Gustavo A. Slafer

Breeding for Yield Potential and Stress Adaptation in Cereals

The need to accelerate breeding for increased yield potential and better adaptation to drought and other abiotic stresses is an issue of increasing urgency. As the population continues to grow rapidly, the pressure on resources (mainly untouched land and water) is also increasing, and potential climate change poses further challenges. We discuss ways to improve the efficiency of crop breeding through a better physiological understanding by both conventional and molecular methods. Thus the review highlights the physiological basis of crop yield and its response to stresses, with special emphasis on drought. This is not just because physiology forms the basis of proper phenotyping, one of the pillars of breeding, but because a full understanding of physiology is also needed, for example, to design the traits targeted by molecular breeding approaches such as marker-assisted selection (MAS) or plant transformation or the way these traits are evaluated. Most of the information in this review deals with cereals, since they include the worlds main crops, however, examples from other crops are also included. Topics covered by the review include the conceptual framework for identifying secondary traits associated with yield potential and stress adaptation, and how to measure these secondary traits in practice. The second part of the review deals with the real role of molecular breeding for complex traits from a physiological perspective. This part examines current developments in MAS and quantitative trait loci (QTL) detection as well as plant transformation. Emphasis is placed on the current limitations of these molecular approaches to improving stress adaptation and yield potential. The essay ends by presenting some ideas regarding future avenues for crop breeding given the current and possible future challenges, and on a multidisciplinary approach where physiological knowledge and proper phenotyping play a major role.

Raising yield potential in wheat

Recent advances in crop research have the potential to accelerate genetic gains in wheat, especially if co-ordinated with a breeding perspective. For example, improving photosynthesis by exploiting natural variation in Rubiscos catalytic rate or adopting C(4) metabolism could raise the baseline for yield potential by 50% or more. However, spike fertility must also be improved to permit full utilization of photosynthetic capacity throughout the crop life cycle and this has several components. While larger radiation use efficiency will increase the total assimilates available for spike growth, thereby increasing the potential for grain number, an optimized phenological pattern will permit the maximum partitioning of the available assimilates to the spikes. Evidence for underutilized photosynthetic capacity during grain filling in elite material suggests unnecessary floret abortion. Therefore, a better understanding of its physiological and genetic basis, including possible signalling in response to photoperiod or growth-limiting resources, may permit floret abortion to be minimized for a more optimal source:sink balance. However, trade-offs in terms of the partitioning of assimilates to competing sinks during spike growth, to improve root anchorage and stem strength, may be necessary to prevent yield losses as a result of lodging. Breeding technologies that can be used to complement conventional approaches include wide crossing with members of the Triticeae tribe to broaden the wheat genepool, and physiological and molecular breeding strategically to combine complementary traits and to identify elite progeny more efficiently.

Raising yield potential of wheat. III. Optimizing partitioning to grain while maintaining lodging resistance

A substantial increase in grain yield potential is required, along with better use of water and fertilizer, to ensure food security and environmental protection in future decades. For improvements in photosynthetic capacity to result in additional wheat yield, extra assimilates must be partitioned to developing spikes and grains and/or potential grain weight increased to accommodate the extra assimilates. At the same time, improvement in dry matter partitioning to spikes should ensure that it does not increase stem or root lodging. It is therefore crucial that improvements in structural and reproductive aspects of growth accompany increases in photosynthesis to enhance the net agronomic benefits of genetic modifications. In this article, six complementary approaches are proposed, namely: (i) optimizing developmental pattern to maximize spike fertility and grain number, (ii) optimizing spike growth to maximize grain number and dry matter harvest index, (iii) improving spike fertility through desensitizing floret abortion to environmental cues, (iv) improving potential grain size and grain filling, and (v) improving lodging resistance. Since many of the traits tackled in these approaches interact strongly, an integrative modelling approach is also proposed, to (vi) identify any trade-offs between key traits, hence to define target ideotypes in quantitative terms. The potential for genetic dissection of key traits via quantitative trait loci analysis is discussed for the efficient deployment of existing variation in breeding programmes. These proposals should maximize returns in food production from investments in increased crop biomass by increasing spike fertility, grain number per unit area and harvest index whilst optimizing the trade-offs with potential grain weight and lodging resistance.

