Manipulating photorespiration to increase plant productivity: recent advances and perspectives for crop improvement.

1 Manipulating photorespiration to increase plant productivity: recent advances and perspectives for crop improvement. Marco Betti1,†, Hermann Bauwe2, Florian A. Busch3, Alisdair R. Fernie4, Olivier Keech5, Myles Levey6, Donald R. Ort7,8, Martin A.J. Parry9, Rowan Sage10, Stefan Timm2, Berkley Walker7,11, Andreas P.M. Weber12 1Departamento de Bioquímica Vegetal y Biología Molecular, Facultad de Química, 41012 Sevilla, Spain. 2Plant Physiology Department, University of Rostock, D-18051 Rostock, Germany. 3Research School of Biology, The Australian National University, Canberra ACT 2601, Australia 4Max-Planck Institute of Molecular Plant Physiology, 14476 Potsdam-Golm, Germany. 5Department of Plant Physiology, Umeå Plant Science Centre, Umeå University, S-90187 Umeå, Sweden. 6Institute of Plant Molecular and Developmental Biology, Heinrich-Heine-University, 40225 Düsseldorf, Germany. 7Global Change and Photosynthesis Research Unit, United States Department of Agriculture/Agricultural Research Service, IL 61801 Urbana, United States. 8Institute for Genomic Biology, University of Illinois, IL 61801 Urbana, United States. 9Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, United Kingdom 10Department of Ecology and Evolutionary Biology, University of Toronto, 25 Willcocks Street, Toronto, Ontario, Canada, M5S 3B2. 11Carl Woese Institute for Genomic Biology, University of Illinois, IL 61801 Urbana, United States. 12Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), HeinrichHeine-University, 40225 Düsseldorf, Germany. †To whom correspondence should be addressed. E-mail: [email protected] Tel: +34 954556917 Fax: +34 954626853 Date of submission: 10th of December 2015 Tables: 1 Figures: 0 Total word count: 10,829 2 ABSTRACT 1 Recycling of the 2-phosphoglycolate generated by the oxygenase reaction of 2 Rubisco requires a complex and energy-consuming set of reactions collectively 3 known as the photorespiratory cycle. Several approaches have been proposed 4 with the aim of producing plants with reduced rates of photorespiration energy or 5 carbon loss, both by screening for natural variation and by means of genetic 6 engineering. Recent works indicate that plant yield can be substantially improved 7 by the alteration of photorespiratory fluxes or by engineering artificial bypasses 8 to photorespiration. However, there is also evidence indicating that, under certain 9 environmental and/or nutritional conditions, reduced photorespiratory capacity 10 may be detrimental for plant performance. Here, we summarize recent advances 11 obtained in photorespiratory engineering and discuss prospects for these advances 12 to be transferred to major crops to help address the globally increasing demand 13 for food and biomass production. 14 15 Keywords 16 Crops, Food production, Genetic engineering, Photorespiration, Rubisco, Yield 17 improvement 18 19 Highlight 20 Manipulation of the photorespiratory pathway may greatly increase plant 21 productivity. Here we summarize recent advances in the engineering of 22 photorespiration and discuss how to use these approaches for crop improvement. 23 24 25 26 27 28 29 30 31 32 33 34 3 Introduction 35 36 There is an urgent demand for increased crop productivity due to the world’s 37 population growth, increasing global affluence, reduction of cultivable soils and 38 higher demand for plant based biofuels. The required increase in agricultural 39 productivity required by 2030 may be in the range of 60 to 120% as compared to 40 the levels of 2005 (Ort et al., 2015). A rapid increase in crop yield, especially for 41 cereals, was obtained in the second half of the 20th century during the so-called 42 “Green Revolution”. Resulting from breeding strategies, this led to the 43 introduction of new crop strains with a greater proportion of biomass partitioned 44 into grains and greater inputs of fertilizer, pesticides and water. However, 45 increases in yield for several major crops such as rice in recent years have been 46 scarce (Zhu et al., 2010), and it is possible that actual crop yield is approaching 47 the ceiling of maximal yield potential (Tilman et al., 2002). Further increases in 48 nitrogen and phosphorous fertilization are unlikely to solve this problem and 49 indeed many countries are currently attempting to reduce the levels of fertilization 50 used in intensive agriculture. For these reasons, attention is being paid to the 51 improvement of photosynthesis, a process that is still far from its theoretical 52 maximum efficiency. Several recent reviews summarise the opportunities that 53 have been so far identified to improve photosynthetic efficiency (Zhu et al., 2010; 54 Raines, 2011; Maurino and Weber, 2013; Long et al., 2015; Ort et al., 2015). 55 Photosynthetic CO2 fixation starts with the carboxylation of ribulose 1,5-56 bisphosphate (RuBP), catalysed by ribulose 1,5-bisphosphate carboxylase-57 oxygenase (Rubisco), to yield two molecules of 3-phosphoglycerate (3PGA). An 58 unavoidable side reaction of Rubisco is the oxygenation of RuBP to produce one 59 molecule of 3PGA and one molecule of 2-phosphoglycolate (2PG). 60 Photosynthetic organisms evolved a complex pathway to recycle 2PG that involve 61 reaction taking place in chloroplasts, peroxisomes, mitochondria and the cytosol, 62 (Bauwe et al., 2010). In this photorespiratory cycle, two molecules of 2PG are 63 transformed into one molecule of 3PGA and one carbon atom is lost as CO2 with 64 an addendant cost of 4 NAD(P)H and 7 ATP. Photorespiration has long been 65 viewed as a target for crop improvement due to the seemingly wasteful nature of 66 the cycle and the high energetic cost that it imposes on plant metabolism. 67 4 The cost of photorespiration is massive at both the leaf and canopy scale. 