Charting New Pathways to C₄ Rice

Foreword.

Preface.

Contents.

Rice is the most important crop in the world for human food. Over the past 40 years, its production has kept pace with the increase in population. However, it is clear that the gains of the first Green Revolution are largely exhausted. Rice with C₄ photosynthesis could make a major contribution to a second Green Revolution. To assess how that change could affect rice, it is necessary to understand how the rice crop works.

In this paper, we examine the properties of individual rice plants both as single individuals and as members of dense crop communities. To estimate the potential of C₄ rice, we compare the yields and radiation-use efficiencies of maize, rice, and a C₄ weed. In that context, the properties of rice canopies with respect to the interception of solar radiation and its effect on leaf temperature are examined. The influence of sink size with respect to source strength is also discussed. It is possible that wild rice types have some of the anatomical features peculiar to C₄ plants and that the wild types may contain C₃-C₄ intermediates. Consequently, we report results obtained from an examination of C₄ characteristics in the 22 species of wild rice.

The case for C₄ rice is summarized as a chain of argument, starting from the need to produce more rice, through higher yield potential, to the consequent improvements in growth and photosynthesis. We conclude that it is essential to make Rubisco, the key enzyme of photosynthesis, work harder, by concentrating carbon dioxide around Rubisco, thus raising the light-saturated rate of photosynthesis and greatly reducing photorespiration, as occurs in C₄ photosynthesis. The additional agricultural benefits of shifting from C₃ to C₄ rice are that C₄ photosynthesis requires (1) less Rubisco and hence less nitrogen, and (2) less water, since a steeper concentration gradient for carbon dioxide diffusion can be maintained through partly closed stomata.

The chain of argument also shows that research should be planned from the perspective of what we wish to achieve, that is, from the top downward, and quantitatively as far as possible. Developments in screening phenotypes, plant breeding, molecular biology, and genetic engineering are proceeding rapidly and need to be directed toward the applied goal. The pathway to success cannot be seen completely but it is very likely that techniques will arise to enable the construction of C₄ rice. The certainties of population growth, climate change, and future shortages of water for agriculture mean that it is essential to start research now.

Rice is the most widely planted crop in Asia and it is the dominant source of calories for Asians living in poverty. Since a majority of the world's poor live in Asia, improving the livelihoods of rice farmers and consumers is therefore critical to global poverty alleviation. Poor farmers need high profits from growing rice, and poor consumers need lower prices so that they can increase the quality and quantity of food consumption and still have money left over for investments in education. Agricultural research is perhaps the most important way to achieve both high profits for farmers and low prices for consumers.

Although some trends in the rice economy give cause for optimism regarding the future path of rice prices, other trends raise concerns that prices will increase in the next 20 years. Uncertainties abound regarding future oil prices, demand for biofuels, water scarcity, climate change, and the pace of slowdowns in population growth and dietary diversification. In the face of such uncertainty, it seems prudent to invest in research for C₄ rice. Without such investments in productivity-enhancing technologies, it is the poor who will suffer the most from adverse shocks that put upward pressure on food prices.

Developments in understanding C₄ photosynthesis since 1999 are reviewed to provide a starting point for the conference on Supercharging the Rice Engine in 2006. It is now clear that all C₃ plants have the enzymes for the C₄ pathway and some plants employ them to capture carbon dioxide in particular tissues, although not as full C₄ photosynthesis with reduced or absent photorespiration. Conversely, parts of C₄ plants, mesophyll cells not close to bundle sheath cells, carry out C₃ photosynthesis. Several types of single-cell C₄ photosynthesis have now been studied; none seem to be highly productive but instead are adaptive for these plants in environments conducive to high rates of photorespiration. Introduction of one or a few genes for C₄ enzymes into rice has become routine but full C₄ photosynthesis or greatly increased growth has not resulted. Kranz anatomy appears to be essential for productive C₄ photosynthesis because there is no other way of confining carbon dioxide at high concentrations around ribulose 1,5-bisphosphate carboxylase–oxygenase (Rubisco), which is the key to success. At least 45 independent origins of C₄ photosynthesis across 19 angiosperm families are known. Study of C₃-C₄ intermediates has led to the definition of seven phases of evolution from C₃ to C₄. Given low atmospheric concentrations of carbon dioxide and high rates of photorespiration, C₄ photosynthesis seems to evolve readily, although not in the subfamily of grasses containing rice. Nevertheless, these recent advances in knowledge are encouraging for those intending to produce C₄ rice.

