Current crop yields will not be enough to sustain today's diets for a growing global population. As plant photosynthetic efficiency has not reached its theoretical maximum, optimizing photosynthesis is a promising strategy to enhance plant productivity. The low productivity of C3 plants is caused in part by the substantial energetic investments necessary to maintain a high flux through the photorespiratory pathway. Accordingly, lowering the energetic costs of photorespiration to enhance the productivity of C3 crops has been a goal of synthetic plant biology for decades. The use of synthetic bypasses to photorespiration in different plants showed an improvement of photosynthetic performance and growth under laboratory and field conditions, even though in silico predictions suggest that the tested synthetic pathways should confer a minimal or even negative energetic advantage over the wild type photorespiratory pathway. Current strategies increasingly utilize theoretical modeling and new molecular techniques to develop synthetic biochemical pathways that bypass photorespiration, representing a highly promising approach to enhance future plant productivity.
Most agronomically important crops use the C3 pathway of photosynthesis to fix atmospheric CO2 through ribulose-1,5-bisphosphate carboxylase oxygenase (rubisco); they are thus compromised by gross inefficiencies in CO2 fixation, which is transduced in high rates of photorespiration . The most productive crops instead use the C4 pathway of photosynthesis, which confers faster photosynthesis and results in better water and nitrogen use efficiencies. C4 photosynthesis acts as a biochemical pump to increase the intracellular CO2 concentration at the site of Rubisco, thereby lowering the flux through the oxygenation reaction. The reduction in photorespiration was likely the major driving force behind the repeated evolution of the C4 pathway of photosynthesis [1,2].
The accelerating growth of the world population, together with a reduction in cultivable soils and an increasing demand for biofuel production, make increased crop productivity a goal of global importance . Current crop yields, resulting from breeding strategies and the use of fertilizers and improved irrigation, show sparse increases  and will not be enough to sustain current diets for the growing global population in the next decades . In the search for alternative strategies to increase plant yield, it was found that plant photosynthetic efficiency is far from its theoretical maximum . Optimizing photosynthesis thus became a strategy with promising potential to enhance plant productivity . Different schemes to optimize photosynthesis were proposed and, at least partially, experimentally pursued. Important examples are the optimization of antenna size [8,9], the improvement of carbon fixation by reducing the oxygenase activity of Rubisco [10,11] or by the introduction of alternative carbon fixation pathways [12,13] or complete carbon-concentrating mechanisms into plant chloroplasts [14–16], the implementation of the C4 pathway in C3 plants [17,18], and the reduction in flux through the photorespiratory pathway by using energy-efficient synthetic biochemical pathways to bypass photorespiration [19–23]. Compared with the other approaches, the implementation of synthetic bypasses to photorespiration offers more immediate potential to improve crop photosynthesis in the near future. This is because their introduction can be achieved with modern technologies and because improvements of photosynthetic performance and plant growth were already reported with model plants under both greenhouse and field conditions [19,21–23].
Photorespiration evolved as a selective response to Rubisco's promiscuity
Rubisco is used by all oxygenic photosynthetic organisms as the primary carboxylase of the Calvin–Benson cycle, also known as the C3 reductive carbon fixation cycle. In this reaction, carboxylation of ribulose 1,5-bisphosphate by CO2 yields two molecules of 3-phosphoglycerate (3-PGA). Rubisco is a promiscuous enzyme that also catalyzes the oxygenation of ribulose 1,5-bisphosphate . The products of this side reaction are one molecule of 3-PGA and one molecule of 2-phosphoglycolate (2-PG). 2-PG is a dead-end metabolite of high toxicity, as it inhibits important chloroplastic enzymes such as triose phosphate isomerase and phosphofructokinase [25–28]. During early plant evolution (∼400 Ma), when the levels of CO2 were high (1500–3000 ppm; ) and the atmosphere was essentially free of O2, there was no selective pressure against Rubisco's potential promiscuity . After the onset of oxygenic photosynthesis, massive amounts of O2 were produced, increasing the oxygenase activity of Rubisco . The photorespiratory pathway, also known as the oxidative photosynthetic carbon cycle, evolved as a selective response to these changes to prevent an accumulation of toxic levels of 2-PG.