68 CO2 is lost from photorespiration under 25°C at about 25% the rate of net CO2 69 fixation (Sharkey, 1985; Sage et al., 2012). For example, photorespiration results 70 in the loss of ~322 trillion Calories annually in the US Corn Belt alone. Even a 71 5% reduction in photorespiration would be worth almost $540 million a year in 72 yield gain in this growing region (Walker et al., submitted for pubblication). This 73 high cost stems in part from the energy used in the reassimilation of the ammonia 74 produced following glycine decarboxylation in the mitochondrion. Moreover, 75 rates of photorespiration increase with temperature and the scarcity of water as 76 these conditions favour increased Rubisco oxygenation (Walker et al., submitted 77 for pubblication). It is thus not surprising that several groups tried to develop 78 plants with reduced rates of photorespiration with the aim of increasing 79 productivity (Peterhänsel et al., 2013a). However, the view of photorespiration as 80 a pathway that only aims at recycling the carbon of 2PG may be simplistic. In 81 addition to photosynthesis, photorespiration interacts with several central 82 metabolic pathways (Foyer et al., 2009; Bauwe et al., 2010; Fernie et al., 2013), 83 and both the relevance and the regulatory aspects of these interactions need 84 further investigations. Furthermore, photorespiration may contribute substantially 85 to the production of serine (Benstein et al., 2013; Ros et al., 2013) and has been 86 implicated in the response to certain biotic (Taler et al., 2004) and abiotic stresses 87 (Wingler et al., 2000; Voss et al., 2013). It was additionally recently demonstrated 88 that there is a positive correlation between photorespiration and productivity 89 (Aliyev, 2012) and between photorespiration and nitrate assimilation (Bloom et 90 al., 2010). While most efforts are aimed at generating plants with reduced 91 photorespiratory rates, the eventual performance of these plants in the field and 92 thus under stress conditions needs also to be considered. Tantalizing results have 93 been obtained by re-engineering photorespiratory pathway in model plants 94 (Kebeish et al., 2007; Timm et al., 2012a) or easy to transform non-staple crops 95 such as tobacco (Lin et al., 2014a), the transfer of these manipulations to our 96 major crops and demonstration of benefits under field conditions is still lacking. 97 In this article we summarise the different approaches that have been used to 98 manipulate photorespiration and their possible application for crop improvement. 99 100 Screening for plants with naturally reduced rates of photorespiration 101 5 102 Screenings of mutagenized plants that showed an altered phenotype under normal 103 air conditions but not under conditions in which photorespiration is suppressed 104 (CO2-enriched atmosphere) were carried in several C3 species, notably barley and 105 Arabidopsis (Sommerville and Ogren, 1992; Blackwell et al., 1988; Foyer et al., 106 2009; Peterhänsel et al., 2010). This approach permitted the identification of the 107 genes that encode for the core enzymes of the photorespiratory cycle. However, 108 the mutants obtained generally show poor performance under normal air 109 conditions associated with different stress symptoms (Timm and Bauwe, 2013). In 110 another approach, natural variants with reduced rates of photorespiration 111 associated with higher yields were screened across broad populations. While 112 preliminary trials carried out with tobacco gave promising results (Zelitch and 113 Day, 1973), subsequent studies failed to identify plants with low levels of 114 photorespiration paralleled by high productivity. Zelitch (1989) successfully 115 isolated plants resistant to high levels of O2 but the trait seemed more related to 116 increased levels of catalase than to reduced rates of photorespiration. Other works 117 of the same author identified tobacco plants with low photorespiratory rates and 118 high catalase activity associated to higher yield, but this increase in yield was not 119 robust across harvests (Brisson et al., 1998; Zelitch, 1992). Similarly, screening of 120 mutagenized tobacco plants identified genotypes with higher yield at low CO2 121 concentrations but the high yield trait could not be related to reduced 122 photorespiration (Medrano et al., 1995). A more recent study that summarized the 123 data obtained over 40 years of field trials using two major crop species, wheat and 124 soybean, concluded that attempts to find highly productive genotypes with high 125 photosynthetic but low photorespiratory rates are inconsistent instead showing 126 that the highly productive cultivars have high rates of photosynthesis 127 accompanied by high rates of photorespiration (Aliyev, 2012). These results, 128 argue against the use natural variation as a strategy to alleviate the yield penalty of 129 photorespiration suggesting that genetic engineering might be the only viable 130 route. 131 132 Enhancing the amount of photorespiratory CO2 scavenging 133 134 6 The CO2 released during the decarboxylation step of photorespiration in 135 mitochondria is not completely lost for the plant. On its way out of the cell, the 136 released CO2 can be refixed while passing through the chloroplasts (Sage and 137 Sage R, 2009; Busch et al., 2013). Some plants optimized this mechanism known 138 as photorespiratory CO2 scavenging by maximizing the likelihood for CO2 to pass 139 the chloroplasts. Firstly, these plants enhanced the surface of chloroplasts via 140 stromules, connecting them to a net like structure (Sage and Sage R, 2009). 141 Secondly, they associated chloroplasts tightly with mitochondria and peroxisomes 142 (Sage and Sage R, 2009; Busch et al., 2013). Rice has such morphological 143 features and it was shown that its CO2 compensation point is lower than that of 144 other C3 crops not showing this morphological adaption (Sage et al., 2009). 145 Similar to rice, the dicot C3 plants Flaveria pringlei and Flaveria robusta also 146 associated all three organelles and showed a reduced CO2 compensation point 147 compared to other C3 Flaveria species (Sage et al., 2013; Sage et al., 2014). 