This article explores some of the structural and physiological changes that may be required for the introduction of C₄ photosynthesis into the leaf of a C₃ plant, such as rice. Transport, both between cells and within cells, is essential to the operation of a C₄ photosynthetic system. In this article, I discuss some of the relationships between mesophyll and bundle sheath cells, and how transport of photosynthetic intermediates via the plasmodesmata is driven by large concentration gradients of metabolites. In turn, the properties of the enzymes of carbohydrate synthesis are modified to accommodate the high concentrations of metabolite precursors. The properties of intracellular metabolite transporters are also modified in C₄ photosynthesis. It is also likely that metabolites moving between the mesophyll and bundle sheath act as the messengers that ensure that individual enzymes are regulated so that photosynthetic fluxes are coordinated between the two cell types.

The advantages of C₄ photosynthesis for plant productivity, particularly in warmer climates, are well characterized. High rates of biomass accumulation and high water-use efficiency and N-use efficiency make the installation of the C₄ pathway (or some other form of CO₂-concentrating mechanism) into C₃ plants an attractive proposition for biotechnologists. Here, we compare anatomical properties of leaves of C₃ and C₄ species to compare characteristics of CO₂ diffusion. We show that leaves of a wide variety of C₃ species are characterized by high exposed mesophyll and chloroplast surface area to leaf area ratios (S_m and S_c). Combining measurements of the internal conductance to CO₂ diffusion (derived from measurements of carbon isotope discrimination) with measurements of S_c shows that the CO₂ conductance across the cell wall, plasma membranes, and chloroplast membrane interface is on average 0.02 mol m^-2 chloroplast area s^-1 bar^-1 for C₃ annual species (including rice) and 0.01 to 0.02 mol m^-2 chloroplast area s^-1 bar^-1 for deciduous and evergreen trees. Measurements of anatomical properties of a number of C₄ species show that S_m is less in C₄ species than in C₃ species, but that high photosynthetic rates require higher conductances for CO₂ diffusion across the C₄ mesophyll cytosol interface. There is little variation in bundle sheath surface area to leaf area ratio (S_b), with average values of 1.77 ± 0.11, such that S_m is from 6 to 10 times greater than S_b. Bundle sheath conductance to CO₂ diffusion cannot be measured directly; however, the efficiency of the C₄ photosynthetic pathway can be assessed through measurements of carbon isotope discrimination. Using a mathematical model of C₄ photosynthesis, we examine the relationship between bundle sheath conductance (or its inverse, resistance) to CO₂ diffusion and the biochemical capacity of the C₄ photosynthetic pathway and conclude that bundle sheath resistance to CO₂ diffusion must vary with biochemical capacity if the efficiency of the C₄ pump is to be maintained. Finally, we use a mathematical model of single-cell C₄ photosynthesis in a C₃ mesophyll cell and examine the importance of CO₂ diffusion on such a C₄ photosynthetic CO₂ pump.

This paper reviews the identification and functional characterization of GLK genes in diverse plant species. In so doing, a hypothesis is proposed to suggest how GLK gene function differs in C₃ and C₄ species. Finally, suggestions are made about how this research and other genetic strategies could be used to produce C₄ rice.

