50 I. I. Introduction The global demand for primary foodstuffs is expected to increase by 60-110% (Tilman et al., ; Alexandratos & Bruinsma, ; Ray et al., ; Long et al., ; Tilman & Clark, ) as a consequence of increased urbanization and the predicted rise in global population from c. 7.3 billion today to 9.7 billion by 2050 (United Nations, ). The projected increase is particularly high for Africa, where the current population of 1.2 billion is expected to reach 5.6 billion by 2100 (United Nations, ). Overlying this scenario are changes in global climate that will affect crop productivity, primarily increases in temperature, increases in the incidence of drought, rising atmospheric CO2 and elevated surface ozone. It has been estimated that climate change by mid-century will depress aggregated cassava production across sub-Saharan Africa by 18% in the absence of adaptation (Schlenker & Lobell, ; Roudier et al., ). This is compounded by loss and degradation of agricultural land and exhaustion of water resources for irrigation (Godfray & Garnett, ). Cassava (Manihot esculenta Crantz) is a perennial woody shrub of the Euphorbiaceae family and is considered a staple food of more than a billion people in c. 105 countries (Chetty et al., ). It is typically cultivated by families for their own consumption on small plots of land, although in Asia and some regions of Latin America it is also grown commercially and on large fields (Nassar & Ortiz, ). In Africa, cassava is the second most important source of calories. Its tuberous roots provide over a quarter of the daily calorie consumption in some African countries, such as Angola and Mozambique, but for the rural poor it represents a much larger proportion of daily calories (FAO, ). In addition to the tuberous roots, the main harvested product, cassava leaves are consumed as a vegetable in at least 60% of the countries in sub-Saharan Africa, providing an important source of proteins, vitamins and micronutrients (Latif & Muller, ). Cassava leaves are also used as a protein supplement for livestock (Lukuyu et al., ). Between 2000 and 2013, the amount of cassava harvested in the world increased by c. 60% (Howeler et al., ). Currently, the largest producer of cassava in the world is Nigeria, followed by Thailand, Indonesia, Brazil and the Democratic Republic of the Congo. Together, these countries account for 52.8% of world production (Table ). Moreover, it is projected that cassava may become an important replacement for crops that are expected to be more vulnerable to climate change, especially in eastern Africa (Schlenker & Lobell, ; Jarvis et al., ). This is in part a result of the fact that rising CO2 concentration ([CO2]) will have a more positive effect on cassava as a C3 crop than it will on the major C4 crops sorghum (Sorghum bicolor), maize (Zea mays) and millets (Pennisetum glaucum) (Roudier et al., ; Rosenthal et al., ). However, there is great uncertainty in these projections, which range from near complete crop loss to large increases in production (Roudier et al., ). This, though, emphasizes the need for the global society to insure against this uncertain future by mobilizing research and development effort to provide germplasm with increased productivity and sustainability potential, under conditions of climate change. Despite its importance, yield improvement in cassava has received relatively little attention or investment (El-Sharkawy, ). This is vividly demonstrated by the fact that between 1961 and 2014 average cassava yields per unit land area did not increase in Nigeria, the largest global producer. Over the same period, maize yields per unit land area in Nigeria increased by 129%, approaching the yield increase of 174% achieved by the world's largest producer of maize, the USA (Fig. ). The efforts to improve tuberous root yields made by breeding programs throughout the world have prioritized drought tolerance, cyanogenic content, low soil fertility conditions, and resistance to a wide range of diseases (El-Sharkawy, ). Concomitantly, efforts have also been centered on improving cassava nutritional quality (Montagnac et al., ; Gonzalez et al., ; Sayre et al., ; Ceballos et al., ) and agronomic practices (CIAT, ). Such efforts in Asia, for instance, have increased cassava yields at a rate of 0.138 t ha-1 yr-1 (on a dry weight basis) between 2004 and 2014. However, in sub-Saharan Africa, where cassava is essential in supplying calories to a large proportion of the population, yields have actually been declining by 0.024 t ha-1 yr-1 (Fig. ). The average yields currently achieved by African farmers are just 2.51 t ha-1 on a dry weight