148 Although the effect of this anatomical adaption is not as big as the one found in 149 C4 or C2 photosynthesis plants, it still accounts as a considerable improvement 150 (Sage et al., 2013). Therefore, installing this anatomy in a C3 crop plant might be 151 an alternative approach to optimize the yield. Compared to other approaches, a 152 modification of cell anatomy should have little impact on cells metabolism. To 153 install this anatomy in a plant, a better understanding of organelle movement and 154 partitioning is needed. Natural varieties of rice and other plants showing an 155 enhanced chloroplast surface and tight connecting of the three organelles should 156 be analysed. Additionally a mutant screen of these varieties combined with RNA 157 sequencing might reveal major regulators for the anatomy of cell organelles. 158 Interestingly, in Arabidopsis thaliana, it was shown that stromules, which are 159 used to enlarge the chloroplast surface, were established when plants were 160 stressed with heat (Holzinger et al., 2007). It would therefore be of interest to 161 study mutant lines affected in stromule formation such as arc(s) (Holzinger et al., 162 2008), or even lines affected in chloroplast movement such as chup1 (Oikawa et 163 al., 2008) and compare the rates of CO2 fixation of these mutants with the wild-164 type ones. 165 166 Introducing C4 metabolism into C3 species 167 168 7 C4 photosynthesis greatly reduces photorespiration by concentrating CO2 at the 169 active site of Rubisco. With the exception of the so-called single-cell C4 170 plants (Sharpe and Offermann, 2014), C4 plants have adopted different 171 biochemical and anatomical modifications. C4 leaves have two distinct layers of 172 photosynthetic tissue (the so called “Kranz” leaf anatomy): mesophyll cells that 173 are in contact with atmospheric CO2 via intercellular air spaces, and bundle sheath 174 cells with cell walls that are less permeable to CO2. CO2 is assimilated into 175 oxaloacetate in the mesophyll cells via PEP carboxylase, which is then converted 176 to a more stable 4-carbon organic acid, malate or Asp, which diffuse to the bundle 177 sheath cells (Gowik and Westhoff, 2011). Here the C4 acid is decarboxylated, 178 releasing CO2 near the active site of Rubisco, which is located only in this cell 179 type in C4 plants. Given the higher efficiency of the C4 photosynthetic mechanism 180 under current atmospheric [CO2], efforts are underway to install C4 181 photosynthesis in C3 plants such as rice (the International C4 rice consortium, 182 http://c4rice.irri.org/) and other crops (www.3to4.org). While the number of genes 183 necessary for the main enzymatic reactions and transporters involved in C4 184 photosynthesis is relatively small, the introduction of C4 photosynthesis into C3 185 crops will also require major changes in leaf anatomy (von Caemmerer et al., 186 2012). Initial progress toward the identification of the genes responsible for C4 187 anatomy has been reported (Feldman et al., 2014; Rizal et al., 2015). On the other 188 hand, terrestrial plants capable to carry out C4 photosynthesis within a single cell 189 were discovered about 10 years ago (Sharpe and Offermann, 2014). While these 190 plants lack the typical Kranz features, they possess a subcellular separation that 191 enables a concentrating of CO2 at the active site of Rubisco. The genes involved 192 in the development of this peculiar subcellular anatomy are unknown. 193 Considering the scarcity of sequence information for single cell C4 species, it is 194 difficult to judge if single cell C4 metabolism can be bio-engineered into C3 195 crops. 196 197 Introduction of CO2-concentrating mechanisms into chloroplasts 198 199 Another strategy to reduce oxygenation and thereby photorespiration is to 200 introduce cyanobacterial CO2-concentrating mechanisms (CCM) into the 201 chloroplasts of land plants (Zarzycki et al., 2013). Cyanobacteria suppress the 202 8 oxygenating reaction of Rubisco by concentrating CO2 inside a proteinaceous 203 microcompartment called carboxysome. The β-carboxysome is constituted by an 204 outer shell composed of several different proteins that enclose Rubisco and 205 carbonic anhydrase, which releases CO2 inside the carboxysome. The high [CO2] 206 obtained near to the active site of cyanobacterial Rubisco suppresses oxygenation 207 thereby increasing the catalytic efficiency of the carboxylation reaction of the 208 enzyme. Furthermore, the use of CCM paves the way to potentially replace the 209 native Rubisco with the cyanobacterial enzyme that has higher catalytic rate but 210 also a lower affinity for CO2 and specificity factor (meaning that is more prone to 211 oxygenating RuBP) compared to the plant one (Zarzycki et al., 2013). This would 212 reduce the amount of Rubisco needed to sustain photosynthesis and permit the 213 allocation of nitrogen for other purposes, thus increasing nitrogen use efficiency 214 (Zhu et al., 2004). The feasibility of introducing carboxysomes into higher plants 215 was boosted by Lin et al., (2014a) demonstration that the shell proteins of the β-216 carboxysome could be assembled in Nicotiana benthamiana chloroplasts 217 producing organized, although empty, microcompartments. The same group was 218 also able to introduce a functional cyanobacterial Rubisco in tobacco chloroplasts 219 together with an internal carboxysomal protein (Lin et al., 2014b). In this instance 220 they replaced the native Nicotiana tabacum gene encoding for the large subunit of 221 Rubisco and replaced it with the large and small subunits of the Synechococcus 222 elongatus Rubisco, an enzyme with lower CO2 affinity but higher catalytic rate 223 compared to the endogenous one. The transformed lines were photosynthetically 224 competent albeit at very high [CO2] and the formation of complexes between the 225 cyanobacterial Rubisco and the carboxysomal protein was observed within the 226 chloroplast stroma as occurs during cyanobacterial β-carboxysomes biogenesis, 227 representing an important step toward the introduction of a CCM into C3 plants. 