The photorespiratory pathway is a multiple-step repair system
The photorespiratory pathway is a multiple-step repair system that converts 2-PG back to 3-PGA in all organisms carrying out oxygenic photosynthesis, although the individual biochemical steps and subcellular localization differs between major photosynthetic lineages [31–33]. In higher plants, photorespiration involves 16 reactions and numerous transport processes distributed across chloroplast, peroxisome, mitochondrion, and the cytosol [31,33] (Figure 1).
The photorespiratory pathway.
In plant chloroplasts, 2-PG is dephosphorylated to glycolate and transported to the peroxisome, where it is oxidized to glyoxylate through glycolate oxidase. The resulting hydrogen peroxide (H2O2) is detoxified by catalase. Glyoxylate is further transaminated to glycine that is transported to the mitochondrion. In this organelle, two molecules of glycine are converted to one molecule of serine, CO2, and ammonia (). Serine is transported back to the peroxisome, where it reacts with glycine to produce hydroxypyruvate, which is further reduced to glycerate. Glycerate is transported to the chloroplasts, where it is phosphorylated to form 3-PGA, which can be fed back into the Calvin cycle.
Why bypass photorespiration?
The conversion of 2-PG to 3-PGA through the photorespiratory pathway recovers ∼75% of the pre-fixed CO2. The remaining CO2 is released in the mitochondria together with an equimolar amount of (Figure 1). Photorespiration thus lowers photosynthetic efficiency in that CO2 and must be re-assimilated with the concomitant consumption of both ATP and reducing power. Under current atmospheric conditions, the rate of CO2 release from glycine decarboxylation through the photorespiratory pathway can reach five times the rate of normal tricarboxylic acid cycle activity . In addition, the activity of the photorespiratory pathway is enhanced under high temperatures and water shortage [35–38]. As a consequence, the high flux through the photorespiratory pathway imposes large energetic costs on C3 plants. These costs are responsible for the lower productivity of C3 plants compared with C4 plants, explaining why lowering the energetic costs of photorespiration has been one of the main goals in the quest to enhance the productivity of C3 crops. One way to achieve this goal is to diminish the flux through the photorespiratory pathway by installing synthetic biochemical bypasses [7,39]. The reasoning behind this approach is to avoid mitochondrial decarboxylation and deamination, as well as to promote the release of CO2 inside the chloroplasts to favor the carboxylase activity of Rubisco.
Synthetic pathways to bypass photorespiration: from the laboratory to the field
It seems obvious that any reduction in photorespiration should increase net CO2 fixation and therefore plant yield. Nevertheless, disruption of photorespiration causes strongly retarded growth of plants. Therefore, the synthetic pathways to bypass photorespiration tested in pioneering studies [19,21] were thought to divert some of the 2-PG into these pathways while still retaining some activity of photorespiration. This was demonstrated by the lower glycine/serine ratios obtained in the transgenic plants relative to the wild type [19,21].
The first bypasses to photorespiration were introduced into chloroplasts of the model C3 plant Arabidopsis thaliana. The bypass tested by the Peterhansel group uses the complete glycolate catabolic pathway of bacteria to convert glycolate into glycerate  (Figure 2). In contrast, the bypass tested by the Maurino group converts glycolate into CO2 through a pathway using transgenic glycolate oxidase, malate synthase, and endogenous NADP-malic enzyme and pyruvate dehydrogenase  (Figure 2). Both approaches led to enhanced biomass production under controlled greenhouse conditions, but only under short days. In addition, transgenic Arabidopsis thaliana plants with either of the two pathways had flatter and thinner leaves, indicating anatomical adaptations in response to the imposed physiological changes.
Synthetic biochemical bypasses to photorespiration analyzed in plants.
In the pathway tested by the Maurino group, the introduction of catalase in the chloroplasts was necessary to eliminate deleterious effects of the accumulation of H2O2, a toxic product of the glycolate oxidase reaction that normally occurs in peroxisomes during photorespiration . Maier et al.  suggested that optimized versions of this basic approach should be tested in plants with agronomical importance after finding a suitable glycolate dehydrogenase that would not produce H2O2 to substitute glycolate oxidase and catalase. Furthermore, it is imperative to analyze plant performance under field conditions, where pot size is not constraining growth and plants are challenged by natural factors.
