. If nitrate reduction enzymes are restricted to the mesophyll of C₂ plants, as they are in C₄ plants (Jobe et al., 2019), this would allow for a novel N feed into glycine (Fig. 3) and justify the abnormally high nitrate levels common to wild rocket (Iammarino et al., 2022). However, this would be in complete contrast to the high NUE efficiency that is associated with many C₄ species. Alternatively, as previously mentioned, tight coordination between other metabolic pathways with photorespiration functioning in an open mode rather than a closed cycle could maintain glycine pools necessary for cellular processes within the mesophyll cell. It is also possible that NH₃ (or NH₄⁺, depending on pH levels) may simply diffuse from the C₂ BS directly into the mesophyll, removing the need for complicated stoichiometry.

Fig. 3.

Hypothetical glutamate feed depicting the leaf N hub in stable state C₂ species. (A) Glutamine synthetase (GS1) and glutamate dehydrogenase (GDH) isoenzymes are located in both the cytosol and companion cells. These mechanisms could work to detoxify NH₃ produced following glycine decarboxylation, with possible signals from H₂O₂ and a potential increase of vein density compared with C₃ relatives facilitating rapid removal of NH₃ and distribution of nitrogen into non-toxic amino acids. This scenario may work independently or co-dependently with amino acid synthesis (as shown in B), in addition to 2-oxoglutarate (2-OG) and γ-aminobutyric acid (GABA) feeds into the TCA pathway (glutamate–2-OG interconversion may operate in cyclic cooperation with GS/GOGAT according to Stitt et al. (2002). Rapid nitrogen transport through the vein to sink tissue may be via glutamate, ornithine, arginine, and citrulline, by cytosolic relocation of metabolites and enzymes.

Similarly to C₃ species, the NH₃ generated by the glycine decarboxylation reaction in C₂ plants must be rapidly assimilated to prevent accumulation and consequent cell damage. Cellular impairment can result from a rise in NH₃ levels, which causes pH levels to become unmanageable (Bittsánszky et al., 2015). More importantly, NH₃ is also a known uncoupler of thylakoid electron transport, which directly restricts photosynthetic capacity (Stitt et al., 2002). With shifts in pH, the NH₃ released from GDC will convert to NH₄⁺, the accumulation of which is also highly toxic and requires amelioration (Bittsánszky et al., 2015). Because photorespiratory flux should be similar in C₂ and C₃ species, running at 30% of Rubisco reactions (Sharkey, 1988), C₂ phenotypes must also have efficient NH₃ and NH₄⁺ remediation mechanisms to mitigate toxicity. The plastidal GS/GOGAT (glutamine synthetase and glutamine:2-oxoglutarate aminotransferase) system, with potential contributions from C₄-related metabolites, is one possible route for rapid re-assimilation (Rawsthorne et al., 1988; Mallmann et al., 2014; Bräutigam and Gowik, 2016) (Fig. 1C). The spatial segregation inherent to the C₂ phenotype may provide additional buffering. By isolating GDC in the BS, NH₃ production is separated from photosynthesis occurring within the mesophyll cells. This separation may provide a benefit for both cell types, whereby the mesophyll must no longer contend with photorespired NH₃, while the BS Rubisco gets CO₂ enrichment with an unknown existing mechanism ameliorating NH₃ prior to accumulation. In theory, the C₂ CCM must simultaneously act as an N concentration mechanism, that may rely on a soil nitrate feed into the mesophyll under photorespiratory conditions (Fig. 6) or alternatively relinquish previously stored nitrate from the vacuole.

Fig. 6.

Proposed integration of carbon, nitrogen, and sulfur metabolisms in C₂ species. A hypothetical C₂ mechanism showing integration of carbon, nitrogen, and sulfur metabolic pathways in the enlarged bundle sheath, including alternative pathways for photorespired NH₄. Yellow circles denote glutamine synthetase isoforms: GS2 found in the chloroplast and the cytosolic GS1.