basis, which is lower than the world average of 3.35 t ha-1 and 2.5 times lower than yields attained in Asia (Fig. ). Sub-Saharan Africa is predicted to see the largest population growth of all world regions, 123% by 2050 (United Nations, ). Cassava is an important crop for subsistence farmers in this region, and, as already mentioned, it is a crucial and favored source of calories by this population, as well as being a cash crop in years of surplus. With limited availability of additional sources of calories, increasing the yield per unit land area of cassava will be critical. Genetic improvement of resource use efficiency, particularly the conversion of available sunlight into biomass, is one opportunity. This review assembles knowledge of the underlying physiology determining yield potential in cassava, and uses this to suggest ways to increase its genetic yield potential. 50 II. II. How might genetic yield potential be increased in cassava? Improvement of the harvest index, or the proportion of total biomass partitioned into the harvested component, was a key factor driving increased yields in the Green Revolution. Because of this improvement, cassava breeding over the past 30-40 yr has understandably focused upon increasing the harvest index, which proved a successful strategy in increasing rice (Oryza sativa) and wheat (Triticum aestivum) yields (Ceballos et al., ). However, because this strategy increases the proportion of total plant biomass partitioned to the harvested product, the total plant biomass places a limit on the absolute increase in yield that may be achieved. What are the prospects for genetically increasing the total biomass per hectare? The total biomass produced by a crop results from the integral of photosynthetic assimilation over the growing season less all respiratory losses. It depends on the efficiencies with which the crop intercepts light and converts that into biomass over the course of the growing season (Monteith & Moss, ; Long et al., ,b; Zhu et al., , ; Parry et al., ; Reynolds et al., ; Long & Zhu, ). Representation of these efficiencies in the form of an equation provides one method to quantitatively assess the opportunity for improvement of yield potential (Monteith & Moss, ; Long et al., ,b). Although almost 50 yr old, this simple, but physiologically well-founded approach remains invaluable. It has revealed similarities across photosynthetic types, allowed evaluation of the effects of atmospheric and climate change on crops, enabled analysis of efficiencies in remote sensing and has shown how genetic improvements of yield potential have been achieved (Beadle & Long, ; Zhu et al., ). This equation states that yield potential is the product of total incident photosynthetically active solar radiation (PAR) over the growing season (St), PAR interception efficiency (i), PAR conversion efficiency (c) and partitioning efficiency or harvest index (p). i is affected by canopy size, architecture, duration and speed of ground coverage after planting; c is defined by the amount of intercepted PAR that is converted into biomass; and p is the proportion of biomass that is partitioned to the harvested plant organ, in the case of cassava, the starchy tuberous roots. Using this equation, the steady increase in soybean (Glycine max) yield of 26.5 kg ha-1 yr-1 over the past 80 yr of breeding was shown to be attributable to an improvement in all three efficiencies. However, while i and p for soybean appear now to be close to their maximum theoretical values, c falls far short of its theoretical maximum and has been improved by far the least of the three. Similar conclusions may be drawn for wheat and rice, where a plateauing of improvement in i and p, as well as failure to improve c, coincides with a stagnation of yield increase at the present time. (Long & Ort, ; Ray et al., ; Long, ; Long et al., ). For cassava, the p achieved by cultivars (Table ) is close to that of the crops that have attracted the most attention and investment in breeding, such as maize, rice, wheat and soybean (Long et al., ,b; Fischer & Edmeades, ; Fischer, ; Koester et al., ; Long & Zhu, ). For grain and seed crops, p is considered to have a theoretical maximum of c. 65%, as at harvest some biomass must remain in the stems and floral structures that contain the seed (Zhu et al., ). In these crops, however, the stem and floral structure are necessary to hold the harvested plant part. As a root crop, cassava, in theory, does not need any part of the shoot to persist through to harvest, and hence a higher theoretical p is conceivable. However, in practice, cassava propagation is predominantly from stem