Inspired by this suggestion, the Ort group recently showed that installing the synthetic bypasses to photorespiration in chloroplasts of the model C3 crop Nicotiana tabacum resulted in enhanced plant growth in the field . This work tested the original two bypasses and an alternative version of the pathway tested by the Maurino group in which glycolate is oxidized by a green algal glycolate dehydrogenase  (Figure 2). In this case, catalase co-expression is not necessary and signaling or oxidative effects of remaining H2O2  are avoided. In addition, this optimized synthetic pathway was co-expressed with an RNAi module targeting the chloroplastic glycolate exporter  to guarantee that most glycolate remains in the chloroplasts, thereby diminishing the flux through photorespiration and enhancing glycolate metabolism through the synthetic pathway. Intriguingly, this last version of the synthetic bypasses showed the largest increases in biomass (up to 40% compared with wild type) in replicated field trials . The use of a glycolate oxidizing enzyme that does not produce H2O2 was of enormous importance for the improvement of the original pathway. We recently discovered a mitochondrial dehydrogenase that oxidizes photorespiratory glycolate in diatoms ; this enzyme represents an alternative new tool to be included in future synthetic photorespiratory bypasses, as it does not use H2O2.
The complete synthetic bacterial glycolate pathway and, in parallel, only the first enzyme of this pathway were also introduced in Camelia sativa, an oilseed crop  (Figure 2). Increased photosynthetic efficiency, vegetative biomass, and seed yield (up to 70%) were observed in both cases. Notably, in both Camelia sativa  and Arabidopsis thaliana, the introduction of only the first enzyme of the bacterial glycolate pathway showed similar effects as the introduction of the whole-pathway. The authors attributed this observation to possible further metabolization of glyoxylate to CO2 through pathways that are still not well-understood [44,45]. It was suggested that plastidic pyruvate dehydrogenase may be able to decarboxylate glyoxylate in the chloroplasts . However, experimental evidence for the participation of this or other pathways in chloroplastic glyoxylate catabolism in the transgenic plants is needed.
Recently, an alternative synthetic bypass to photorespiration was tested in rice . In this approach, glycolate is completely metabolized to CO2 in the chloroplasts through the sequential action of glycolate oxidase, oxalate oxidase, and catalase (Figure 2). Field trials showed that the transgenic rice showed significant increases in photosynthetic efficiency, biomass yield, and nitrogen content relative to the wild-type plants . In contrast, grain yield varied with the season. The improvements observed with this pathway are attributed to a CO2-concentrating effect rather than to an improved energy balance.
The Parry group proposed an alternative bypass to photorespiration in which glyoxylate is converted to hydroxypyruvate in the peroxisomes through the action of glyoxylate carboligase and hydroxypyruvate dehydrogenase . This synthetic pathway bypasses the mitochondrial reactions; with this, it avoids decarboxylation and deamination. Although this pathway is potentially energy efficient, its impact on plant growth remains to be thoroughly tested.
Another promising alternative pathway that still has to be experimentally tested in plants is the bacterial 3-hydroxypropionate pathway that converts glycolate to pyruvate . This CO2-fixing synthetic photorespiratory bypass, which needs the expression of six enzymes, was already tested in the model cyanobacterium Synechococcus elongatus . This pathway not only prevented the loss of but also resulted in a net gain in carbon fixation rate; nevertheless, transformants showed no obvious changes in growth and development relative to the wild type.
Designing energy-efficient bypasses to photorespiration
It is intriguing that although energy balances and kinetic modeling indicate that the already tested synthetic pathways result in a minimal or even negative energy advantage over the wild-type photorespiratory pathway [51,52], in all cases enhanced plant growth was observed at least in some conditions. Metabolic modeling suggests that an increase in CO2 levels in chloroplasts could result in increases in photosynthesis and biomass . It is thus conceivable that not only an energetic advantage but also a concomitant increase in CO2 levels in chloroplasts is responsible for the growth stimulation observed in the synthetic pathways tested in vivo.