An alternative pathway to alleviate pressure on plastidal GS/GOGAT assimilation is to employ a secondary cycle in the cytosol to fix NH₃ to supply glutamine synthesis, using an isoenzyme of glutamine synthetase (GS1) (Fig. 3A). It has been found that the cytosolic GS1 enzyme is up-regulated in response to sudden increases in NH₃ influx to prevent toxicity (Guan et al., 2016). Pérez-Delgado et al. (2015) also identified an increase in gene expression of cytosolic GS1 in response to photorespiratory NH₃, which was found to be concomitant with glutamate dehydrogenase (GDH) expression for glutamate–2-oxoglutarate (2-OG) interchange in the absence of the plastidal GS2, similar to findings by Ferreira et al. (2019). Additionally, asparagine synthetase gene expression was also elevated under photorespiratory conditions, suggesting that asparagine serves as another player in N mobilization under stress (Pérez-Delgado et al., 2015). It is likely, however, that there are several GS isoenzymes that function to maintain cellular GS activity, which are regulated in response to levels of cellular N and nutritional type, especially within the vasculature (Bernard and Habash, 2009).

Under photorespiratory conditions, the influx of NH₃ could be fundamental for C₂ plants, as the NH₃ molecule influences the regulation and distribution of multiple GS1 isoenzymes in response to plant N status and environmental stimuli in some C₃ species, thus providing a protective role through N remobilization (Bernard and Habash, 2009). Indeed, Hirel et al. (1983) found significant GS1 activity in both C₂ and C₄ species of Panicum that was notably absent in the C₃ phenotype and directly correlated GS1 content with photosynthetic type. Frequent periods of high photorespiratory flux could induce GS1 expression in mesophyll and BS cell types of stable-state C₂ species by mitigating NH₃ accumulation via the sharing of N assimilation between both cytosolic and companion cell GS1 and the chloroplastic GS2. Collectively increasing BS N metabolism activity in close proximity to the vasculature (and potentially also coupled with an increase in vein density) means that the C₂ BS will supply a substantial amount of amino acids to the phloem stream (Weibull et al., 1990; Sandstrom and Pettersson, 1994; Leegood, 2008). This export gateway of the leaf acts as a potential compensatory C₂ feature to support plastidal GS2 activity under photorespiration. The resulting excess glutamine (from GDH amination) can then act as an export substrate as well as the main substrate for ornithine, arginine, and polyamines (Majumdar et al., 2016) once homeostasis of the cellular glutamine pool is achieved (Fig. 3B). A good proportion of the glutamine could then be exported to improve N mobilization for developmental processes. This export to the phloem opens any auxiliary cycle that may have operated (e.g. any smaller metabolic pathway in which it acts as an amino group donor for growth) and, consequently, necessitates de novo 2-OG synthesis. Any de novo synthesis of 2-OG from citrate in turn releases additional CO₂ into the BS. A proportion of the aspartate produced should also be exported to the phloem and removed from any future cycles, necessitating de novo C backbone synthesis. These export pathways to the phloem underscore the critical role of the mitochondria to supply C backbones for amino acid synthesis in both C₂ and C₃ plants.

In all plants, glutamate and glutamine are also precursors to proline, ornithine, and arginine synthesis, and therefore also to polyamine biosynthesis (Fig. 3B) (Winter et al., 2015). Ornithine is a light-stable amino acid that often interchanges with citrulline. The role of citrulline has not been previously linked to C₂ species but could be a strategic molecule for the phenotype, as citrulline and ornithine have been shown to accumulate under photorespiration and low concentrations of CO₂ (Blume et al., 2019). Research in C₃ species has shown that citrulline is a major long-distance N transporter involved in nutrient release to maintain osmotic pressure gradients, moving from source to sink tissue under stress (Kawasaki et al., 2000; Song et al., 2020). It is also a prominent reactive oxygen species (ROS) scavenger, thus preventing oxidative disruption to the electron transport chain through inhibition of hydroxyl radicals (Akashi et al., 2001; Yokota et al., 2002). Synthesis of citrulline through ornithine–arginine interconversion is mediated in the chloroplast in response to plant N status, and prompts crosstalk between the chloroplast, mitochondria, and the cytosol to maintain N homeostasis (Urbano-Gámez et al., 2020; Chen et al., 2022). These reactions would take place in the chloroplast-rich mesophyll of C₃ plants, causing the amino acids to diffuse through the BS to reach the phloem stream for export (Leegood, 2008). In contrast, if the same principle is applied to C₂ plants, then synthesis of N-rich amino acids will also be common in their chloroplast-rich BS cells (Khoshravesh et al., 2016), which would conveniently expedite N transport (Leegood, 2008). Maintaining functional chloroplast activity in both mesophyll and BS cells should theoretically permit flexibility of ornithine or arginine synthesis between the two cell types in C₂ species. This versatility may decline as mesophyll chloroplast concentrations wane with the emergence of a C₄ phenotype in some lineages (e.g. Stata et al., 2016). The role of arginine may be particularly fundamental to N distribution as it has a high N:C ratio, which is useful for N storage and transport under abiotic stress (Winter et al., 2015; Blume et al., 2019). Ornithine, arginine, and citrulline levels are thought to increase under high photorespiratory flux to reduce NH₃ toxicity, promote photoprotection, provide drought mitigation, and improve NUE (Joshi and Fernie, 2017). This suggests that C₂ phenotypes may strategically balance their C:N ratio through these three amino acids, uniting a TCA-derived C skeleton with photorespired NH₃ for glutamate synthesis prior to conversion.