In C₃ leaves, photosynthetic electron transport is closely coupled to the carbon reduction cycle in each chloroplast due to the small pools of nicotinamide adenine dinucleotide phosphate (NADP) and adenosine triphosphate (ATP). In contrast, in C₄ leaves, photosynthetic electron transport occurring in two different cell layers is buffered via extensive metabolite exchange and larger pools of metabolites. The demand for NADPH and ATP in the mesophyll and bundle sheath cells depends on the decarboxylation type. At one extreme, NADP malic enzyme (ME) species such as Zea mays produce almost no NADPH by linear electron flux in the bundle sheath. The NADPH required by the photosynthetic carbon reduction (PCR) cycle is either transferred into the bundle sheath via malate or phosphoglyceric acid (PGA) is cycled into the mesophyll for reduction. At the other extreme, nicotinamide adenine dinucleotide (NAD)-ME species such as Panicum miliaceum only need to produce ATP in the mesophyll to supply the C₄ cycle. These two extremes restrict linear electron flux to either the mesophyll (NADP-ME) or the bundle sheath (NAD-ME), with the remaining ATP requirement being generated from cyclic electron flux in either cell type. Biochemical diversity within C₄ species means that intermediate solutions with some linear electron flux in both mesophyll and bundle sheath cells also exist. Additional flexibility is also required for any given decarboxylation type because the requirements change depending on the leakiness of the bundle sheath to CO₂. Leakiness tends to increase at lower irradiance and under fluctuating light.

To maximize quantum yields, the different cellular locations for linear electron flux between the decarboxylation types require different distributions of light absorption between the mesophyll and bundle sheath cells. The majority of chlorophyll is co-located within the cells with linear electron flux. However, light absorption is not simply proportional to chlorophyll distribution because of the complex leaf anatomy. We visualized profiles of light absorption through leaves of Flaveria bidentis and Z. mays by imaging chlorophyll fluorescence emerging from the transverse face of a cut leaf. Green light was absorbed throughout the leaf. In contrast, blue light was strongly absorbed near the surface, with little light penetrating the bundle sheath. This resulted in the rate of CO₂ assimilation under blue light being half that under green light of the same photon irradiance. The decline in the rate of CO₂ assimilation after switching from green to blue light occurred over 100 s and represented a change in metabolite pool size of 400 μmol m^-2. We predict that leakiness is greater under blue light than under green light for a given photon irradiance. Engineering a single-cell CO₂-concentrating mechanism would be simpler than a Kranz-type C₄ system as it would require little cellular-specific adjustment to thylakoid composition and function.

The photosynthetic characteristics of four transgenic rice lines overexpressing maize phosphoenolpyruvate carboxylase (PEPC; line PC), pyruvate, orthophosphate dikinase (PPDK; line PK), PEPC + PPDK (line CK), and NADP-malic enzyme (NADP-ME; line ME) were investigated using outdoor-grown plants. Relative to untransformed wild-type (WT) rice, PC transgenic rice exhibited high PEPC activity (a 25-fold increase) and enhanced activity of carbonic anhydrase (more than a twofold increase). The PC transgenic plants also showed a higher CO₂ uptake rate and carboxylation efficiency, and slightly reduced CO₂ compensation point. Furthermore, PC transgenic rice produced 22% more grains than WT plants. Labeling with ¹⁴CO₂ for 20 s showed more ¹⁴C distributed to C₄ primary photosynthate aspartate and feeding with exogenous C₄ primary products such as oxaloacetate (OAA), malate (MA), or phosphoenolpyruvate (PEP) showed an increment of photosynthetic rate in PC transgenic rice, suggesting that a limited C₄ cycle exists in leaves of transgenic rice.

Introduction of the maize PEPC gene could activate or induce activities of the key enzymes scavenging active oxygen, such as superoxide dismutase (SOD) and peroxidase (POD).

The line JAAS45 manifested higher photosynthetic rates and photochemical efficiency of PS II (F_v/F_m). The value of δ¹³C in PC transgenic rice was similar to that in untransformed rice, demonstrating that transgenic rice is still a C₃ plant.