cuttings, which means that a significant amount of live mass must remain in stems under current methods of propagation (El-Sharkawy, ). Although p in cassava, like other crops, is probably close to the maximum that could be achieved while still providing stems for propagation, this is not true for i. Calculated i values from four cassava cultivars were 52.3-64.1% (Table ), well below those of modern cultivars of major grain and seed crops, as well as the theoretical limit of 90% (Long et al., ,b; Koester et al., ). Even for cultivars from Asia, the continent where the greatest increases in yield per unit land area have been achieved (Fig. ), the cassava season-long i is c. 64% (Leepipatpaiboon et al., ) while for soybean, for instance, i may approach 90% (Long et al., ,b). Thus, although i seems to be near to its theoretical maximum in the crops that have been most intensively bred (Long & Zhu, ), in cassava there is still considerable room for improvement. Based on Table , selection or engineering of forms that approach the theoretical 90% could give a 55% increase in yield potential. The conversion efficiency ([epsi]c), which is determined by canopy photosynthetic carbon uptake less whole-plant respiration, appears to present the greatest opportunity for improving cassava yield potential. In terms of mass per unit intercepted energy, [epsi]c ranges from 0.69 to 0.94 g MJ-1, corresponding to an average [epsi]c in terms of energy transduction of intercepted PAR into biomass of just 1.4%. That is only one-seventh of the theoretical efficiency of C3 photosynthesis (Table ; Fig. ; Zhu et al., ). Similar values were reported previously (Beadle & Long, ; Pellet & El-Sharkawy, ). By contrast, the most productive soybean cultivars can reach an [epsi]c of 2.9-4.3% (Koester et al., ). However, even these values are well below the theoretical maximum of [epsi]c predicted for C3 plants of 9.4% (Zhu et al., ) (Fig. ). It should also be noted that this theoretical maximum assumes typical photorespiratory rates, which, if successfully reduced, would raise this theoretical maximum. These low values of [epsi]c by comparison both to other crops and to theoretical values highlight the lack of incorporation of photosynthetic traits in breeding and current engineering programs. Although El-Sharkawy & De Tafur provide a compelling argument for the inclusion of photosynthetic efficiency in breeding traits, the available data for landraces and cultivars suggest that there has been no improvement in photosynthetic rate through breeding (Table ). This might be partially explained by the fact that there has been little research into cassava photosynthesis compared with the major field and glasshouse crops of the developed world. That improving photosynthesis in cassava will increase yield is vividly demonstrated by the open-air [CO2] enrichment study by Rosenthal et al. . Here, an average season-long 30% increase in leaf photosynthetic rates under elevated [CO2] resulted in a 104% increase in tuberous dry mass. This greatly exceeds the c. 15% increase in the yields of wheat, rice and soybean observed with similar enhancements of photosynthesis to that seen in cassava, under open-air elevation of [CO2] (Long et al., ,b; Ainsworth et al., ,b). In reviewing the growth of a wide range of crops under elevated [CO2], it was observed that, in general, root crops showed a greater stimulation of yield. This may reflect the more indefinite nature of the size and number of roots or tubers in these crops relative to most grain and seed crops. Genetically increasing photosynthetic efficiency might therefore be expected to provide even larger benefits in cassava than in our major grain and seed crops. 52 III. III. Modifying the cassava canopy to achieve higher i The i of a given crop is defined by the size and architecture of the canopy as well as its duration and speed of closure. In cassava, many parameters related to the canopy, such as leaf area index (LAI), leaf retention and branching habit, have already been shown to positively correlate with yield (Okogbenin & Fregene, ; El-Sharkawy & De Tafur, ; Lahai, ). The cassava canopy starts to develop c. 15 d after planting of the stem section propagules and reaches maximum light interception at c. 4-5 months (Alves, ; El-Sharkawy, ). This would appear slow compared with maize and soybean crops, which can achieve closed canopies within 4 wk (Singer et al., ). Yet crops sown from seed have a very small reserve, which clearly limits the speed at which closure can be achieved. Cassava stem segments should in theory represent a much larger reserve which should power rapid development of leaves