The photorespiratory pathway is highly interconnected with other central metabolic routes, such as nitrate assimilation [54,55], amino acid metabolism [34,56,57], folate metabolism , and the tricarboxylic acid cycle [59,60]. Moreover, by transporting excess reducing equivalents out of the chloroplast, photorespiration may be a mechanism for preventing photoinhibition [61,62]. Such a cross-talk of the photorespiratory pathway with other routes might need co-ordinated regulation. In addition, post-translational modifications of core players of the photorespiratory pathway are known [63–67], and most probably many more remain to be elucidated; efficient engineering of photorespiratory bypasses may require an increased understanding of photorespiratory regulation.
Moreover, it is conceivable that the bypasses to photorespiration might have different effects in individual species, as specific physiological and anatomical features influence particular photosynthetic aspects, such as the refixation of CO2 produced by photorespiration and mitochondrial respiration .
Theoretical models of photorespiration and of new synthetic bypasses to photorespiration need to consider all these interactions and regulations. As models can only be as good as their input data, the achievement of a fundamental understanding of photorespiratory fluxes will depend on the generation of comprehensive experimental data on metabolic interactions and the regulation of photorespiratory enzymes in specific species.
Finally, most pathways tested until now rely on known enzymatic activities and pathways. As in the case of alternative CO2-fixing pathways [12,13], designing novel synthetic pathways to bypass photorespiration could be achieved through modeling-based strategies [69,70]. Recently, through in silico design and kinetic-stoichiometric modeling, bypasses to photorespiration were designed that metabolize 2-PG without releasing CO2 . Most of these synthetic carbon-conserving photorespiratory bypasses make use of enzymatic activities that are not present in plants or even were never identified in nature. Before these computer-designed synthetic pathways can be introduced in vivo, the implementation of the heterologous enzymes might need their engineering to adapt them to the new cellular and metabolic environment, for example through detailed computational design or directed in vitro evolution.
An important point to take into consideration for the implementation of energy-efficient bypasses to photorespiration is that all existing implementations used traditional genetic engineering technologies. Although genetically modified crops could potentially solve many of the world's hunger problems and help preserve the environment, public concerns over potential risks for human health, ecosystems, and biodiversity limit the acceptance of the use of transgenic plants. To address these concerns, it would be highly desirable to implement bypasses to photorespiration using new molecular techniques, such as genome editing via the CRISPR-Cas9 technology .
Importance of the field: Plant photosynthetic efficiency is a trait that has not reached its theoretical maximum, hence manipulation of photorespiration to improve photosynthesis offers enormous potential to enhance plant productivity.
Current thinking on the field: Photorespiration is essential for all oxygenic photosynthetic organisms and thus cannot be completely eliminated. Most of the already implemented synthetic bypasses to photorespiration in model plants showed an improvement of plant photosynthetic performance and growth under laboratory and field conditions. Together with the current feasibility to design, synthesize, and introduce biochemical pathways in vivo, these arguments support the use of synthetic biochemical pathways to bypass photorespiration as a highly promising approach to enhance plant productivity. Nevertheless, possible drawbacks of altering photorespiration, such as perturbed redox metabolism and signaling as well as negative impacts on primary nitrogen assimilation, must be taken into consideration in such endeavors.
Future directions: The future implementation of new or improved bypasses to photorespiration relies on overcoming several challenges, such as (i) the improvement and establishment of effective transformation strategies of major crops; (ii) the optimization of transgene expression to balance the enzymatic activities required, which is not only important for the primary flux of the synthetic pathway but also to avoid negative side effects of byproducts (e.g. H2O2); (iii) the implementation of new molecular techniques to overcome public concerns about the use of genetically modified plants; (iv) the use of protein engineering to develop novel enzymatic activities suggested by in silico design of synthetic pathways; and (v) the consideration of metabolic interactions of photorespiration with other cellular processes, as well as their post-translational regulation, when modeling energy-efficient synthetic biochemical pathways. An improved understanding of this last point may help close a critical knowledge gap, which is exposed by the fact that current modeling efforts still do not predict the observed improvements in plants engineered to express energy-efficient synthetic bypasses to photorespiration.
This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy — EXC 2048/1 — Project ID: 390686111 and EXC 1028 to VGM.
The Author declares that there are no competing interests associated with this manuscript.