How can we redesign C₄ rice from the limited C₄ features of photosynthesis reached currently? In future work, introduction of the PEPC gene from a CAM plant into C₄-enzyme transgenic rice could carry out higher photosynthesis day and night. Simultaneously, the enhancement of endogenous ATP in PK transgenic rice through genetic engineering would increase its operation of the C₄ cycle. Most importantly, recent advanced techniques such as laser capture microdissection enable us to study the mechanisms of cellular differentiation, for example, of bundle sheath cells. From the above suggestions, these techniques might shed light on a new green revolution in rice breeding.

C₃ plants lose a significant part of previously fixed CO₂ in the process of photorespiration. Reduction in photorespiration is expected to increase the productivity of crop plants and reduce the requirements for irrigation and fertilization. For more than ten years, research at our institute has focused on the genetic engineering of dicotyledonous crop plants toward improved CO₂ fixation. In this paper, we summarize results form our work vis-à-vis reports from other laboratories and define future challenges. Furthermore, we introduce an alternative approach based on the installation of a bypass of photorespiration in the chloroplast.

C₄ photosynthesis is characterized by a division of labor between two different photosynthetic cell types, mesophyll and bundle sheath cells. Relying on phosphoenolpyruvate carboxylase (PEPC) as the primary carboxylase in the mesophyll cells, a CO₂ pump is established in C₄ plants that concentrates CO₂ at the site of ribulose 1,5-bisphosphate carboxylase/oxygenase in the bundle sheath cells. The C₄ photosynthetic pathway evolved polyphyletically, implying that the genes encoding the C₄ genes originated from nonphotosynthetic progenitor genes that were already present in the C₃ ancestral species. To establish a C₄ cycle in a C₃ plant, detailed knowledge about the components of C₄ photosynthesis and the differences of these components in C₃ and C₄ plants is needed.

C₄ photosynthesis has evolved naturally over 50 times in 19 families of flowering plants. This repeated evolution of a complex trait indicates that it is either relatively easy or was under an intense directional selection pressure. Either way, the repeated evolution of C₄ photosynthesis indicates that it should be feasible to create C₄ rice plants by engineering C₄ genes into C₃ rice and replicating strong selection pressures for C₄ traits that we think exist in nature. Studies of the natural lineages, particularly those using species with intermediate characteristics of C₃ and C₄ photosynthesis, reveal the probable phases and selection pressures in the evolution of C₄ photosynthesis. A key early step is the formation of leaves with close vein spacing and slightly enlarged bundle sheath cells. Following this, the photorespiratory enzyme glycine decarboxylase is localized to the bundle sheath tissue, which allows for CO₂ to be concentrated into the bundle sheath, thereby improving photosynthetic efficiency at low atmospheric CO₂ concentration. Localization of glycine decarboxylase to the bundle sheath tissues facilitates the creation of Kranz-like anatomy and elaborate transportation networks between the mesophyll and bundle sheath tissues. This prepares the lineage to evolve the C₄ cycle, and is thus considered a key link in the evolutionary bridge to C₄ photosynthesis. Natural populations of C₃-C₄ intermediate species indicate that photorespiratory conditions (high temperature and low CO₂) were the main selection agent favoring the evolution of C₄ photosynthesis. By establishing screens based on high rates of photorespiration, genotypes transformed by natural selection and mutagenized populations could be bred for increasing expression of C₄-like characters. Biotechnology approaches could accelerate the breeding process by introducing critical genes; however, a screening approach will likely be needed to improve the many unknown traits involved in the evolution of C₄ plants.

Placing the efficient C₄ pathway into rice appears ambitious because it would involve modifications to biochemistry, leaf development, and cell biology. The biochemical modifications need to include high expression of genes encoding carbonic anhydrase, phosphoenolpyruvate carboxylase, malate dehydrogenase, and pyruvate orthophosphate dikinase in the mesophyll, while a decarboxylase and Rubisco are specifically needed in the bundle sheath. Alterations in leaf development required are increased venation, larger bundle sheath cells, and fewer mesophyll cells. Changes in cell biology include chloroplast proliferation and expansion in the bundle sheath, and increased plasmodesmatal connectivity between mesophyll and bundle sheath cells.

