Neglecting legumes has compromised human health and sustainable food production

The United Nations declared 2016 as the International Year of Pulses (grain legumes) under the banner ‘nutritious seeds for a sustainable future’. A second green revolution is required to ensure food and nutritional security in the face of global climate change. Grain legumes provide an unparalleled solution to this problem because of their inherent capacity for symbiotic atmospheric nitrogen fixation, which provides economically sustainable advantages for farming. In addition, a legume-rich diet has health benefits for humans and livestock alike. However, grain legumes form only a minor part of most current human diets, and legume crops are greatly under-used. Food security and soil fertility could be significantly improved by greater grain legume usage and increased improvement of a range of grain legumes. The current lack of coordinated focus on grain legumes has compromised human health, nutritional security and sustainable food production.

Year 19651970197519801985199019952000Year 1965197019751980198519901995  legume production increases have mostly been due to increases in the land area planted (Fig. 2). For soybean and the major grain legumes such as chickpea, groundnut and lentil, yield has increased proportionately with land area planted. Year-on-year increases in soybean yields are slowing while area planted is increasing, suggesting that more marginal land is being planted and improvements in genetic potential are not keeping pace. By contrast, yields of faba bean and peas are increasing while the area harvested is decreasing (Fig. 2), resulting in no net production increases in these two crops.
Intercropping and rotation of grain legumes with cereals or other non-leguminous crops have many benefits, such as enhanced yield, increased nitrogen-use efficiency (NUE, calculated by dividing mass grain dry yield by mass nitrogen fertilizer applied: G W /N S ), reduced occurrence of disease and, in some cases, improved access to other essential elements such as phosphorus 3 . The nitrogen-fixing ability of legumes affords complementarity through natural soil fertilization. Grain legumes favour reduced greenhouse gas emissions in agricultural cropping systems -for example, emissions declined by 56% on a per-hectare basis when a lupin crop preceded wheat 4 . A global analysis of historical data shows that cereals have greater NUE when a larger proportion of nitrogen inputs come from residues of a preceding legume crop with symbiotic nitrogen fixation, than when from synthetic fertilizer 5 .
About 21 Mt nitrogen is fixed annually by the crop legume-rhizobia symbioses 6 , returning 5-7 Mt of nitrogen to soils from about 190 million ha of grain legumes 7 , saving US$8-12 billion (Box 1). Moreover, the protein content of cereal grains produced following a legume crop is increased, and the soil structure and health are improved after growing legumes. Accordingly, nitrogen-fixing legumes provide unparalleled opportunities for minimizing future nitrogen fertilizer usage. The inclusion of grain legumes in cropping systems can enhance annual productivity 8 as well as increase diversity in cropping systems, thereby reducing the reliance on a cereal monoculture. While legume crops can favour soil acidification due to an imbalance in cation over anion uptake, this can be managed by varying the legume species used in the crop rotations and by application of lime to the soil. However, the sustainable development of intensive agriculture is limited when grain legumes are intensively cultivated on the same farmland year after year 9 , for example through the build-up of autotoxins that influence soil microbe communities 10 . The cultivation of grain legumes is a very promising way for resource-poor farmers to increase income, especially when the comparatively low input cost compared to cereals is considered 2,11 . For example, a formal survey of farmers in Bangladesh indicated awareness of the economic advantages of using integrated crop management practices for chickpea 11 .