The decarboxylation of ornithine or arginine by the enzymes ornithine decarboxylase and arginine decarboxylase, respectively, is the first catalytic step in the process of polyamine synthesis (Fig. 4). Polyamines such as putrescine, spermidine, and spermine are important in plants to stimulate DNA synthesis, growth, and physiological development (Chen et al., 2019). However, they are also known to play an important role as substantial N sinks and C:N regulators through hydrogen peroxide signalling (Moschou et al., 2012). Additionally, polyamines are thought to improve abiotic stress tolerance, as their levels increase in response to high salinity, high or low temperatures, and drought (Tiburcio et al., 2014; Liu et al., 2015) (Fig. 4). Two forms of polyamines are found in plants, free amines and amide conjugates, such as hydroxycinnamic acid amides (Tiburcio et al., 2014). Wild rocket (D. tenuifolia), a C₂ crop species, has been found to up-regulate the N-rich arginine metabolism under stress conditions, which very well could serve as an effective N distribution pathway whilst supporting a stress tolerance strategy (Cavaiuolo et al., 2017). Its roles in supplying export pathways and in being a precursor to stress-relevant pathways supports the idea that BS NH₃ functions as a leaf N hub from which the release of NH₃ into the BS has many beneficial outcomes. This view is in contrast to previous suggestions that NH₃/NH₄⁺ is a problem that must be solved by shuttling it back to the mesophyll (Mallmann et al., 2014).

Fig. 4.

Polyamine synthesis from glutamate showing relevant outcomes, noted at each stage (green). Input from cysteine is also shown. Amino acids are outlined in black and polyamines are shown in blue (tropane alkaloid synthesis may only be relevant to C₂ species in the Brassicaceae).

Plants have evolved various methods to detoxify both NH₃ and NH₄⁺, such as reformation into non-toxic derivatives or vacuole sequestration. NH₃ detoxification in C₂ plants may use GDH, either exclusively or co-dependently, alongside the ornithine pathway as previously described (Fig. 3B). These methods are also thought to be active under high levels of NH₄⁺ in C₃ plants (Bittsánszky et al., 2015). Therefore, under environmental conditions that promote a high photorespiratory flux and generate NH₄⁺, the C₂ phenotype may have evolved a highly efficient detoxification approach. Studies on C₃ plants have found increased expression of the GDH enzyme in both leaves and roots under high photorespiratory conditions, indicating an important role in NH₃ sensing and whole-plant tolerance (Turano et al., 1997; Cruz et al., 2006; Skopelitis et al., 2006; Sarasketa et al., 2014). If the activation energy requirements were low enough, this enzyme could be a prominent player in NH₃ detoxification by a cytosolic isoform interconverting 2-OG and glutamate. The process may also present in a cyclic form, interacting with GS/GOGAT, as suggested by Stitt et al. (2002) (Fig. 3A). This would regulate both glutamate pools and 2-OG TCA levels, maintaining C flow through the non-cyclic TCA chain, with 2-OG being the predominant C acceptor for NH₃ (Stitt et al., 2002). Glutamate would then present either as a precursor for other amino acids or for vascular export, maintaining a constant pool size. Interestingly, isoenzymes of both GS and GDH convey metabolic flexibility and spatial variability for adaptation to environmental and developmental changes under varying N nutrition to optimize growth (Fontaine et al., 2009). Evidence of GDH has been found not only in the cytosol in response to high NH₃ levels, induced by photorespiration, but also in the companion cells of the vasculature, especially in the midrib (Tercé-Laforgue et al., 2004). Therefore, the increase in vein density initiated at the very early stages of C₄ photosynthesis evolution (Christin et al., 2013; Griffiths et al., 2013) could primarily be for rapid NH₃ detoxification, with any hydraulic improvement being a secondary benefit (Sack and Holbrook, 2006).