and canopy closure. This suggests that there has been little selection for accumulation of reserves in the stem, which could greatly improve the speed of canopy closure in cassava. [epsi]i is determined largely by LAI and leaf angle. [epsi]i shows a rectangular hyperbolic response to LAI, with [delta][epsi]i:[delta]LAI increasing as the average leaf angle becomes more horizontal (Drewry et al., ). Typically, cassava LAI peaks at c. 5 months when senescence of lower leaves begins to counteract further leaf production at the top of the canopy. LAI typically begins to decline at 9 months when senescence outpaces new leaf production and [epsi]i declines accordingly (Fig. ) (El-Sharkawy et al., 1992a; Pellet & El-Sharkawy, ; El-Sharkawy & Cadavid, ; El-Sharkawy & De Tafur, ). Breeding strategies have selected cultivars that have long-lived leaves and individual leaves with an increased leaf area (Lenis et al., ; Lebot, ). Cassava shows simultaneous shoot and tuberous root development in which photoassimilates are partitioned between leaves and tuberous root growth (Fukai et al., ; Alves, ). Tuberous root development starts c. 2 months after planting, before maximum investment in leaf biomass (Fig. ). Although tuberous roots are bulking throughout this period, shoot development is dominant and appears to have priority over root growth (Lian & Cock, ). In other words, photoassimilates are not preferentially partitioned to tuberous roots until shoot growth nears completion, which usually occurs c. 6 months after planting (Fig. ). This preference may reflect a delicate balance between shoot and tuber growth, imposing a limit on canopy development for maximum yields. Excessively large canopies may actually reduce yields in cassava (Lahai, ). This could result from the fact that the lower leaves might be starved of light to the extent that they respire more carbon than they assimilate in photosynthesis, and represent investment of resource that could have been used in tuber growth. Conversely, early bulking of tuberous roots will slow canopy development and lower [epsi]i. Selection of genotypes in which bulking does not begin until canopy closure has occurred, followed by a switch to sink dominance by the tuberous roots, could greatly increase yield. Dominance of the shoot would appear to result in an over-investment in leaves. The comparison between improved cultivars and landraces exemplifies this dominance interaction. Improved cultivars have higher tuberous root yields than landraces, yet the LAI of improved cultivars is lower for most of the growing season (Table ). This suggests that breeders have inadvertently selected for this trait. Maximum yield appears to occur with an LAI between 2.5 and 3.5 (Cock et al., ; Ramanujam, ; Lebot, ), and a leaf longevity of c. 100 d has been suggested to be optimal for maximizing yield (Cock et al., ). Canopy architecture in cassava varies considerably with cultivar, ranging from nonbranching types, also called erect types, to bush or highly branched types (Ekanayake et al., ). Branching genotypes usually form a better canopy that can intercept more light than nonbranching genotypes, resulting in higher tuberous root yields. However, if branching occurs very early during development, resulting in multiple shoot sinks, competition for photoassimilates between shoot and tuberous root development can reduce the final yield. Indeed, computational simulations suggest that late branching is ideal for maximizing yield (Cock et al., ). In addition to the importance of canopy architecture for photoassimilate partitioning between the shoot and tuberous roots, the canopy is also important in terms of agronomic practices (CIAT, , ). For instance, more than one-third of cassava world-wide is intercropped (Lebot, ), where an unbranched stem may reduce shading of the adjacent crop. In monocultures, however, improvement may be gained by increasing leaf angles toward the vertical and by selection for lighter green leaves in the upper canopy (Long et al., ,b; Drewry et al., ; Ort et al., ). This would allow a more effective distribution of light between upper and lower leaves, increasing net canopy photosynthesis and in turn yield. Lighter green leaves would also serve to cool the canopy, relative to the current dark green leaves, as a strategy to deal with rising temperatures (Drewry et al., ). The strategy of increasing leaf angle was demonstrated to be effective in other crops such as maize, wheat and rice (Sakamoto & Matsuoka, ; Isidro et al., ), especially as it allows an increase in plant density which would again allow faster canopy closure. Although the ideal characteristics