Mitigating climate change
Future legume production will be influenced by climate change factors, such as: (1) increased atmospheric CO 2 levels favouring carbon gain because legumes use C 3 photosynthesis; (2) faster plant developmental rates due to the predicted higher temperatures, which would allow a shorter growing season and reduce exposure to drought that is often experienced at the end of the cropping season; (3) accelerated canopy decay due to extreme temperatures; (4) reduced photosynthetic efficiency, increased pod and flower abortion, and reduced production of reproductive structures due to more frequent droughts; (5) defective pollination due to hightemperature-induced pollen sterility; and (6) reduced seed quality 12 .
Regional yield changes will depend on the local manifestation of climate change (for example, interactions between high CO 2 , temperature and drought) as well as other factors. Indeed, the negative impacts of climate change have already affected soybean production: soybean yields in the USA between 1994 and 2013 declined by 2-4% for every degree rise in temperature over the growing season, resulting in losses of US$11 billion 13 . Rising global temperatures could Figure 1 | World production of cereal and grain legumes over the past 50 years. a,b, Total world production (a) and yield (b) of dry grain legumes lags behind that of cereals. c, Increased production of grain legumes is associated with expansion of land area planted with the crops, compared with cereals whose increased production is due to yield improvements while land area has remained the same. Data from ref. 79. Cereals include wheat, rice, barley, maize, rye, oats and millet, and grain legumes includes 11 of the 12 major categories of grain legume in FAO data: bambara bean; broad bean and faba bean; chickpea; cowpea; groundnut; lentil; lupin; miscellaneous grain legumes; pea; Phaseolus spp.; pigeon pea; and soybean. All production values are dry clean weights, excluding pod weights. Production of string beans were neither dried nor shelled, so data were excluded. Data for groundnuts in shells were converted assuming 70% of dry weight data in FAO was due to the grains. process, requires a highly reducing environment. A stable lowoxygen environment is achieved within the nodules by the presence of an oxygen diffusion barrier 21 . A continual oxygen flux to support bacteroid respiration is facilitated by high concentrations of leghaemoglobin. The nodules deliver reduced nitrogen to the host plant either as amides or ureides (depending on the legume species) in return for dicarboxylic acids. While the residual nitrogen present in most agricultural soils can have a negative impact on nodule formation and lifespan 1 , the sensitivity of this response varies between legume species and needs better characterization. Nitrate acts as a signalling molecule that negatively influences susceptibility to nodulation via nitrate-specific peptide signalling cascades 22 . Consequently, one of the challenges facing scientists seeking to expand legume productivity is to maximize symbiotic nitrogen fixation while allowing nitrogen acquisition from the soil. To maximize beneficial nitrogen inputs from legumes as cover crops and as green manures 23 , a molecular understanding of the nitrogen-sensing components that lead to repression of nodulation is essential. The families of membrane transporters that can double as nitrogen sensors identified in non-legumes also occur in legumes 24,25 .
Of the 400,000 plant species in existence today, only Actinorhizal plants and legumes have evolved nitrogen-fixing nodules. However, at least part of the genetic platform that facilitates the legumerhizobia interaction is conserved with other symbioses 26 . Little is therefore reduce the areas suitable for bean production 14 . Future climate conditions are predicted to be more favourable to common bean cultivation in the Northern Hemisphere than the Southern Hemisphere 15 ; new grain legume growing areas should open up in Canada, northern Europe and Russia. Already, pea and faba bean are becoming increasingly important crops in Finland 16 . Predicted changes in climate should increase yields of dry pea, chickpea, broad bean, lentil, lupin and grasspea in developed countries such as Canada and France, but yields will decrease in developing countries in the tropics and sub-tropics such as India, Pakistan and Ethiopia 17 .
The yield gap for legume crops in Africa is more than 300%, with cowpea yields being only 10-20% of their genetic potential. Moreover, while legume cultivation is an integral part of the Indian agriculture, legume crop production has remained low and unstable, with a yield gap for soybean ranging from 850 to 1,320 kg ha −1 , for groundnut 1,180 to 2,010 kg ha −1 , for pigeon pea 550 to 770 kg ha −1 and for chickpea 610 to 1,150 kg ha −1 . Such data indicate that substantial productivity improvements might be gained through improved crop management practices 18,19 .

