You have full access to this article via Jozef Stefan Institute.

Photorespiration is a key determinant of plant response to changing climates, connecting atmospheric concentrations of CO and O to net carbon assimilation and central metabolism. Photorespiration stems from a secondary reaction of rubisco, the enzyme responsible for photoautotrophic CO fixation. Rubisco also catalyses a reaction with O, producing 2-phosphoglycolate (2-PG), an inhibitory byproduct that is recycled by photorespiration. Photorespiratory recycling of inhibitory 2-PG to Calvin-Benson cycle intermediates involves a series of energy-consuming reactions and CO loss across multiple organelles. Rates of rubisco oxygenation are sensitive to both the atmospheric CO concentrations and the leaf temperature, directly connecting photorespiration to key aspects of climate change. Photorespiration is a high-flux pathway, and its flux has large consequences for plant metabolism and plant productivity; for example, under current atmospheric CO, ~20% of photosynthetically fixed carbon is released by photorespiration, making it a limiting factor of crop yield. It also consumes 30-40% of leaf ATP and reductant from the light reactions and involves additional nitrogen cycling. However, photorespiration is also an essential biochemical process for normal growth and development of plants under oxygenated environments; for example, deletion of core photorespiratory enzymes inhibits photosynthesis. In addition, large fractions of the CO released by photorespiration is re-captured by the chloroplast, which still requires additional energy but partially mitigates the consequences of photorespiration from a carbon-balance perspective.

Because of its high flux, photorespiration has been co-opted into other aspects of plant metabolism including nitrogen assimilation, sulfur assimilation, amino acid synthesis, reactive oxygen species signalling and one-carbon (C) metabolism. While a general connection between photorespiration and these other pathways is clear, it is less clear how much flux from photorespiration goes towards these alternative pathways. For example, recent studies indicated that ~40-30% of photorespiratory carbon can be diverted from the pathway as serine in tobacco and sunflower leaves, but similar values are not resolved for C metabolism. These findings challenge the view that photorespiration is only important for returning carbon into the Calvin-Benson cycle, but the quantity of flux from photorespiration to other pathways such as C metabolism remain unclear. A quantification of these fluxes is also important for understanding the fundamental connections between metabolic pathways, benchmarking future engineering efforts and understanding to what degree metabolic flux will be shaped by changing climates.

Folates (tetrahydrofolate (THF) and its C derivatives) are essential compounds for plant metabolism and in the diets of humans and animals (as B vitamers). These folates are important for (1) reactions that require C units other than methyl groups (biosynthesis of purines and thymidylate, protein biosynthesis initiation in plastids and mitochondria, B biosynthesis), (2) reactions that require methylation via the universal methyl group donor S-adenosyl methionine (AdoMet, such as lignin, DNA and histone methylation) and (3) reactions that require AdoMet as non-C precursor (ethylene, vitamer B). The first committed step in pantothenate (vitamer B) biosynthesis requires a C unit and is thus needed for production of coenzyme A, potentially impacting lipid metabolism and oilseed crop productivity. C units in THF derivatives come primarily from serine through the activity of serine hydroxymethyl transferase (SHMT) in various cellular compartments. This serine may come from either photorespiratory-dependent or -independent pathways, including a recently described chloroplastic pathway in poplar. Photorespiratory-independent pathways include a phosphorylated pathway which produces serine from 3-phosphoglycerate from the Calvin-Benson cycle of photosynthesis or glycolysis and are important in tissues without photorespiration, with mutants in this pathway showing growth and metabolic phenotypes. While non-photorespiratory serine production is thus important, photorespiratory metabolism involves the flux of much larger pools of serine.

There is a well-established link between photorespiration and C metabolism in mitochondria through the glycine decarboxylase complex (GDC) and SHMT, coupled via a common pool of 5,10 methylene-THF. Mutants in these genes result in clear photorespiratory phenotypes, showing the importance of mitochondrial C metabolism to photorespiration. While this work firmly establishes the dependency of photorespiration on C metabolism, it is less clear how much carbon from photorespiratory intermediates supplies C metabolism for use outside of photorespiration. Several lines of evidence suggest that mitochondrial 5,10 methylene-THF is passed exclusively to serine synthesis; for example, the pool of mitochondrial 5,10 methylene-THF does not equilibrate with the overall pool of the cell, at least in heterotrophic tissues. Previous work also suggests that rather than being released from enzyme protein surfaces, 5,10 methylene-THF is channelled directly from the T protein of the GDC complex to SHMT. Taken together, these findings suggest that other mechanisms are needed to explain a biosynthetic connection between photorespiration and C metabolism in the cytosol. This is especially important given the greatest demand for C units is assumed to be in the cytosol, the exclusive site for synthesis of AdoMet.