Although these modifications appear complex, C₃ species have the ability to accumulate proteins needed for C₄ photosynthesis in defined cell types, and it also appears that they possess trans-factors needed for the expression of genes specifically in mesophyll or bundle sheath cells. When intact genes from a C₄ species are placed in a closely related C₃ plant, they are expressed in the correct cell type for C₄ photosynthesis, but in more distantly related species this is less likely. It should therefore be possible to integrate enzymes needed for C₄ photosynthesis into rice if genes are sourced from a closely related C₄ plant.

I propose a dual-track approach to the challenge of integrating C₄ traits into rice. First, studies of rice leaf development are needed. Second, fundamental work is needed on C₄ photosynthesis itself, and the species used should depend on the particular question being asked. The hypothesis that introducing the biochemistry of C₄ photosynthesis into a C₃ plant leads to leaf development associated with the C₄ pathway should be tested. It would be fastest to do this by placing genes from Cleome gynandra into Arabidopsis thaliana. If this hypothesis is supported, a shortcut to the whole process of generating a C₄ rice could be found. If the hypothesis is not supported, many phenotypes of C₄ plants are shared, and so loci controlling these could be identified in systems other than rice and maize. For example, bundle sheath enlargement and increased plasmodesmatal connectivity should be investigated with A. thaliana and Cleome because resources and generation times are favorable.

Plans to make a C₄ rice plant date back to a document in 1987 and the first patent application for C₄ rice submitted in 1991. In addition, an attempt to make a C₄ rice plant was made in collaboration with Japan Tobacco Inc. during the 1990s. This collaboration recognized the importance of two compartments in C₄ photosynthesis, normally provided by mesophyll and bundle sheath cells. However, a single-cell system was devised in which the endogenous compartments of the cytosol and the chloroplast of C₃ plants were used to mimic the two C₄ compartments. Phosphoenolpyruvate carboxykinase (PEPCK) was used as the C₄ acid decarboxylating enzyme and was synthesized with a transit peptide to ensure location in the chloroplasts. The PEPCK gene from Urochloa panicoides was transferred to rice and was expressed successfully: carbon flow was altered toward a C₄ pathway but without appreciable increases in photosynthesis or growth. The properties and location of enzymes postulated to be required to convert a C₃ plant to a C₄ plant (carbonic anhydrase, phosphorenolpyruvate carboxylase, PEPCK, and pyruvate, orthophosphate dikinase) are reviewed. Further modifications to maximize the efficiency of a C₄ pathway in C₃ plants are discussed.

A common feature of photosynthesis in practically all organisms is the assimilation of CO₂ into organic matter via a catalyst called ribulose 1,5-bisphosphate carboxylase oxygenase (Rubisco) in the carbon assimilation cycle. One of the constraints on the process in terrestrial plants is conditions where CO₂ becomes limiting because of high temperature, drought, or soil salinity. This can occur by restricting the entry of CO₂ into leaves, by decreased stomatal conductance, by decreased cytoplasmic solubility of CO₂, and by increased photorespiration (a process resulting from O₂ competing with CO₂ in Rubisco catalysis). In response to CO₂ limitations, some terrestrial plants evolved mechanisms to concentrate CO₂ around Rubisco through a C₄ cycle that requires spatial separation of fixation of atmospheric CO₂ into C₄ acids, and the donation of CO₂ from C₄ acids via decarboxylases to Rubisco (called C₄ plants). The paradigm for C₄ photosynthesis in terrestrial plants for more than 35 years was that a dual-cell system, called Kranz leaf anatomy, is required for spatial separation of these functions. Surprisingly, recent research on species in family Chenopodiaceae has shown that C₄ photosynthesis can occur within a single photosynthetic cell. Two very novel means of accomplishing this evolved in subfamily Suaedoideae. These systems function by spatial development of two cytoplasmic domains, which contain dimorphic chloroplasts. Emerging information on the biochemical and structural strategies for accomplishing C₄ has promise for improving the productivity of rice, which lacks a CO₂-concentrating mechanism, and for securing this important crop as a food supply under CO₂-limited conditions predicted with global warming.