for a cassava canopy have been simulated using computational models (Cock et al., ; Fukai & Hammer, ; Gutierrez et al., ; Gijzen et al., ; Matthews & Hunt, ; Gabriel et al., ) with some of those characteristics incorporated into breeding programs, the i of modern cassava cultivars is still far below the theoretical maximum (Table ; Fig. ). This shortcoming might be because none of these models have succeeded in incorporating an adequate solution to describing the dynamic partitioning of biomass (Gray, ), a crucial parameter for cassava. Further, none of these models have considered the link between canopy, leaf-level photosynthesis and productivity. Thus, the inclusion of mechanistic processes such as leaf photosynthesis, stomatal closure, and energy balance in the next generation of models will be vital to identify the best character to select for in improving yield. This will provide a physiological and morphological basis to link with emerging information on gene function and their associated gene networks affecting canopy architecture and leaf properties. 54 IV. IV. Increasing [epsi]c in cassava through photosynthesis Carbon assimilation through the photosynthetic process is, of course, crucial for cassava tuberous root production. As mentioned in the previous section, shoots have preference over tuberous root growth in the competition for photoassimilates, so affecting final yield. However, under conditions of increased canopy photosynthesis, it appears that the excess is allocated to the tuberous roots (Rosenthal et al., ). Increasing crop photosynthesis may therefore result in a larger than expected increase in yield, given that a 30% increase in photosynthesis at the leaf level resulted in a more than three-fold increase in the harvested yield of tuberous roots. A very valuable approach to further increase cassava yields would therefore be to increase the efficiency of photosynthesis. Indeed, this has already been recommended by other authors (Pellet & El-Sharkawy, ; De Tafur et al., 1997b; Flood et al., ), but, to date, genetic improvement of photosynthesis appears a rather unexplored field in cassava research. Studies of photosynthesis in cassava are limited. Most of the published research on cassava photosynthesis comes from the International Center for Tropical Agriculture (CIAT), in Colombia. Therefore, the available knowledge about photosynthesis in cassava is mostly limited to Latin American cultivars. Some studies on these cultivars have suggested that cassava uses a C3-C4 intermediate form of photosynthesis on account of high leaf photosynthetic rates, low apparent rates of photorespiration, a chlorenchymatous bundle sheath and a high photosynthetic nitrogen (N) use efficiency (El-Sharkawy & Cock, ; El-Sharkawy, , ). Parallel work feeding 14CO2 to leaves of cassava showed an apparent intermediate pattern of initial C4 and C3 products (Cock et al., ). However, a subsequent detailed analysis of first products failed to show such an intermediate pattern, but rather a very typical C3 pattern. This later study also showed that the photosynthetic CO2 compensation points of 10 different cultivars were between 55 and 62 µmol mol-1, typical of C3 species, whereas an intermediate would be expected to show a value of c. 25 µmol mol-1 (Edwards et al., ). Carbon 13 isotope fractionation ([delta]13C) in cassava is also identical to that of C3 species, ranging from -23[per thousand] to -26[per thousand] (Burns et al., ), compared with -12[per thousand] and -16[per thousand] found in C4 plants (O'Leary, ). Moreover, although cassava has a chlorenchymatous bundle sheath, unlike C4 species it is not surrounded by mesophyll cells (Edwards et al., ). In total, these findings show cassava to be a typical C3 species (Edwards et al., ; Angelov et al., ; Gleadow et al., ; Rosenthal et al., ). The compilation of data from several studies in which photosynthetic rates were reported shows that the average net photosynthetic rate is significantly smaller than the highest rates observed, calling into question the concept that cassava has unusually high rates for a C3 species (Table ). In optimal growing conditions, the highest reported photosynthetic rate for field-grown improved cassava cultivars was 50 µmol m-2 s-1 at a photon flux of over 1800 µmol m-2 s-1. However, the seasonal average is about half of this value (Table ) (El-Sharkawy, ) and, although some high values may be observed, photosynthetic rates varied little in the field over the growing season (Bhagsari, ; Pellet & El-Sharkawy, ; De Tafur et al., 1997a; El-Sharkawy & De Tafur, ; Rosenthal et al., ). The average in vivo capacity for Rubisco