228 Simpler CCM mechanisms have been also considered for the transformation of C3 229 plants. For example, a recent work described the introduction of a cyanobacterial 230 bicarbonate transporter into tobacco chloroplasts (Pengelly et al., 2014). The 231 transformed plants expressed ample amount of the foreign transporter but 232 displayed the same CO2-assimilation rates than the WT, implying that the 233 transporter had little or no in vivo activity. 234 235 Rubisco engineering and screening for natural variation 236 9 237 Despite its central role in plant metabolism, Rubisco is a relatively inefficient 238 enzyme (Carmo-Silva et al., 2014). In addition to its oxygenase activity, Rubisco 239 also shows a relatively low kcat value for CO2 that obliges plants to produce very 240 high amounts of the enzyme in order to sustain adequate photosynthesis, 241 representing a large nitrogen investment (Zhu et al., 2007). Understandably, 242 considerable effort has been made to address these inefficiencies by trying to 243 engineer a more efficient Rubisco. One first challenge for replacing the plant 244 endogenous Rubisco with a more efficient one is that the large subunit of the 245 enzyme is encoded by a single chloroplastic gene and the small one by several 246 nuclear genes. Transformation of both the nuclear and chloroplast genomes of the 247 same plant is thus required in order to substitute the endogenous enzyme with a 248 more efficient one. Given that the active sites of Rubisco are on the chloroplast-249 encoded large subunit (Andersson, 2008), it may be possible that changing only 250 the large subunit will improve enzyme efficiency, but this would require the 251 transformation of the chloroplast genome, a technique that is currently available 252 only for a small number of species. High-resolution crystallographic structural 253 data are available for several plant Rubiscos and were used in site-directed 254 mutagenesis approaches in order to try to improve Rubisco efficiency. However, 255 this effort was hindered by the propensity of plant Rubisco to form insoluble 256 aggregates when expressed in E. coli, probably caused by the lack of the complex 257 network of chaperonins needed for the correct folding of the plant enzyme in the 258 bacterial host (Saschenbrecker et al., 2007; Liu et al., 2010; Feiz et al., 2012). 259 For this reason, structure-function studies were carried out mainly with the 260 enzymes from cyanobacteria and from the alga Chlamydomonas reinhardtii 261 (Whitney et al., 2011a; Parry et al., 2013 and references therein). Another 262 limitation to rational Rubisco engineering is our poor knowledge of the 263 mechanism of Rubisco-catalysed oxygenation (Tcherkez, 2015). To overcome 264 these technical difficulties, Whitney et al. (2011b) used transplastomic tobacco 265 lines that expressed WT and mutated genes encoding the large Rubisco subunit 266 from either C3 or C4 plants as well as from C3-C4 intermediate species. Using this 267 approach, the investigators were able to identify a single amino acid residue 268 responsible for the different catalytic properties of the Rubiscos from C3 and C4 269 plants (low kcat combined with low Km for CO2 and high kcat combined with high 270 16 photorespiratory enzymes could undergo oxidative modifications for some of 474 their cysteine residues, and were therefore identified as potential targets for redox 475 regulations (Keech et al., submitted for publication). Undoubtedly, the next step 476 will be to determine primarily the extent to and the conditions for which the 477 proteins or cysteines are modified, the type of modifications that occur, and 478 secondly whether these modifications positively or negatively regulate enzyme 479 activities, and how they are controlled at the cellular level. Altogether, this 480 clearly indicates that a rational bio-engineering of plants with modified levels of 481 photorespiratory enzymes would also benefit from an increased knowledge of the 482 biochemical regulations inherent to this cycle. 483 484 Perspectives for crop improvement 485 486 As summarized in the above sections and in Table 1, several approaches have 487 been used in order to manipulate photorespiration in attempt to increase plant 488 yield. However, most of these efforts have been carried out using model plants 489 (with some notable exceptions like the consortia working on the transformation of 490 rice into a C4 plant, see http://c4rice.irri.org/). In the light of the results obtained 491 by recent field trials (Aliyev, 2012), it would appear unlikely that crops with 492 improved photorespiratory performance can be obtained by screening for natural 493 genetic variation, but they should be rather generated by means of genetic 494 engineering. Unfortunately, transformation of our major crops is still a difficult 495 and time-consuming process, even if is getting easier and more successful every 496 year. Moreover, some promising approaches such as the engineering of the large 497 subunit of Rubisco require the transformation of chloroplast DNA, a technique 498 that is available only for a few crop species: notably tobacco, potato, tomato and 499 perhaps soybean, but as yet not cereal species (Scharff and Bock, 2014). As a first 500 step, organisms for which transformation is more tractable such as algae and 501 cyanobacteria can be used in order to obtain clues on the metabolic and 502 physiological consequences of a targeted genetic manipulation. A second step 503 may be the use of tobacco; a plant that is especially easy to transform both in the 504 nuclear and plastid genomes and forms canopies in the field that are similar to 505 those of food crops (Long et al., 2015). Even after careful experimental design 506 and test in intermediate plant models, several challenges would need to be 507 17 overcome before new genes and pathways can be introduced into crops. As 508 mentioned before, nuclear and especially plastid transformation techniques are 509 still inefficient or unavailable for most staple crops. In addition to that, promoters 510 and