Symbiotic nitrogen fixation
The symbiotic relationship between nitrogen-fixing bacteria (which are housed in root nodules) and legumes benefits both partners 20 . The bacterial enzyme nitrogenase, which catalyses the fixation known about the drivers for nodulation, particularly among diverse soil rhizobial populations. The management of rhizobial populations under hostile soil conditions remains a challenge, particularly in new or expanded cropping areas 27 . Our current understanding of the factors that determine nodule lifespan is superficial. Nodule senescence is a programmed, complex process that is controlled by developmental factors and environmental triggers. Activation of the senescence program by environmental stress can lead to premature loss of nitrogen-fixing activity, increased proteolysis and ultimately the death of the infected cells 28 .
Global grain legume production in 2013 was 399 Mt, with soybean (278 Mt) contributing a significant portion to the agricultural exports of the Americas (Fig. 3). However, legume crop yields tend to vary more than cereal crops 29 , largely due to environmental constraints such as drought 30,31 , which limits symbiotic nitrogen fixation 32,33 . However, nitrogen fixation is extremely tolerant to soil drying in cowpea 34 . The incorporation of improved drought tolerance and nitrogen fixation traits into elite lines of grain legumes is anticipated to generate better-yielding cultivars that can be grown on marginal land. Data comprises the grain legumes as cited in Fig. 1 plus string bean. Production of the 12 categories are presented as a stacked column graph by the ten net highest-producing countries (inset

Technologies for legume improvement
Gregor Mendel used a legume, the common garden pea, to demonstrate the 'particulate nature of inheritance' in 1865 35 . Nevertheless, many grain legume breeding programs suffer from low genetic diversity and low rates of genetic progress. For example, several bottlenecks both during and after domestication in chickpea have resulted in limited genetic diversity in the crop gene pool 36 . For soybean, 55% of the pedigree in public soybean cultivars in the USA in the 1990s was accounted for by five introductions 37 . Innovative methods of crop breeding, based on the animal model 38 , may conserve genetic diversity while accelerating grain legume genetic improvement, and may help to bridge the genetic gap between grain legumes and their wild and landrace relatives (Fig. 4). When combined with genomics-assisted breeding 39 , it should be possible to unlock valuable genes such as drought and heat stress tolerance in wild legumes, and move them efficiently into cultivated varieties 40 . Rapid introgression of important genes is a major challenge facing grain legume breeders. Grain legumes are a vital part of the response to the 2009 Declaration of the World Summit on Food Security, which requires a 70% increase in agricultural output by 2050 to keep pace with population increase, while adapting to climate change through sustainable use of genetic resources for food and agriculture. The importance of biodiversity in seed banks is widely recognized, with much attention paid to rice and other cereal crops 41 . Large genetic and phenotypic variation exists in the world collections (Table 1 and Fig. 5). It is therefore important to have a systematic inventory of legume germplasm centres and their collections. Most of the publicly available information can be found in GENESYS (Global Gateway to Genetic Resources) 42 . In addition to major CGIAR (Consultative Group for International Agricultural Research) institutes listed in Table 1 This germplasm list is not exhaustive because information is often hard to retrieve. Moreover, several accessions are duplicated across genetic resource centres. The format of the data should be standardized to facilitate easy access.
Whole-genome sequencing is an affordable and powerful tool to delineate genomic information of core germplasm 45 , and to generate high-resolution genetic maps for important agronomic traits, develop molecular markers for breeding and identify important genes for crop improvement 39 . High-resolution genetic maps are available for 10 legumes with de novo sequence information, and low-resolution maps available for all but bambara bean, tepary bean and lima bean. These resources will accelerate the development of genomics-assisted breeding strategies for legume crop improvement (Fig. 6).
Global cereal production has almost tripled over the past 50 years but grain legume production has only increased by about 60%. The relatively low rate of yield improvement in grain legumes The relationship between the use of synthetic nitrogen fertilizers and global population growth belies the untapped potential of biological nitrogen fixation by grain legume crops. Some headline facts in this debate: Synthetic nitrogen fertilizers sustain 30-50% of present crop yields and will need to increase with further population growth 86 . Global ammonia capacity is projected to grow by 16% between 2014 and 2019, with total industrial nitrogen demand set to grow by 28% over this time, compared with a 6% increase across the fertilizer sector 87 .
The environmental impact of nitrogen fertilizers is manifold, including: • Loss of biodiversity through eutrophication. Recent studies show biodiversity recovering with more environmentally stringent practices since the 1980s 89 . Eutrophication also increases production of bacterial nitrous oxide (N 2 O), which is one of the most toxic greenhouse gases 90 . • Production of N 2 O and other reactive nitrogen by-products of fossil fuel combustion, including from ammonia synthesis. The average lifetime N 2 O in the atmosphere is >100 years 90 .
The global atmospheric N 2 O concentration is now 18% higher than in pre-industrial times, and it is estimated that >30% of all atmospheric N 2 O results from agriculture. The nitrogen-use efficiency of cereals decreased from around 80% to around 30% between 1961 and 2000 86,91 . More than 50% applied nitrogen fertilizer was lost from cereal crops between 1961-2010 5,92 , and in some cases >80% of applied nitrogen is lost through runoff 90 . versus cereals is at least partly explained by low genetic diversity in grain legume breeding programs 46 . It is important to increase genetic diversity in elite breeding programs if we are to capitalize on biotechnology for legume yield improvement. Genomic selection relies on allelic diversity in the breeding population for complex traits, and may improve long-term genetic progress if accompanied by high effective population size with minimal inbreeding 47 . In animals, selection for complex traits was shown to be more efficient when based on genomic relationship information 48 . For grain legumes, many of which are self-pollinating crops, genomic selection offers the prospect of accelerating genetic progress for yield, when analysis is based on the animal model 49 . Advanced phenotyping technologies are available to measure morphological and physiological traits 50 , and high-throughput image-based phenotyping platforms will make a significant impact in plant phenomics 51 . Accurate physiological phenotyping of specific and well-defined traits will also contribute to improved breeding outcomes 52 .