carboxylation (Vc,max) is 133.3 µmol m-2 s-1 and the maximum rate of whole chain electron transport (Jmax) is 225.95 µmol m-2 s-1 for improved cassava cultivars. These are comparable to the Vc,max and Jmax observed for rice in the field (Borjigidai et al., ), but somewhat higher than averages observed for other C3 crops, shrubs, trees, grasses, and legumes (Ainsworth & Long, ; Ainsworth & Rogers, ). The significant variation in these parameters within cassava suggests an opportunity for selection. Several genetic strategies have been proposed for crop and model plants to increase [epsi]c and yield through increasing photosynthesis, and a limited number of these have been realized (Long et al., ). One of the most extensively explored strategies has been the optimization of enzyme activity within the Calvin cycle, in particular the up-regulation of sedoheptulose-1,7-bisphosphatase (SBPase) and fructose-1,6-bisphosphate aldolase (Raines, ; Rosenthal et al., ; Simkin et al., ) as predicted by computer simulation (Zhu et al., ). In wheat, genetic variation in SBPase expression correlating with leaf photosynthetic rates has been shown, suggesting that conventional breeding in which high expression of SBPase is selected would also increase productivity (Driever et al., ). The catalytic properties of the enzyme ribulose-1,5-bisphoshate carboxylase/oxygenase (Rubisco) are a key factor determining light-saturated photosynthetic rates of C3 crops (Portis & Parry, ). Variation within and between species suggests an opportunity to engineer or select for improved kinetic properties that would improve canopy photosynthesis without requiring more protein or nitrogen (Zhu et al., 2004b). Under global change-driven conditions of elevated [CO2] and elevated temperature, efficiency gains could be achieved by altering the balance between the capacity for regeneration of ribulose-1,5-bisphoshate (RubP) and the amount of Rubisco (Kromdijk & Long, ). This is particularly relevant to cassava, given the higher temperature conditions of the tropics. Thus, exploring variation in the kinetic properties of Rubisco between cassava cultivars could be of particular value (Galmes et al., ; Carmo-Silva et al., ). Synthetic photorespiratory bypass systems engineered into the chloroplast and designed to decrease CO2 losses, have been shown to effectively increase photosynthesis and production in model species (Kebeish et al., ; Peterhansel et al., ). Consideration of stochiometries shows that synthetic photorespiratory bypasses will reduce the energetic costs and increase [CO2] within the plastid, so also serving to decrease oxygenation and hence photorespiration (Xin et al., ). Furthermore, the increase in plastid [CO2] will increase the temperature optimum of photosynthesis, and increase water use efficiency (Kromdijk & Long, ). This approach might be especially relevant in cassava in view of the tropical conditions under which it is cultivated, while serving as a means to counteract the impacts of increasing water vapor pressure deficit with climate change (Lobell et al., ; Ort & Long, ). In high light, generation of a trans-thylakoid pH gradient and de-epoxidation of the xanthophyll violaxanthin to zeaxanthin are associated with dissipation of excess excitation energy as heat, termed nonphotochemical quenching (NPQ). This protects the photosynthetic apparatus against the generation of destructive oxidizing radicals (Long et al., ). However, on transfer to shade it takes a considerable time, many minutes, for these processes to relax. As a result, even though light is now limiting, a large proportion of the absorbed light energy continues to be dissipated as heat rather than being used to drive CO2 assimilation. The diurnal course of the sun on a clear day transfers leaves below the canopy top into and out of shade. This change occurs in a second at the level of individual chloroplasts. Using ray tracing, computational analysis showed that this slow recovery at the canopy level could cost 30% of potential carbon assimilation over the course of a day (Zhu et al., 2004a). The loss would be even greater under intermittent cloud. Selecting or engineering traits for faster relaxation could therefore considerably improve canopy photosynthetic efficiency and [epsi]c. As stomatal conductance adjusts slowly, on transition to shade, this would also significantly increase crop water use efficiency. 