vectors that can permit high expression of transgenes and a correct subcellular 511 localization of the protein product should be available, together with strategies to 512 avoid gene silencing and random insertion in the genome (see Ort et al., 2015 for 513 a more detailed discussion on this topic). It should also be taken into consideration 514 that crops with engineered photorespiratory pathways will be considered as 515 genetically modified plants (GMP), and the potential use of such GMPs will 516 remain limited under the current legislation, which furthermore can vary greatly 517 between countries. For example in the European Union the authorization 518 procedure for placing a GMP on the market is a long, complex and expensive 519 procedure regulated by directives that were approved more than 10 years ago 520 (more details in Hartung and Schiemann, 2014). Furthermore, due to social and 521 political rejection of GMPs, even those transgenic plants that have been approved 522 are not cultivated in most EU countries. On the other hand, several millions of 523 hectares of GMPs are growing in countries with less restrictive regulations such as 524 the United States, Canada, Brazil, India and China. That said, several new 525 molecular techniques, like TALENS (transcription activator-like effector 526 nuclease(s)) or the CRISPR/Cas9 system, have been developed in the recent 527 years. The use of these genome editing techniques can lead to the production of 528 plants which cannot be classified as GMPs under current legislations. The 529 European Commission is currently evaluating these techniques together with 530 cisgenesis and intragenesis, RNA-dependent DNA methylation, grafting 531 (production of chimeric plant with a wild-type scion inserted on a genetically 532 modified rootstock), reverse breeding and agro-infiltration in order to determine 533 the extent to which they should lead to genetically modified organisms (Lusser et 534 al., 2012). Promising steps towards the regulation of these techniques are being 535 given, for example mutant plants obtained with the CRISPR/Cas9 system have not 536 been considered as GMPs in a recent decision of the Swedish Board of 537 Agriculture (http://www.upsc.se/about-upsc/news/4815-green-light-in-the-tunnel-538 swedish-board-of-agriculture-a-crispr-cas9-mutant-but-not-a-gmo.html). 539 540 Should we really look for plants with lower rates of photorespiration? 541 18 542 Regardless of the difficulties that we may face to obtain plants with modified 543 photorespiratory rates, some changes in photorespiration in the field will happen 544 anyway because of the rise in atmospheric [CO2], which is predicted to double by 545 2100 (Intergovernmental Panel on Climate Change, 2014). On one hand, this 546 increase in [CO2] will reduce photorespiration by increasing CO2 fixation by 547 Rubisco. On the other hand, photorespiration should be stimulated by the 548 predicted increase of the average atmosphere temperature, and subsequently of 549 leaf canopy. Moreover, the expected increased stomatal closure caused by 550 elevated CO2 will contribute to further increase in leaf temperature. Thus, 551 photorespiratory losses are still expected to be high even in a high CO2 world. 552 Photorespiration has been traditionally considered as a wasteful and unavoidable 553 process that needs to be minimized in order to improve plant yield. However, 554 different lines of evidence suggest that reducing photorespiration may not 555 necessarily always have beneficial effects. 556 1) Plant productivity may be improved by engineering more efficient ways to 557 recycle 2PG but also by an increased capacity for photorespiratory flux. The 558 introduction of bypasses to photorespiration can lead to up to 30% of increase in 559 plant biomass (Kebeish et al., 2007; Maier et al., 2012; Nölke et al., 2014). 560 However, these beneficial effects were observed only under short day conditions 561 and/or controlled temperature and humidity, which may not always reflect the 562 conditions that crops will face in the field. Further testing of these GMPs under 563 different conditions would be needed in order to determine if photorespiratory 564 bypasses may be beneficial also under field conditions. By contrast, several 565 studies indicated that a higher capacity for photorespiratory flux is paralleled by 566 increased plant yield (see the section “Optimization of the levels of 567 photorespiratory enzymes”). A higher photorespiratory capacity would reduce the 568 levels of photorespiratory metabolites that may inhibit the Calvin-Benson cycle as 569 well as increase the rate at which photorespiratory carbon is returned to the 570 chloroplast in form of 3-PGA, thus facilitating CO2 assimilation. Therefore, CO2 571 assimilation may be improved either by bypassing photorespiration or by the 572 overexpression of bottleneck enzymes of the cycle. The best engineering strategy 573 to use will depend on the crop considered and the environmental conditions at the 574 field level. 575 19 2) Energetically wasteful and useful are not necessarily antithetic to one another. 576 As mentioned before, under stress conditions such as drought, salinity, cold, high 577 light, heat or a combination of them, an excess of NADPH may be produced that 578 could lead to an increase of reactive oxygen species (ROS). Photorespiration can 579 act as a sink for this excess of reducing power, and this welcome effect can be 580 even more important considering that different stress conditions can increase 581 photorespiratory rates. Drought and salinity for example trigger a decrease in 582 stomatal conductance, thus decreasing the CO2:O2 ratio and increasing 583 photorespiration (Kangasjärvi et al., 2012). Heat also leads to increased 584 photorespiration of decreased Rubisco specificity and secondarily due to the 585 changes in the relative solubility of CO2 and O2. It is not surprising then that 586 attention has been paid to the role of photorespiration in the response to stress 587 (Wingler et al., 2000; Voss et al., 