Ribulose bisphosphate carboxylase–oxygenase (Rubisco) is inhibited by O₂ and, as a consequence, atmospheric CO₂ does not saturate C₃ photosynthesis. The O₂ effect has two components: direct inhibition of carboxylation and an oxygenase reaction that initiates photorespiration. C₄ photosynthesis concentrates CO₂ for Rubisco, which minimizes both components, and increases photosynthesis up to 50%. Although atmospheric [CO₂] is projected to reach 550 μbar by 2050, it will not eliminate adverse O₂ effects. Rice yields will increase, but the benefit may be offset by projected higher temperatures and reductions in rice Rubisco protein. Hydrilla verticillata is a monocot that operates a facultative, single-cell C₄ system. Based on this single-cell premise, rice plants have been transformed with C₄-cycle enzymes to improve photosynthesis and yield, but the results have been disappointing. The Hydrilla system can provide clues to the essential elements needed for an effective CO₂-concentrating mechanism (CCM) because the C₄ and C₃ cycles operate in series in the same C₃ cell, without the bundle sheath anatomy of terrestrial C₄ plants. In Hydrilla, phosphoenolpyruvate carboxylase (PEPC) in the cytosol is segregated from Rubisco and the decarboxylase, NADP-dependent malic enzyme (NADP-ME), in the chloroplasts, where CO₂ is concentrated. Multiple isoforms of PEPC and NADP-ME exist in Hydrilla, but hvpepc4 and hvme1 are up-regulated in C₄ leaves and encode proteins with characteristics specific to C₄ photosynthesis. A β-carbonic anhydrase (CA) is also up-regulated, presumably in the cytosol to aid PEPC fixation, but we hypothesize that CA is down-regulated in C₄ chloroplasts. To maintain the NADPH/NADP⁺ ratio in the granal chloroplasts of C₄ leaves, oxaloacetate and/or aspartate may be imported and reduced to malate for decarboxylation. A major unknown is how the Hydrilla chloroplasts, which in the C₃ state must maximize CO₂ conductance for Rubisco, minimize this permeability to reduce leakage from the CCM when the C₄ system is induced. The down-regulation of chloroplast envelope aquaporins may be involved, and the Hydrilla system provides a means to study this crucial component. If chloroplast leakage is not regulated in a single-cell C₄ rice plant, even a high-capacity C₄ pump will be ineffective, and low quantum yields will compromise productivity. Hydrilla studies indicate that transporter and permeability issues, and the nuances of enzyme regulation, should be incorporated in the design of a single-cell C₄ rice plant to produce an effective CCM.

The 10,000 or more species of diatoms are microscopic photosynthetic organisms of the class Bacillariophyceae in the phylum Heterokontophyta. They are dominant primary producers in marine and inland water habitats, and may account for up to 20% of global primary productivity. The core carboxylation enzyme in their photosynthesis is Form ID Rubisco (ribulose bisphosphate carboxylase–oxygenase), which, if it replaced rice Form IB Rubisco on a molecule-for-molecule basis, would give slightly lower rates of photosynthesis at extant CO₂ concentrations. These kinetic characteristics, along with the low conductance for CO₂ of aqueous boundary layers, rationalize the occurrence of CCMs (inorganic carbon-concentrating mechanisms) in all diatoms investigated. It was assumed that these mechanisms, which increase the CO₂ concentration around Rubisco, were all based on active transport of CO₂, HCO₃^- or H⁺ across membranes. It now appears, from recent extensions of earlier work, that there is a C₄-like photosynthetic carbon metabolism in certain diatoms. However, more work is needed to determine the extent to which diatoms have photosynthesis analogous to that of single-cell C₄ higher plants. The relevance of this work to producing C₄ rice probably comes more from concepts than from the direct introduction of diatom genes in rice. One such concept is the possibility that C₄-like photosynthesis in diatoms involves no carbonic anhydrases (CAs), and so needs less Zn. However, this requires HCO₃^- entry, so decreased Zn costs of growth may be less readily achieved in rice unless phosphorenolpyruvate carboxykinase (using CO₂) replaces phosphoenolpyruvate carboxylase (using HCO₃^-) as the C₄ carboxylase.