One of the side effects associated with photorespiration is the stimulation of nitrate uptake (Bloom et al., 2010) (Fig. 3). As mentioned earlier, the C₂ salad crop D. tenuifolia often contains dangerously high leaf nitrate levels that must be closely monitored by growers prior to harvest (Weightman et al., 2012). The crop itself benefits from this accumulation because the influx of nitrate is a sink for excess energy dissipation under high light conditions, as the reduction of nitrate has a high energy requirement of eight electrons initially provided by NAD(P)H as reductant (Long et al., 2015). Storage of nitrate in the vacuole also helps to maintain the osmotic status of a leaf, which would be particularly important under hot and dry conditions where photorespiratory fluxes will be high (Bloom, 2015). This hyperaccumulation of nitrate causes an alkaline pH within the cell. To address this, phosphoenolpyruvate carboxylase (PEPC) and malate dehydrogenase (MDH) are stimulated to produce higher levels of malate to neutralize the pH (Stitt et al., 2002). This malate would then be transported into the roots for decarboxylation in C₃ species. However, in C₄ species, the malate would be decarboxylated in the BS to enhance the C concentration there. From a C₂ biochemistry perspective, this suggests that the original purpose of malate decarboxylation in a proto-C₄ BS may be to address cellular pH homeostasis following excessive nitrate uptake under photorespiration. Indeed, when photorespiration is absent, a plant would have to devote a quarter of photosynthate to nitrate uptake and reduction (Bloom, 2015).

Sulfur assimilation

Sulfur (S) is an essential nutrient for plant growth that is integrally entwined within the C and N metabolic framework. S assimilation is directly linked to photosynthetic capacity and consequently photorespiration, owing to the effect of S metabolism on redox regulation and electron consumption (Abadie and Tcherkez, 2019). Photorespiration plays a major role in supplying serine to the plant by stimulating S uptake (Abadie and Tcherkez, 2019). In turn, this influx of serine under conditions that promote photorespiration increases supply of C₁ units to provide the basic components needed for amino acids such as methionine and cysteine (Fig 5), which are essential for crops with high demand for these metabolites, such as those in the Brassicaceae for glucosinolate synthesis (Koprivova and Kopriva, 2016). In fact, there are three major routes for serine synthesis: through photorespiration, glycerate, and the phosphorylated pathway (Anoman et al., 2019), with photorespiration being the dominant serine source in both C₃ and C₂ leaves. Serine plays a fundamental role in cellular processes, as the cytosolic serine pool is the principal source of C₁ units for synthesis of methionine and S-adenosylmethionine (AdoMet). AdoMet is the predominant methyl donor for many methylation reactions such as lignin biosynthesis, DNA methylation, and other crucial processes (Fernie et al., 2013). Therefore, through serine, photorespiration supports at least four essential plant processes.

Serine biosynthesis will be different in C₄ compared with C₂ or C₃ plants. The reduced photorespiratory flux in C₄ leaves will require serine biosynthesis to be re-routed, either through its import via a different organ, the glycerate pathway, or via the phosphorylated pathway of serine synthesis (Ros et al., 2014; Zimmermann et al., 2021). Notably, serine synthesis has been found to be transferred to the root when photorespiration is reduced in C₄Flaveria (Gerlich et al., 2018). Alternatively, the glycerate pathway may provide serine for C₄ plants, which could be afforded by their increased C assimilation and crucial for their environmental stress tolerance (Igamberdiev and Kleczkowski, 2018).

The role of serine in the C₂ BS could be a particularly vital one, as serine acts as the precursor for cysteine synthesis, coupled with a reduced sulfide ion (Takahashi et al., 2011). The location of C₂ serine synthesis within the BS could improve the efficiency of the signalling mechanism required for cysteine application in the coordination of stress responses, as its close proximity to the vein would allow for rapid export. Cysteine acts as the S currency for all plant primary and secondary compounds, including DNA, amino acids, proteins, and defence compounds (Abadie and Tcherkez, 2019), and functions as the precursor for the photoprotective glutathione and for methionine for protein synthesis (Kopriva et al., 2019) (Fig. 5).

Fig. 5.

Proposed sulfate assimilation in C₂ species. Sulfate is taken up from the soil and delivered to the bundle sheath via the xylem. Glycine and γ-glutamylcysteine (γEC) form glutathione (GSH), a vital plant compound for abiotic stress resistance. Cysteine and methionine are exported for other important plant compounds. Cys, cystathionine; Hcyst, homocysteine; OAS-TL, O-acetylserine; APS, adenosine 5ʹ-phosphosulfate.