2013). Barley mutants with reduced levels of 588 different photorespiratory enzymes as well as Arabidopsis mutants of the 589 peroxisomal hydroxypyruvate reductase (HPR1) enzyme were more sensitive to 590 drought (Wingler et al., 1999b; Li and Hu, 2015). On the other hand, rice plants 591 with increased photorespiratory capacity showed enhanced tolerance to salt stress 592 (Hoshida et al., 2000). A protective role of photorespiration in the dissipation of 593 excess energy has been already hypothesized long time ago (Heber and Krause, 594 1980) and a demonstration to this hypothesis was provided later by Kozaki and 595 Takeba (1996), who showed that photorespiration protects against photoinhibition 596 caused by high light. A more recent work demonstrated that when the 597 photorespiratory cycle is impaired, the excess of reducing power and the 598 consequent over-production of ROS prevent the repair of photosystem II, thus 599 leading to accelerated photoinhibition (Takahashi et al., 2007). A role for 600 photorespiration in the response to other kinds of stress such as chilling or 601 exposure to heavy metals has also been proposed (Voss et al., 2013 and references 602 therein). Interestingly, several photorespiratory genes are co-expressed with genes 603 involved in the resistance to Al, that although not technically a heavy metal is also 604 a stressor that constrains plant productivity (Nunes-Nesi et al., 2014a). Since 605 abiotic stress is one of the factors that most frequently limits crop productivity 606 worldwide (Mittler, 2006), the performance of plants with reduced rates of 607 photorespiration should be tested carefully under different stress conditions. This 608 should be carried out also for plants expressing bypasses to photorespiration, since 609 20 the sink effect for excess reducing power exerted by photorespiration under stress 610 conditions may be lost in such organisms. Moreover, since most of the high 611 quality soils available are already farmed, the rising demand for food would 612 probably lead to farm crops in marginal lands with poorer soil and adverse 613 climatic conditions. In such a scenario, the use of crops with high resistance to 614 abiotic stress, and not only high yield under optimal conditions, would seem to be 615 desirable. 616 Interestingly, photorespiration has also been shown to play a significant 617 role in biotic stress responses, where the H2O2 produced by the reaction of 618 glycolate oxidase in the peroxisome plays a central role in the defence from 619 pathogen attack (Taler et al., 2004; Rojas et al., 2012) and is part of the signalling 620 route that leads to programmed cell death (Mateo et al., 2004). Plants with 621 reduced rates of photorespiration or engineered with alternative routes that bypass 622 the peroxisomal part of the pathway may show increased sensitivity to pathogen 623 attacks and should also be tested carefully. In a recent report it was also showed 624 that some photorespiratory enzymes are highly expressed in plant roots (Nunes-625 Nesi et al., 2014b), so it is possible that changes in the levels of photorespiratory 626 enzymes may also affect the physiology of heterotrophic tissues. 627 3) Rates of photorespiration correlate with nitrate assimilation in hydroponically 628 grown Arabidopsis and wheat (Rachmilevitch et al., 2004; Bloom et al., 2010). 629 This relationship has even been proposed to explain the lower-than-expected 630 growth increases in plants grown under elevated CO2 and explain why many C3 631 crops and trees grow more slowly when fed with nitrate as a sole nitrogen source 632 (Bloom et al., 2011). Recent evidence suggests that these hydroponic-based 633 observations may occur at larger scales when it was shown that wheat grown 634 under free-air CO2 enrichment had higher nitrate pools and a greater 15N 635 enrichment of both total nitrogen and nitrate, observations consistent with a 636 decrease in nitrate assimilation (Bloom et al., 2014). The exact mechanism that 637 underpins this co-dependency is still unknown but it may be related to the 638 photosynthesis-dependent export of malate from the chloroplast (the ‘malate 639 valve’), which increases the levels of cytosolic NADH thus providing reducing 640 equivalents for nitrate reduction (Bloom et al., 2010). Additionally, increased 641 rates of photorespiration further result in excess NAD(P)H since photorespiration 642 consumes more ATP relative to NAD(P)H than CO2 fixation (Kramer and Evans, 643 21 2011; Walker et al., 2014). This results in excess NAD(P)H that must be 644 consumed to balance the energy demands of central metabolism with energy 645 production from the light reactions. C4 plants on the other hand assimilate NO3646 independently of atmospheric CO2 concentration since the cytoplasmic NADH for 647 nitrate reduction can be produced by the same C4 pathway instead of by 648 photorespiration (Bloom, 2015). 649 Nitrate is the most abundant form of N in agricultural soils and is the 650 major N source for most higher plants. This is despite the higher amount of 651 energy that is needed for the assimilation of NO3into organic compounds 652 compared to other N sources such as NH4+ or organic forms of nitrogen. Taking 653 this into consideration, it is possible that a reduction of the photorespiratory rates 654 in crops that use mainly NO3may lead to nitrogen deprivation. Reliance on NH4+ 655 fertilizers may not always be possible in order to circumvent this since many 656 plants show symptoms of toxicity when grown on NH4+ as the sole N source 657 (Britto and Kronzucker, 2002). 