Achieving yield gains in wheat

Wheat provides 20% of calories and protein consumed by humans. Recent genetic gains are <1% per annum (p.a.), insufficient to meet future demand. The Wheat Yield Consortium brings expertise in photosynthesis, crop adaptation and genetics to a common breeding platform. Theory suggest radiation use efficiency (RUE) of wheat could be increased ~50%; strategies include modifying specificity, catalytic rate and regulation of Rubisco, up-regulating Calvin cycle enzymes, introducing chloroplast CO(2) concentrating mechanisms, optimizing light and N distribution of canopies while minimizing photoinhibition, and increasing spike photosynthesis. Maximum yield expression will also require dynamic optimization of source: sink so that dry matter partitioning to reproductive structures is not at the cost of the roots, stems and leaves needed to maintain physiological and structural integrity. Crop development should favour spike fertility to maximize harvest index so phenology must be tailored to different photoperiods, and sensitivity to unpredictable weather must be modulated to reduce conservative responses that reduce harvest index. Strategic crossing of complementary physiological traits will be augmented with wide crossing, while genome-wide selection and high throughput phenotyping and genotyping will increase efficiency of progeny screening. To ensure investment in breeding achieves agronomic impact, sustainable crop management must also be promoted through crop improvement networks.

Changes in yield and yield stability in wheat during the 20th century

Abstract An analysis of trends in yield and yield stability throughout the century was made for 21 countries (Algeria, Argentina, Australia, Canada, Chile, Egypt, France, Germany, India, Italy, Japan, Mexico, New Zealand, South Africa, Spain, Sweden, Tunisia, UK, Uruguay, USA and the former USSR). Regressions (linear, bi-linear or tri-linear fitted with an optimisation technique) were used to evaluate the trends in yield during the century. Residuals and relative residuals of these regressions were used to evaluate in absolute and relative terms, respectively, trends in yield stability. Countries varied greatly in their yields and yield gains as well as in changes in harvested area. But almost all of them showed a remarkable lack of yield gain during the initial 3 to 5 decades of this century, followed by noticeable increases in yield. Yield trends for relatively young agricultural wheat-exporting countries, such as Argentina, Australia, Canada and USA, reveal an important breakpoint ca. 2 decades earlier than European countries with longer tradition in wheat production. In addition, yield gains in many countries have apparently been levelling off during the last decade. Trends in yield residuals during the present century revealed a decrease in yield stability in 14 of the 21 countries analysed, but the increase in yield residuals was relatively small (≤0.3 Mg ha−1) compared with increases in yield. Therefore, relative yield residuals indicated that yield stability, as a percentage of yield, increased or at least did not change for most of the analysed countries. Moreover, it is suggested that wheat production systems have been, in general, highly successful in increasing yield while maintaining or increasing relative yield stability with respect to that existing at the beginning of the century. Finally, no relationship was found between variations in yield stability, both in absolute and relative terms, and the increase in yield comparing the present values and those at the beginning of the century.

Physiological attributes related to the generation of grain yield in bread wheat cultivars released at different eras

Abstract Several authors have concluded that the main physiological changes produced by the genetic improvement of bread wheat grain yield were an increased harvest index associated with an increased number of grains. However, the underlying causes of these changes have been less considered, especially regarding crop-based measurements. Two irrigated field experiments were carried out to study the accumulation and partitioning of the dry matter and the associated generation of number of grains per m 2 and grain yield of three bread wheat cultivars released in 1920 (Klein Favorito), 1940 (Eureka FCS) and 1980 (Buck Pucara) in Argentina. Grain yield of the modern cultivar was higher than that of the intermediate and old cultivar (755, 585 and 460 g m −2 ; mean of the two years, respectively). These yields were closely associated with the number of grains per m 2 (30 150, 22 115 and 16 665; mean of the two years, respectively), confirming previous results. These differences in number of grains per m 2 were more associated with differences in number of grains per spikelet than with differences in any other component. At anthesis the cultivars had accumulated similar above-ground biomass, but the proportion partitioned to spikes was higher as the year of release increased. The number of fertile florets produced by Buck Pucara was greater than those produced by other cultivars, which did not differ in this trait. The grain set percentage was the cause of the differences in number of grains per m 2 between Eureka FCS and Klein Favorito. The number of fertile florets was related to spike dry weight at heading rather than at anthesis, emphasizing the period of floret primordia death as crucial for the generation of a higher number of fertile florets. The differences in the fate of florets among cultivars were related to spike growth during a short period (approximately 25 days) before anthesis. Spikes of Buck Pucara grew faster than those of the other cultivars from approximately 15 days before anthesis, while the spikes of Eureka FCS grew faster than those of Klein Favorito only in the final period of spike growth, between heading and anthesis. These differences will still more evident in the spike to stem dry weight ratio. The major reason for the increased ability of the modern cultivar to partitioning more biomass to spikes was that its spikes began their growth earlier when the stems were lighter than those of the old cultivar, leading to differences in their competitive ability for obtaining photoassimilates.