C₄ photosynthesis is a system that uses resources present in C₃ plants. C₄ photosynthesis has evolved numerous times, in widely separated phylogenetic groups. All existing species that are able to fix carbon dioxide through one of the variety of C₄ schemes appear to rely on enzymatic activities and other factors present in most or all plant species, but regulated to exhibit an extreme intercellular or intracellular compartmentalization that supports the delivery of CO₂-derived carbon to Rubisco in an environment that disfavors competition from oxygen. C₄ species are particularly numerous in certain subfamilies of grasses, suggesting that the resources required for C₄ physiology are present and predisposed to this re-regulation. Rice lacks the dense leaf venation, bundle sheath (BS) differentiation, high BS plasmodesmatal density, and compartmentalization of photosynthetic activities that characterize nearly all C₄ grass species. To what extent are these C₄ resources already networked together in C₃ grasses such as rice and how might we find the targets and means for engineering the re-regulation of this network? A systems biology approach that compares the development of cell types in rice leaves to those in C₄ grasses could provide these targets and means. Emerging techniques such as laser microdissection of cell types and microarray profiling can provide the comprehensive data needed for a systems approach.

Previous studies of photosynthesis made in the field under irrigated conditions have shown that there are several potential sites of limitation at the leaf level, including light saturation of photosynthesis in upper leaves, mid-morning depression of photosynthetic capacity, and photoinhibition of photosynthesis (reduction in quantum yield). The response of rice photosynthesis in the field to a high CO₂ concentration (measured in a leaf chamber) indicates the central importance of Rubisco chemistry and photorespiration in limiting the assimilatory capacity and potential biomass production of rice under tropical conditions in the field. In this paper, we discuss the relationship between C₃ photosynthesis and rice leaf morphology and how this research may be incorporated into a program to produce a rice plant with C₄ features.

Rice crops typically demand high inputs of N fertilizer to achieve high grain yield and this is reflected in the Rubisco concentrations observed in field-grown rice leaves. We have described an inconsistent relationship in the field between Rubisco content and in situ rates of rice leaf photosynthesis in some genotypes and postulated the role of Rubisco in forming part of an N store for later remobilization to the grain. Rice leaf morphology (in this case thickness and area) is a feature of rice crops important for canopy efficiency and integrity, photosynthetic rate, and N content. However, the relationship among leaf thickness, N content, and photosynthesis is not clear. We have adopted a number of lines of research that explore the factors responsible for leaf thickness determination in rice. First, using differences in morphology induced by acclimation to irradiance, we suggest that this results from a signal provided by mature leaves. We postulate that these changes are a "fine-tuning" of cellular morphology and that the establishment of Kranz anatomy in rice may not require such signals. Second, we are exploiting genotype differences and rice mutant collections. The exploitation of new mutant resources for rice will be essential if the goal of C₄ rice is to be achieved. Although high-throughput screening of rice mutant populations is still largely impracticable, this may not apply to IRRI's IR64 deletion mutant collection.