Glutathione is known to accumulate under intensive light levels to facilitate acclimation and mitigate photo-damage (Speiser et al., 2015). The primary role of glutathione is as a principal ROS scavenger, and it also plays a significant role in redox signalling and enzyme regulation (Kopriva et al., 2019; Li et al., 2020). To produce glutathione, three amino acids—cysteine, glutamate, and glycine—are amalgamated in a two-part reaction, catalysed by γ-glutamylcysteine synthetase and glutathione synthetase (Zechmann, 2014) (Fig. 5). Glutathione then pairs with ascorbate in an interactive pathway to mitigate ROS damage to organelles and photosynthetic machinery under stress conditions (Foyer and Noctor, 2011; Zechmann, 2014). Consequently, each cell is then dependent on glutathione for initiating defence gene activation through redox signalling. Synthesis of glutathione is under regulation of the hormone abscisic acid and is often induced by drought stress. Crucially, glutathione homeostasis is needed for compartment-specific plant defence sensing and signalling, especially within the chloroplast (Hasanuzzaman et al., 2017). Appropriate application of this process in C₂ species is found when the C₂ mechanism is activated under photorespiratory conditions, as the BS glycine pool can provide one of the substrates for glutathione by siphoning a fraction prior to serine conversion (Fig. 5). Interestingly, glutathione concentrations show a gradual increase along the C₃ to C₄ evolutionary continuum in the genus Flaveria (Kopriva, 2011).

Carbon, nitrogen, and sulfur integration

Photorespiration, and by extension the C₂ glycine shuttle, acts as an important metabolic junction with many of its intermediates linking to other pathways that are central to primary metabolism (Fig. 6). In this section, we discuss how the metabolisms of C, N, and S discussed above are integrated into a central hub through photorespiration.

In C₃ species, C metabolism is tightly linked to N pathways, with nitrate and NH₃ assimilation at the forefront. Changes in nitrate uptake and influxes of NH₃ affect the transcription of metabolic enzymes, such as PEPC, pyruvate kinase (PK), citrate synthase (CS), and isocitrate dehydrogenase (ICDH) (Scheible et al., 2000). Adjustments to these enzymes consequently allow for the cellular pH to be balanced in relation to nitrate uptake and C diversion, owing to the poor buffering capacity of the leaf (Stitt et al., 2002). Interactions between C and N can therefore be mediated appropriately by the TCA pathway in response to metabolite signalling to prioritize the cellular C:N balance. The uptake of nitrate can also have profound effects on cellular metabolism via the down-regulation of the AGPS gene, which encodes the starch synthesis regulatory subunit enzyme, ADP-glucose pyrophosphorylase (Scheible et al., 2000) and leads to an increase in PEPC activity and organic acids. Perhaps the localization of nitrate reduction to the mesophyll leads to an increase in PEPC activity that confines starch synthesis to the BS of C₄ species. This observation of C partitioning in C₄ species has been recently reviewed by Furbank and Kelly (2021). In C₂ species, however, the partitioning of photorespiration between the mesophyll and BS could potentially invoke consequences for the energy consumption needed for nitrate reduction, with NAD(P)H supplied by either the chloroplast or mitochondria depending on environmental conditions (Gardeström and Igamberdiev, 2016). One option would be that in high light environments, NAD(P)H would be amply supplied by mesophyll chloroplasts, despite nitrate reductase activity taking place. This relies on chloroplasts being interactive with other organelles, with cellular stoichiometry being flexible especially under stress or fluctuating conditions. Alternatively, TCA pathway activation in the light would also be another source of ATP under high stress conditions, alongside other ATP-generating processes within the mesophyll cellular metabolism, contributing to energy balancing (Noctor and Foyer, 1998).