Source—sink relationships and grain mass at different positions within the spike in wheat

Abstract Studies manipulating the source-sink relationships during postanthesis in wheat have shown inconsistent results, concluding that wheat yield could be either sink- or source-limited. Most of the studies reported data of grain mass as an average of all grains preventing the recognition of different sensitivities to the availability of assimilates of grains placed at different positions within the spike. The response in growth and final mass of grains from the base, middle and apical spikelets to increases in assimilate availability was studied by modifing source-sink relationships during the grain growth period of wheat. In addition, a quantitative analysis of the responses in grain mass to manipulations in the source—sink relationships was attempted. Two wheat cultivars, ‘Condor’ (semi-dwarf) and ‘Thatcher’ (standard-height) were grown in 1992 at the experimental station of The University of Melbourne under two different photoperiods between emergence and heading at field conditions to have two different environments during grain-filling period (early and late). Twenty main shoots in each plot were chosen and detillered at anthesis. Ten days after anthesis, the source—sink relationship was altered in half of them by removing all the spikelets along one side of the spike. Grain growth (averaged over all grains per spike) in both cultivars and under the two grain-filling periods appeared to be largely insensitive to the increase in the availability of assimilates. Neither grain growth rate nor the affective period of grain-filling was significantly affected by potentially doubling the source capacity per grain by trimming. Similarly, there was no effect of trimming on grain mass and its components when grains from apical, central or basal spikelets were analysed separately for the two grain-filling periods. This lack of effect of changes in source—sink relationships on grain growth would imply that the assimilate availability in control plants in this study was sufficient to fully satisfy grain growth requirements. These results were similar when grains of different size (those at different position within the spike) were individually considered, suggesting that there is no association between the size of the grains in the control treatment and their response to changes in the assimilate availability. Comparing these results and many others from the literature, in which there was not agreement about the fact that grain yield was source- or sink-limited, and using a quantitative approach, we suggest that during post-anthesis period, grain yield of wheat is either sink-limited or co-limited by both source and sink but never source-limited.

Floret development in near isogenic wheat lines differing in plant height

Abstract The effects of Rht1 and Rht2 alleles on the dynamics of floret development in isogenic lines (dwarf, DD; semi-dwarf, SD and standard height, SH) of spring wheat were investigated. Studies were conducted on wheat grown in the field in each of 4 years and where water and nutrients were non-limiting. The number of grains per spike was significantly greater in the lines with Rht alleles than in the SH lines. Grain number for each line was such that DD>SD>SH. Grains per spike varied with the number of grains per spikelet rather than number of spikelets per spike. Grains per spikelet in turn varied with the number of fertile florets at anthesis. Florets were considered fertile when male and female reproductive organs had developed green anthers and bifidum stigma, respectively. The dwarfing genes had no effect on the percentage of fertile florets setting grain. Increased number of fertile florets per spikelet due to the presence of Rht1 and Rht2 alleles in the genome was a consequence of the higher number of relatively distal primordia, to progress to the stage of fertile floret at anthesis in the DD and SD than in the SH lines. This ability to allow that a greater proportion of distal florets maintain a normal rate of development was related to the fact that Rht alleles produced a more favourable assimilate partitioning to the spike during the pre-anthesis period associated with the reduction in stem growth imposed by Rht alleles. This allowed a higher proportion of the later-initiated floret primordia to produce fertile florets at anthesis.

Shading effects on the yield of an Argentinian wheat cultivar

Shading treatments of 50% of the incident radiation were applied to the semidwarf wheat cultivar Leones INTA before and after anthesis in two field experiments in Argentina in 1987 and 1988. The treatments reduced biological (above-ground dry matter) yield, grain yield and number of grains/m2. Number of grains/m2 was closely and linearly correlated with ear dry weight at anthesis and with the photothermal quotient, calculated from 20 days before to 10 days after anthesis. Grain yield was sink limited, and the shading treatments reduced sink strength. The contribution of preanthesis assimilates to grain yield was smaller in the shaded crops than in the unshaded controls; in unshaded crops, almost 40% of grain yield was contributed by preanthesis assimilates whilst in preanthesis shaded crops this contribution was negligible. The proportion of preanthesis assimilates contributed to the grain was closely related to the decrease in stem dry weight during grain filling. The effects of shading on main stems and tillers were the same.

PAPER PRESENTED AT INTERNATIONAL WORKSHOP ON INCREASING WHEAT YIELD POTENTIAL, CIMMYT, OBREGON, MEXICO, 20–24 MARCH 2006 Sink limitations to yield in wheat: how could it be reduced?

Ponencia presentada al International Workshop on Increasing Wheat Yield Potential, CIMMYT, Obregon, Mexic, del 20 al 24 de marc de 2006.