55 V. V. Does cassava have the sink capacity for an increased influx of photoassimilates? Limited sink capacity can feed back on any photosynthetic enhancements. Thus, efforts to increase conversion efficiency through improved photosynthetic rates in cassava could be unsuccessful without sufficient sink capacity. Pellet & El-Sharkawy and Rosenthal et al. found that individual tuberous roots have limited sink capacity, but this is offset by the initiation of additional tuberous roots. Earlier work has suggested that genotypes in which fewer than nine tuberous roots form are sink limited (Cock et al., ). This finding has driven interest in analyzing the genes, gene networks and gene products that control tuberous root initiation and bulking. Improving the sink capacity of cassava to increase yield may depend on these molecular targets. Mitprasat et al. , for instance, found a down-regulation of a glyceraldehyde-3-phosphate dehydrogenase with a concomitant up-regulation of a UDP-glucose pyrophosphorylase in cassava leaves after 8 wk, corresponding to initiation of tuberous root development or bulking. The authors hypothesized that changes in these two enzymes will favor sucrose synthesis for sink supply. They also observed a decrease in an antioxidant enzyme from weeks 4 to 7 after planting, and proposed that reactive oxygen species (ROS) formed by this enzyme may be functioning as a signaling molecule for tuber growth regulation through gibberellic acid. Li et al. found a series of differentially expressed genes in cassava roots between 2 and 4 months after planting by using a cDNA microarray. These root-specific genes might be responsible for initiating the root bulking process, given associated changes in transcripts involved in signal transduction, protein metabolism, starch and sucrose metabolism, and glycolysis-related processes. In addition to sink capacity, sucrose and glucose concentrations in the leaf play a large role in the regulation of expression of genes coding for the proteins of the photosynthetic apparatus. Sucrose and glucose accumulation in source leaves enhance the expression of genes involved in carbon storage and utilization, and cause down-regulation of some key genes coding for the photosynthetic apparatus, including Rubisco (Cho et al., ; Kunz et al., ). This is clearly illustrated when the leaf petiole is heat-girdled, preventing export and causing a large accumulation of starch and soluble carbohydrates, in turn rapidly down-regulating expression of genes encoding enzymes of photosynthetic carbon metabolism, N metabolism and chlorophyll synthesis. Simultaneously, expression of genes encoding enzymes involved in the tricarboxylic acid (TCA) cycle, mitochondrial electron transport, and flavonoid biosynthesis were up-regulated. Thus, increasing photosynthetic rates in cassava could also increase the sucrose negative feedback loop unless there is sufficient sink demand and transport capacity to remove the additional sucrose formed in the leaves (Zhang et al., ). Strategies to increase sugar transport and reduce carbohydrate accumulation in leaves may involve increasing the expression of sucrose transporters (Ainsworth & Bush, ) and the recently discovered SWEET (Sugars Will Eventually be Exported Transporter) transporters (Chen et al., ) in sink tissues to optimize sugar flux and increase phloem-loading capacity. Another approach could be to increase capacity for starch formation in leaves. Modeling of leaf photosynthetic carbon metabolism has shown that up-regulation of starch synthesis would also support greater rates of light-saturated CO2 assimilation (Zhu et al., ). This would allow increased export over the dark period, to make more efficient use of phloem capacity. In cassava, the overexpression of AGPase in the tuberous roots increased the amount of starch accumulated in this organ. This modification is also likely to lessen feedback inhibition of photosynthesis by decreasing the risk of accumulation of nonstructural carbohydrates in the leaves (Ihemere et al., ). The large increases in tuber yield seen in the one open-air elevated [CO2] experiment so far conducted may suggest that sink limitation is not a barrier (Rosenthal et al., ). However, this experiment concerned one cassava clone over a relatively short growing season. It will be important to establish with a wider range of open-air [CO2] elevations, ideally within the regions where the crop is normally produced, whether this finding applies more broadly to cassava. If it does, then it suggests photosynthetic improvement would be of great value in increasing yield potential. 57 VI. VI. Environmental stress effects on photosynthesis and development Environmental stresses will increase as global climate change unfolds, particularly with respect to temperature and soil moisture in the tropics (IPCC, ). In addition, to meet further demand, cul
View source