Efficient functioning of diurnal respiration could equip C₂ species to optimize photorespired NH₃ by integrating glutamate with C skeletons of anaplerotic organic acids for amino acid formation, with an integral flexibility of a dual branched TCA pathway to adjust to cellular C:N status in response to NH₃ influx and other forms of stress (Fig. 6) (Vega-Mas et al., 2019; and see Ji et al., 2020 for a similar mechanism in response to heavy metal stress). Interestingly, verification of a diurnal TCA pathway was found in the C₂ species Salsola divaricata by Gandin et al. (2014) who successfully showed the C₂ mechanism and day respiration working in tandem to acclimate the plant to both high and low temperatures. There is a necessity for the TCA cycle to operate a diurnal open branched system to mitigate degradation of α-ketoglutarate dehydrogenase and succinyl-CoA under oxidative stress conditions (Sweetlove et al., 2002). This is supported by evidence that mitochondria can adapt their metabolism in accordance to variation in light intensity to coordinate responses with those seen in photorespiration (Huang et al., 2013; Reinholdt et al., 2019).

A C₂ solution could be an open flux TCA flow straddling the mitochondrial–cytosolic membrane, using cytosolic isoenzymes to offer flexibility for metabolite pools to compensate for redox regulatory reactions and address the cellular C:N balance (Fig. 6). Although this strategy is sometimes employed by certain C₃ plants, operating at a low to moderate flux, to improve C-use efficiency in specific environments (Tcherkez et al., 2017), C₂ phenotypes may depend on optimization of the diurnal respiratory pathway to increase C and balance the C:N ratio for acclimation as photorespiration occurs under challenging conditions. This may even be a characteristic of C₂ or C₄ species with the potential to adjust diurnal respiration at a higher flux in dynamic response to environmental conditions in comparison with C₃ species. This is supported by the idea of a diurnal TCA pathway being regulated by a CO₂/O₂ ratio (i.e. rates of photosynthesis and photorespiration in response to C and NH₃ flux), and a direct correlation between high photorespiration flux and an increase in day respiration (Tcherkez et al., 2008). If C₂ species partially relocate the TCA reactions to the cytosol, as some C₃ species do, then this could potentially increase the efficiency of signalling cascades for stress response mechanisms and source–sink transport given its close proximity to the vein in C₂ plants. This strategy may also improve metabolic coordination between photosynthesis, respiration, photorespiration, and N assimilation in response to hostile environments, a synchronicity that could be a requirement to enhance biochemical adjustments along the C₃ to C₄ evolutionary trajectory, with N acquisition at the forefront.

Within the mesophyll cells of C₃ species, the arrival of N that has been assimilated through glutamate is regulated by γ-aminobutyric acid (GABA), with subsequent Ca²⁺ and GABA signalling (and metabolism) to minimize disruption to the synthesis of amino acids, DNA, chlorophyll, and secondary metabolites under abiotic stress and utilize photorespired C in amino acids (Nunes-Nesi et al., 2010; Busch et al., 2018). GABA is synthesized in the cytosol from glutamate by glutamate decarboxylase. On its return to the mitochondria, GABA is transaminated by GABA transaminase to succinic semialdehyde. It is finally oxidized by succinate semialdehyde dehydrogenase to form succinate to rejoin the TCA pathway. This process incorporates glutamate into GABA, storing N before joining the TCA pathway, to recycle the C framework and help balance the C:N ratio (Bouche and Fromm, 2004) (Fig. 3B) (GABA can also be derived from polyamines, as shown in Fig. 4). Following the incorporation of succinate into the TCA pathway, it is then converted by succinate dehydrogenase to fumarate, and from fumarate to malate by fumarase. Indeed, a 3-fold elevation of fumarase was found to be present in the BS of the C₂ species Moricandia arvensis (Rawsthorne, 1988), which could indicate the presence of a GABA mechanism (i.e. the GABA shunt) under photorespiratory conditions that functions to continue TCA pathway operation and boost malate pools in response to stress and NADH ratios (Igamberdiev, 2020). The inclusion of GABA synthesis to boost fumarase activity could be an important mechanism, as a reduction of this enzyme leads to photosynthesis inhibition by defective stomatal movement (Nunes-Nesi et al., 2007).