658 In conclusion, different lines of evidence have shown that engineering of 659 photorespiration may greatly improve plant CO2-assimilation and growth. Several 660 recent advances have been made in reducing photorespiratory losses in model 661 organisms as well as in some plants of agricultural relevance. A great challenge 662 will be the transfer of these advances to our major food crops, which are generally 663 more recalcitrant to genetic manipulation. Nonetheless, a rational bio-engineering 664 of plants with altered photorespiration should also take into consideration that this 665 pathway is tightly connected with several other aspects of plant metabolism and a 666 reduction of photorespiration may not always be beneficial, especially for plants 667 growing under adverse environmental conditions. Finally, taking into 668 consideration that NO3assimilation depends on photorespiration, the 669 manipulation of the photorespiratory pathway may also affect the rates of N 670 assimilation and may favour the use of one N source over another. 671 672 Acknowledgements 673 674 This article was conceived during the discussion session “Round table on future 675 avenues of photorespiration research: crop improvement” held at the meeting 676 “Photorespiration – Key to better crops” in Warnemünde in June 2015. This work 677 22 was supported by FEDER-Ministerio de Economía y Competitividad, Spain, 678 [project AGL2014-54413-R to M.B.]. 679 References Aliyev JA. 2012. Photosynthesis, photorespiration and productivity of wheat and soybean genotypes. Physiologia Plantarum 145, 369-383. Allan WL, Clark SM, Hoover GJ, Shelp BJ. 2009. Role of plant glyoxylate reductases during stress: a hypothesis. Biochemical Journal 423, 15-22. Andersson I. 2008. Catalysis and regulation in Rubisco. Journal of Experimental Botany 59, 1555-1568. Avila C, Suárez MF, Gómez-Maldonado J, Cánovas FM. 2001. Spatial and temporal expression of two cytosolic glutamine synthetase genes in Scots pine: functional implications on nitrogen metabolism during early stages of conifer development. The Plant Journal 25, 93-102. Bauwe H, Hagemann M, Fernie AR. 2010. Photorespiration: players, partners and origin. Trends in Plant Science 15, 330-336. Benstein RM, Ludewig K, Wulfert S, Wittek S, Gigolashvili T, Frerigmann H, Gierth M, Flügge U-I, Krueger S. 2013. Arabidopsis phosphoglycerate dehydrogenase1 of the phosphoserine pathway is essential for development and required for ammonium assimilation and tryptophan biosynthesis. The Plant Cell 25, 5011-5029. Betti M, García-Calderón M, Pérez-Delgado CM, Credali A, Pal’Ove-Balang P, Estivill G, Repčák M, Vega JM, Galván F, Márquez AJ. 2014. Reassimilation of ammonium in Lotus japonicus. Journal of Experimental Botany 65, 5557-5566. Blackwell RD, Murray AJS, Lea PJ, Kendall AC, Hall NP, Turner JC, Wallsgrove RM. 1988. The value of mutants unable to carry our photorespiration. Photosynthesis research 16, 155-176. Bloom AJ, Burger M, Rubio Asensio JS, Cousins AB. 2010. Carbon dioxide enrichment inhibits nitrate assimilation in wheat and Arabidopsis. Science 328, 899-903. 23 Bloom AJ, Asensio JSR, Randall L, Rachmilevitch S, Cousins AB, Carlisle EA. 2011 CO2 enrichment inhibits shoot nitrate assimilation in C3 but not C4 plants and slows growth under nitrate in C3 plants. Ecology 93, 355-367. Bloom AJ, Burger M, Kimball BA, Pinter JPP. 2014 Nitrate assimilation is inhibited by elevated CO2 in field-grown wheat. Nature Climate Change 4, 477480. Bloom AJ. 2015. The increasing importance of distinguishing among plant nitrogen sources. Current Opinion in Plant Biology 25, 10-16. Blume C, Behrens C, Eubel H, Braun H-P, Peterhänsel C. 2013. A possible role for the chloroplast pyruvate dehydrogenase complex in plant glycolate and glyoxylate metabolism. Phytochemistry 95, 168-176. Brisson LF, Zelitch I, Havir EA. 1998. Manipulation of Catalase Levels Produces Altered Photosynthesis in Transgenic Tobacco Plants. Plant Physiology 116, 259-269. Britto S, Kronzucker HJ. 2002. NH4+ toxicity in higher plants: a critical review. Journal of Plant Physiology 159, 567-584. Busch FA, Sage TL, Cousins AB, Sage RF. 2013. C3 plants enhance rates of photosynthesis by reassimilating photorespired and respired CO2. Plant, Cell & Environment 36, 200-212. Carmo-Silva E, Scales JC, Madgwick PJ, Parry MAJ. 2014. Optimizing Rubisco and its regulation for greater resource use efficiency. Plant, Cell & Environment 38, 1817-1832. Carvalho JdFC, Madgwick PJ, Powers SJ, Keys AJ, Lea PJ, Parry MAJ. 2011. An engineered pathway for glyoxylate metabolism in tobacco plants aimed to avoid the release of ammonia in photorespiration. BMC Biotechnology 11:111. Cousins AB, Pracharoenwattana I, Zhou W, Smith SM, Badger MR. 2008. Peroxisomal malate dehydrogenase is not essential for photorespiration in Arabidopsis but its absence causes an increase in the stoichiometry of photorespiratory CO2 release. Plant Physiology 148, 786-795. Dalal J, Lopez H, Vasani NB et al. 2015. A photorespiratory bypass increases plant growth and seed yield in biofuel crop Camelina sativa. Biotechnology for Biofuels 8:175. Eisenhut M, Ruth W, Haimovich M, Bauwe H, Kaplan A, Hagemann M. 2008. The photorespiratory glycolate metabolism is essential for cyanobacteria 24 and might have been conveyed endosymbiontically to plants. Proceedings of the National Academy of Sciences of the United States of America 105, 1719917204. Feiz L, Williams-Carrier R, Wostrikoff K, Belcher S, Barkan A, Stern DB. 2012. Ribulose-1,5-bis-phosphate carboxylase/oxygenase accumulation factor 1 is required for holoenzyme assembly in maize. The Plant Cell 24, 3435-3446. Feldman AB, Murchie EH, Leung H, Baraoidan M, Coe R, Yu S-M, Lo S-F, Quick WP. 2014. Increasing leaf vein density by mutagenesis: laying the foundations for C4 rice. PLoS ONE doi:10.1371/journal.pone.0094947. Fernie AR, Bauwe H, Eisenhut M et al. 2013. Perspectives on plant photorespiratory metabolism. Plant Biology 15, 748-753. Foyer CH, Bloom AJ, Queval G, Noctor G. 2009. Photorespiratory metabolism: genes, mutants, energetics, and redox signaling. Annual Review of Plant Biology 60, 455-484. Galmes J, Flexas J, Keys AJ, Cifre J, Mitchell RAC, Madgwick PJ, Haslam RP, Medrano H, Parry MAJ. 2005. Rubisco specificity factor tends to be larger in plant species from drier habitats and in species with persistent leaves. Plant, Cell & Environment 28, 571-579. Gowik U, Westhoff P. 2011. The path from C3 to C4 photosynthesis. Plant Physiology 155, 56-63. Grodzinski B, Butt VS. 1977. The effect of temperature on glycollate decarboxylation in leaf peroxisomes. Planta 133, 261-266. Haimovich-Dayan M, Lieman-Hurwitz J, Orf I, Hagemann M, Kaplan A. 2015. Does 2-phosphoglycolate serve as an internal signal molecule of inorganic carbon deprivation in the cyanobacterium Synechocystis sp. PCC 6803? Environmental Microbiology 17, 1794-1804. Hartung F, Schiemann J. 2014. Precise plant breeding using new genome editing techniques: opportunities, safety and regulation in the EU. The Plant Journal 78, 742-752. Häusler RE, Blackwell RD, Lea PJ, Leegood RC. 1994. Control of photosynthesis in barley mutants with reduced levels of glutamine synthetase and glutamate synthase. II control of electron transport and CO2 assimilation. Planta 194, 406-417. 