If mitochondrial GDC and SHMT activity do not link photorespiration to C biosynthesis to processes outside the mitochondria, what intermediates of photorespiration could and how does that link respond to rates of rubisco oxygenation? For the C folate derivatives needed for other metabolism, photorespiration may still be a key supplier as photorespiratory serine, glycine and formate serve as potential sources of C units for metabolism, with serine hypothesized as the main C donor. A biosynthetic link between photorespiration and C metabolism has been proposed where cytosolic SHMT contributes C units for needs of cytosolic metabolism by catalysing the conversion of serine to glycine and methylene-THF; the former is metabolized in mitochondria by GDC/SHMT pathway, while the latter can be reduced to 5-methyl-tetrahydrofolate (5-methyl-THF) (Fig. 1), which serves as the methyl donor for the conversion of homocysteine to methionine. Evidence for activity of this route is clearly shown in C-glycine feeding studies under different CO and O conditions, which show C enrichment in methylene-THF and AdoMet. Interestingly, these C-glycine feeding experiments did not show an increase of this labelling into methylene-THF and AdoMet with increasing rates of rubisco oxygenation, but interpretation of this relationship is complicated by the large diluting pool of C-glycine produced as native photorespiration increased. This observation inspires determining whether C biosynthetic flux out of photorespiration scales with rates of rubisco oxygenation, or whether there is a 'baseline' rate of C production from photorespiration under ambient conditions. In addition, it is not clear whether exogenously supplied glycine increases cytosolic pools of glycine and artificially decreases rates of cytosolic SHMT due to a more favourable equilibrium against glycine production. Subsequent work with CO labelling demonstrates that carbon 5 on methionine that originates from THF increases in C enrichment as rates of rubisco oxygenation increase, further supporting this hypothesis, but the single time point measurement after 2 h of CO feeding does not allow a flux calculation of the amount of carbon that left photorespiration or what the contribution from other serine biosynthetic pathways could be. It also does not clarify whether that carbon came from serine or alternative non-serine routes connecting photorespiration to C metabolism.

An alternative potential pathway facilitating the biosynthetic flux of photorespiratory carbon towards C metabolism is the synthesis of formate through the non-enzymatic decarboxylation (NED) of glyoxylate (Fig. 1). It has been hypothesized that increased CO release under high photorespiratory conditions is due to NED of hydroxypyruvate or glyoxylate by HO in plant leaves lacking peroxisomal catalase, the enzyme that converts HO resulting from glycolate oxidation into O and water. This NED would produce formate which could then either be oxidized to CO in the mitochondria by formate dehydrogenase (FDH) or incorporated into cytosolic C metabolism through formate tetrahydrofolate ligase (FTHFL). This putative reaction would compete with at least one other NED between hydroxypyruvate and HO which produces glycolate, which would not feed into C metabolism. Indeed, the concentrations of glyoxylate and hydroxypyruvate relative to the rate constants of their reaction with HO provide evidence against the NED of glyoxylate to form formate, but a clearer demonstration of its absence would help resolve this question more fully. It should also be noted that the peroxisomal catalase mutants (cat) used in these studies also show increased oxidative damage and defence signalling responses.

The quantitative biosynthetic link between photorespiration and C metabolism is important for understanding basic plant metabolism as well as understanding how future climates will alter plant nutrition. For example, C metabolites and other vitamin B complexes decrease in response to rising atmospheric CO, potentially due in part to decreased photorespiration. Despite this compelling evidence, it is unclear whether carbon flow directed to C metabolism is affected by changes in atmospheric CO availability generally or the photorespiratory conditions specifically. To tease these two mechanisms apart, experimental treatments that more specifically impact photorespiration relative to carbon assimilation (like decreased O) are needed. Lowered O more effectively minimizes photorespiration with minimal changes to flux through other aspects of carbon assimilation relative to changing CO concentrations. In addition, it remains unclear what portion of the assimilated carbon leaves photorespiration to maintain cytosolic C metabolism, which pathway is predominant, and how much carbon each pathway contributes to C metabolism.

To address these questions, we used CO labelling in wild-type and cat knockout Arabidopsis thaliana leaves and isotopically non-stationary metabolic flux analysis (INST-MFA) in wild type under varying O concentrations. INST-MFA is a powerful approach that quantifies fluxes from CO labelling time courses and has been used to quantify fluxes in metabolism downstream of carbon capture through photorespiration, the tricarboxylic acid cycle, to amino acid synthesis and remobilization of cellular sugar. In this Article, we demonstrate that ~5.8% of assimilated carbon passes to C metabolism under ambient CO and O, but this flux decreases by ~5-fold when photorespiration is suppressed to minimal rates at 2% O where ~1.2% of fixed carbon passes to C metabolism. Furthermore, we demonstrate that the primary carbon flux from photorespiration to C metabolism is mainly through serine. Our study provides compelling evidence quantifying how photorespiration is integrated with C metabolism in autotrophic leaves, with possible implications for understanding the mechanism behind C metabolic response to elevated CO and benchmarking future engineering efforts.