In C₃ species, elevated GABA levels stimulate the cell’s stress response for photoprotection, osmoregulation, C:N balancing, regulation of stomatal opening, and gibberellic acid synthesis (Li et al., 2021). The molecule is known to be a thermo-protectant that can improve C assimilation under heat stress and maintain leaf water levels (Priya et al., 2019). It also has a role in root proliferation and primary root growth (Li et al., 2021), which would aid in water and nutrient acquisition. GABA can be synthesized by the citrate branch of the dual branched TCA pathway in the light (Fig. 6). The TCA pathway can provide either energy or backbones for amino acid synthesis, in response to C:N balance status, by switching flux modes between the dual branches of either the citrate or malate lines (Figs 2B, 6), with the conventional cyclic mode more likely to be used in darkness (Vega-Mas et al., 2019). Within C₃ plants, operation of the malate and citrate valves of the diurnal TCA pathway forms a dynamic coordination between chloroplast and mitochondria, which maintains energy balance through redox regulation, with chloroplastic malate acting as a redox signal of imbalance within the cell (Scheibe, 2019). Meanwhile, the mitochondrial citrate valve can also support organic acid synthesis and N assimilation in the cytosol in accordance with cellular demand, resulting in regulation of gene transcription (Igamberdiev, 2020). C₂ species may utilize this approach, where straddling both mitochondria and cytosol would allow for greater metabolic speed and efficiency within the BS (Fig. 6). Additionally, the synthesis of ATP produced in the mitochondria will also have a direct influence over subsequent transcription and translation of organelle genomes as a response to energy status needed for metabolic and assimilatory co-dependence (Liang et al., 2015).

While the coordination of regulatory mechanisms for C, N, and S assimilation remains undefined in the literature, its importance is clear, as each element is crucial for cysteine and glutathione synthesis to provide ROS protection through redox signalling and regulation in tandem with an NH₃–glutamate feed to prevent NH₃ accumulation. Indeed, both N and S uptake and assimilation are improved when NH₃ is the N source, implying a link between these metabolic pathways (Coleto et al., 2017), especially under hostile environments, thus implicating the involvement of photorespiration (Eisenhut et al., 2015).

C₂ biochemistry may provide enhanced protection to the plant. The high flux through photorespiration that C₂ plants experience will also create high fluxes in many metabolites required for secondary defence compounds, such as flavonols and glucosinolates, when mitigating for excessive heat, drought, and oxidative stress (Nakabayashi et al., 2014; Essoh et al., 2020). Indeed, these secondary defence compounds have been found in two C₂Brassicaceae species that lack C₄ relatives—D. tenuifolia and Moricandia arvensis (Martínez-Sánchez et al., 2007; Marrelli et al., 2018). Moreover, the rich Ca²⁺ content and high levels of antioxidants, polyphenols, glucosinolates, alkaloids, and amino acids found in D. tenuifolia leaves suggest that GABA synthesis could be in operation to protect the plant (Brock et al., 2006; Bell et al., 2015). Secondary metabolites rich in N and S are not confined to the Brassicaceae. High levels of polyphenols, glycosides, alkaloids, etc. have also been identified in the C₂ species Parthenium hysterophorus (Asteraceae) and Mollugo verticillata (Molluginaceae), both of which, interestingly, are highly invasive and also lack C₄ relatives (Lundgren, 2020; Kumar and Shariff, 2021; Jaiswal et al., 2022). Lineages that do synthesize these compounds would need to increase both C gain and TCA flux to provide a C framework over and above that needed for growth and reproduction. Within C₂ plants, de novo nitrate uptake must also increase to supplement photorespired NH₃ fixation in order to maintain plant growth while also synthesizing the N-containing secondary compounds. Of course, these interesting metabolites are also present in many C₃ species, providing nutrient-rich crops and plant stress resistance.

If C₂ species use their consistent photorespiratory flux to balance incoming C, N, and S with their secondary metabolite sink, on top of that which is already required for primary metabolism within their environment, then that would be sufficient to trap these lineages in a C₂ state. Progression towards C₄ would limit photorespiration and therefore decrease the availability of substrates for defence and mitigation pathways. A stable C:N balance within a certain criterion of environmental conditions could allow C₂ phenotypes to thrive by facilitating enhanced plasticity between its C₃ and C₂ cycles. Under conditions that promote a high photorespiratory flux, the C₂ mechanism could feed into an N sink for protective secondary metabolites, which could then be stored within the vacuole. Under non-photorespiratory conditions, the C₃ cycle would be sufficient to provide an adequate C:N ratio to fuel growth and reproduction within a C₂ plant. This would of course depend on the ability for C₂ cycle engagement to self-optimize under variable environmental conditions that could oscillate dramatically.

Photorespiration serves to dissipate energy (Eisenhut et al., 2019) and therefore provides protection against ROS. The much lower photorespiratory flux experienced by C₄ plants would consequently require an alternative metabolism to enable crosstalk between organelles to reach optimal protection and synthesize secondary compounds for plant protection.