The genus Oryza, to which cultivated rice (O. sativa) belongs, has 22 wild species (2n=24, 48) representing 10 genomes. These wild species show tremendous diversity in plant morphology, life cycle (perennial, annual), growth habit, light requirement (full sun, full shade, partial shade, etc.), including agronomic traits (resistance to biotic and abiotic stresses). Wild species are an important reservoir of useful genetic variability to broaden the gene pool of rice for tolerance of major biotic and abiotic stresses, for cytoplasmic diversification, and to introgress yield-enhancing loci/QTLs. Low crossability, increased sterility, reduced recombination, and linkage drag limit the introgression of genes from wild species into rice. Recent advances in tissue culture and molecular markers have facilitated alien introgression in rice. At IRRI, using direct crosses, embryo rescue, anther culture, molecular markers, and flourescence in situ hybridization (FISH), a series of interspecific hybrids, alien introgression lines, monosomic alien addition lines (MAALs), and chromosome segmental substitution lines (CSSLs) have been produced. Genes for resistance to brown planthopper (BPH), bacterial blight (BB), blast, tungro, acid sulfate soils, and iron toxicity have been introgressed from AA, BBCC, CC, CCDD, EE, and FF genomes into rice. Some of the introgressed genes (Bph10, Bph18, Xa21, Pi-9) introgressed from wild species have been mapped and used in marker-assisted selection (Xa21, Bph18). Some of the breeding lines with genes introgressed from wild species have been released as varieties. Opportunities exist in wild species germplasm for novel genetic variability for traits such as C₄ or C₃-C₄ intermediates. Once variability for C₄ traits is identified in Oryza, it should be possible to exploit the transfer of such traits following wide hybridization techniques already used successfully for the transfer of resistance to biotic and abiotic stresses and thus enhance the photosynthetic efficiency and yield of rice by modifying it from C₃ to C₄.

Rice is one of the most important cereal crops in the world. Rice production has been increasing at a steady pace since the adoption of modern high-yielding rice cultivars. However, demand for rice continues to increase and so does the number of rice consumers. It is estimated that we need to produce an extra 200 million tons of rice by 2025 to meet the demand. To achieve this objective, we have to breed rice varieties with higher yield potential. Several breeding strategies are being employed at IRRI for developing rice varieties with higher yield potential, such as empirical breeding, heterosis breeding, ideotype breeding, and wide hybridization. Modern approaches such as the development of C₄ rice for increasing the yield potential of rice are being investigated. To maximize the benefit of C₄-ness in rice, breeders should target associated yield component traits. Furthermore, rice breeders can play a major role in transferring C₄-ness (once available) from donor or engineered genotypes into commercially acceptable backgrounds.

The delivery of improved rice varieties with C₄-like yield advantages will require a precise modeling of the genome. Interestingly, the novelty of C₄ photosynthesis appears to have evolved from a key reservoir of genes present also in C₃ plants through gene duplication and neofunctionalization. Both specific patterns of gene expression and kinetic properties of the enzymes together with anatomical changes are observed in C₄ plants. Improved photosynthesis in rice can probably be achieved by engineering alleles involved in biochemical pathways and plant development. Key limiting steps in C₃ photosynthesis have also been envisaged as a possible strategy to lead to yield increases. From these results, I believe that a dual approach based on directed evolution and allele engineering could boost the yield potential of rice. Here, we discuss briefly some possible options from allele engineering to phenotype.

With the sequencing of the rice genome and pending sequences of several C₄ cereals (i.e., maize and sorghum), functional genomics analysis may elucidate the critical genetic differences contrasting C₄ and C₃ photosynthetic biochemical and anatomical systems. Recent genomics findings may also deepen our understanding of the evolution of the C₄ syndrome, pointing at the evolutionary processes involved and possibly the plasticity of the rice genome for conversion to C₄ photosynthesis. Bioinformatics systems and methodology serve as critical glue for cross-linking such genomic information into an interpretable framework.

The conference on "Supercharging the Rice Engine," held at the International Rice Research Institute in July 2006, discussed various ways of constructing C₄ rice, given its desirability for substantially increasing rice yields along with better use of water and nitrogen. It was agreed to set up a consortium to stimulate and coordinate research to produce C₄ rice. The organization of the consortium, stages of work, proof of concept, and the need for sustained funding are outlined. The two main pathways to C₄ rice are with Kranz anatomy, as in maize, or as a single-cell system. The features of each system are summarized, as currently understood, and questions for research are identified. Both pathways deserve attention and research on each will benefit the other; there is no clear single route to success at present. Other possibilities for improving photosynthesis are mentioned but only C₄ rice brings the whole package of high productivity plus better use of resources that is necessary to help alleviate hunger and poverty in countries dependent on rice.

Index.