25 Heber U, Krause GH. 1980. What is the physiological role of photorespiration? Trends in Biochemical Sciences 5, 32-34. Heineke D, Bykova N, Gardeström P, Bauwe H. 2001. Metabolic response of potato plants to an antisense reduction of the P-protein of glycine decarboxylase. Planta 212, 880-887. Hodges M, Jossier M, Boex-Fontvieille E, Tcherkez G. 2013. Protein phosphorylation and photorespiration. Plant Biology 15, 694–706. Holzinger A, Buchner O, Lütz C, Hanson MR. 2007. Temperature-sensitive formation of chloroplast protrusions and stromules in mesophyll cells of Arabidopsis thaliana. Protoplasma 230, 23-30. Holzinger A, Kwoy EY, Hanson MR. 2008. Effects of arc3, arc5 and arc6 mutations on plastid morphology and stromule formation in green and nongreen tissues of Arabidopsis thaliana. Photochemistry and Photobiology 84, 1324-1335. Hoshida H, Tanaka Y, Hibino T, Hayashi Y, Tanaka A, Takabe T, Takabe T. 2000. Enhanced tolerance to salt stress in transgenic rice that overexpresses chloroplast glutamine synthetase. Plant Molecular Biology 43, 103-111. Kangasjärvi S, Neukermans J, Li S, Aro E-M, Noctor G. 2012. Photosynthesis, photorespiration, and light signalling in defence responses. Journal of Experimental Botany 63, 1619-1636. Kebeish R, Niessen M, Thiruveedhi K et al. 2007. Chloroplastic photorespiratory bypass increases photosynthesis and biomass production in Arabidopsis thaliana. Nature Biotechnology 25, 593-599. Keech O, Zhou W, Fenske R, Colas des Francs-Small C, Bussell JD, Badger MR, Smith SM. 2012. The genetic dissection of a short term response to low CO2 supports the possibility for peroxide-mediated decarboxylation of photorespiratory intermediates in the peroxisome. Molecular plant 5, 1413-1416. Kozaki A, Takeba G. 1996. Photorespiration protects C3 plants from photooxidation. Nature 384, 557-560. Kramer DM, Evans JR. 2011. The importance of energy balance in improving photosynthetic productivity. Plant Physiology 155, 70-78. Igamberdiev AU, Bykova NV, Kleczkowski LA. 1999. Origins and metabolism of formate in higher plants. Plant Physiology and Biochemistry 37, 503-513. 32 Zelitch I. 1992. Control of plant productivity by regulation of photorespiration. Bioscience 42, 510-516. Zhu C, Fan Q, Wang W, Shen C, Meng X, Tang Y, Mei B, Xu Z, Song R. 2014. Characterization of a glutamine synthetase gene DvGS2 from Dunaliella virdis and biochemical identification of DvGS2-transgenic Arabidopsis thaliana. Gene 536, 407-415. Zhu X-G, Portis AR, Long SP. 2004. Would transformation of C3 crop plants with foreign RUBISCO increase productivity? A computational analysis extrapolating from kinetic properties to canopy photosynthesis. Plant, Cell & Environment 27, 155-165. Zhu X-G, de Sturler E, Long SP. 2007. Optimizing the distribution of resources between enzymes of carbon metabolism can dramatically increase photosynthetic rate: a numeric simulation using an evolutionary algorithm. Plant Physiology 145, 513-526. Zhu X-G, Long SP, Ort DR. 2010. Improving photosynthetic efficiency for greater yield. Annual Review of Plant Biology 61, 235-261. 33 Table 1. Summary of the approaches that can be used to improve crop yield through manipulation of photorespiration. Strategy Advantages (A) / Disadvantages (D) Screening for plants with reduced rates of photorespiration A: -No need for genetic manipulation. D: -Highly improbable to find plants with high levels of PS and low of photorespiration in field trials. Enhancing the amount of photorespiratory CO2 scavenging A: -Does not imply changes in cellular metabolism. D: -Genetic determinants of organelle partitioning and connection are not completely understood. Introduce C 4 photosynthesis into C3 plants A: -Great theoretical potential for increase in crop yield. D: -Major changes in leaf anatomy are required. -The genes responsible for C4 anatomy not completely identified. Introduction of CCM into chloroplasts A: -Should greatly reduce the rates of photorespiration. - Should allow replacing endogenous Rubisco with enzymes with higher catalytic rates and lower CO2 affinity. D: -Requires transformation of the chloroplast genome. -Complex CCM requires the transformation of multiple genes and the correct assembly of multiprotein complexes. Rubisco engineering and screening for naturally occurring more efficient Rubisco A: -Rubisco has several catalytic inefficiencies. This implies several opportunities for engineering. -Naturally occurring more efficient Rubiscos have been found in some species. D: -Structure-function studies with Rubisco are hampered by different technical difficulties. -The exact mechanism of the oxygenating reaction is still not completely understood. Photorespiratory bypasses A: -Successfully engineered in both model and crop species. -Can increase yield up to 30%. -Possibility of complete oxidation of 2PG in the chloroplast, thus raising the [CO2] near Rubisco active site. D: -Need transfer of multiple genes. -Increased yield is seen only under short day in some bypasses. -The lower energy cost of some bypasses may prevent the protective role of PR under stress conditions. Optimization of the levels of photorespiratory enzymes A: -Relatively easy genetic manipulation. D: -Transcriptional and posttranslational regulation of photorespiratory genes and enzymes is still poorly characterised.