Implications for stress tolerance in C₂ plants

Our proposed integration of C, N, and S metabolism in C₂ plants (Fig. 6) is underpinned by the diurnal, dual branched TCA pathway being able not only to function, but also to optimize and coordinate responses between photosynthesis, respiration, photorespiration, and ROS mitigation. We suggest that confinement within the BS cytosol and mitochondria is key to the success of C₂ species. Through this novel organization of primary metabolism, a C₂ crop should theoretically be able to maintain yields and nutritional integrity better than a C₃ crop as CO₂ levels rise under the advancement of climate change. An efficient, integrated optimization of metabolic pathways, supplemented by retaining C₃ gene expression in mesophyll cells, would let the CBB cycle, TCA pathway, N metabolism, and S metabolism operations be maintained under temperate conditions, giving C₂ plants the flexibility to withstand cooler conditions and low light levels better than C₄ plants (Bellasio and Farquhar, 2019). Indeed, C₂ plants occur in cooler regions despite an overall preference for warmer climates (Lundgren and Christin, 2017), displaying evidence of C₂ climatic versatility. C₂ plants may have alternative routes of photorespiratory metabolism, such as another open flux system that could replenish glycine and serine pools for auxiliary metabolic processes (Timm et al., 2012; Busch, 2020). There may also be multiple photorespiratory phenotypes within a species, as has been identified in the C₃ model species Arabidopsis thaliana (Timm et al., 2012). If also present in C₂ species, this intraspecific photorespiratory diversity would facilitate tolerance of unfavourable or fluctuating environments. Indeed, under specific combinations of light and temperature, C₂ species are predicted to simultaneously optimize CO₂ assimilation and resource use (Blätke and Bräutigam, 2019; Johnson et al., 2021). This efficiency therefore reduces evolutionary pressure and facilitates the stable state version of the C₂ phenotype.

Conclusion

For a C₂ lineage to transition towards C₄, evolution needs to find a solution for serine supply, for the sufficient production of export amino acids in the absence of high flux through NH₃, and potentially for substrate production to maintain—even at a lower level—synthesis of metabolites necessary for abiotic stress resistance. The primary advantage of the C₂ phenotype may therefore be efficient NH₃ detoxification in response to increased photorespiratory flux. With a rapid influx of N, the plant would require improvements in C backbone synthesis to provide a C framework for protective secondary metabolites in addition to amino acids. Therefore, the strong need for an effective N sink would take priority over C assimilation under photorespiratory conditions, for rapid N remobilization and stress protection under conditions that incite high photorespiratory flux. Indeed, plant lineages that operate pathways to successfully integrate N storage and transport (e.g. via increased investment in the shikimate or ornithine pathways) may be more likely to adopt the C₂ phenotype on the strength of their N sink. Based on the competence of the N sink, our theory may explain why some C₂ lineages do not transition to a C₄ state (i.e. the stronger the N sink, the more stable the C₂ state becomes). These C₂ populations could thrive under nutritionally stable soil types and may have co-evolved alongside C₄ species to optimize the C₂ pathway towards a ‘super C₂’ phenotype. For species that do not specialize in synthesizing N-rich compounds, their alternative may be to increase amino acid production and consequently require more C, resulting in the amino acid shuttling we associate with C₄ phenotypes. Interestingly, this theory gives secondary metabolites a more prominent role within plant biochemistry and evolution, in addition to their well-known role in stress response. Importantly, a high performing version of C₂ metabolism may provide a lifeline for our crop resources under climate change as a solution to compromised food security and quality.

Glossary

Abbreviations

BS

bundle sheath

CBB

Calvin–Benson–Bassham cycle

CCM

carbon-concentrating mechanism

GDC

glycine decarboxylase complex

GDH

glutamate dehydrogenase

GS/GOGAT

glutamine synthetase and glutamine:2-oxoglutarate aminotransferase

NUE

nitrogen use efficiency

PEPC

phosphoenolpyruvate carboxylase

PK

pyruvate kinase

ROS

reactive oxygen species

RuBP

ribulose 1,5-bisphosphate

SHMT

serine hydroxymethyl transferase

TCA

tricarboxylic acid

Conflict of interest

The authors have no conflicts of interest.

Funding

CAW is funded by Biotechnology and Biological Sciences Research Council (BBSRC) (BB/V509711/1) and Waitrose Agronomy Group as part of the Waitrose Collaborative Training Partnership. MRL has been funded by a Leverhulme Early Career Fellowship (ECF–2018–302) and a UK Research and Innovation Future Leaders Fellowship (MR/T043970/1).