Bringing in orphans
Orphan crops are minor crops with regional importance that have been largely neglected by researchers and industry due to limited economic importance in the global market. However, many people, particularly in developing countries, rely on these crops not only as food and feed crops, but also for their daily healthcare needsdespite advancements in modern medicine. They often fill ecological niches unoccupied by major crop plants, resulting in a greater genetic diversity and plasticity. Orphan food legumes such as cowpea, grass pea, the 'dolichos' bean, the tepary bean and the marama bean are usually grown in arid regions, often on marginal land unsuitable for major crop species. They have heat-and droughttolerance traits, high nutritional value and are extensively used by subsistence farmers. Cowpea is particularly valuable to people who have limited access to animal protein. The seeds have a high protein content (25% of dry weight), and the leaves are also consumed. The protein content of cowpea leaves consumed annually in Africa and Asia is equivalent to the amount in 5 million tons of dry cowpea seeds, which equates to about 30% of total food legume production in lowland tropics 53 . The 'dolichos' bean, one of the most ancient legume crops among cultivated plants, is grown as a multipurpose crop pulse, vegetable and forage. The bean is a major protein source in diets in the southern states of India. The tepary bean originated from dry subtropical areas of Mexico and the southwestern US. Tepary bean is well adapted to drought and high-temperature stresses; the major drought stress adaptation mechanisms are deep rooting for more water uptake, small leaves for reduced water use and less stomatal conductance 54 . The oil-rich marama bean, a perennial legume growing in the Kalahari Desert of southern Africa, can be more nutritious than soybean 55 . A major drawback of all these legume crops (or potential crops such as marama bean) is inefficient harvesting techniques due to the shape and density of branchesground creepers occupying large areas with limited seed yield and low propagation rates. Therefore, the promise of orphan legume crops remains largely unexplored, even though they may represent a treasure trove of undiscovered and potentially unique traits due to their great genetic diversity.

Nutrition and health
Legumes are a crucial source of a variety of phytochemicals that are important for human health. These include protein, low GI (glycaemic index) carbohydrate, fibre, minerals, vitamins, carotenoids and polyphenols 56 . Consequently, legumes hold a near-unique position among foodstuffs because of their health-determinant properties 57 . Studies in Taiwan involving people of Chinese origin revealed that all-cause mortality was increased in individuals on a legume-free diet 58 . Moreover, the mortality hazard ratio declined by 8% in older people globally for every 20 g increase in daily grain legume intake 57 .
The first study to assess the link between health and a Mediterranean diet, which included a 20 g intake of legumes per day, found a 10% reduction in all-cause mortality 59 . Legumes contribute to reduced risk of mortality because of their benefits against major chronic diseases and their risk factorsincluding cardiovascular disease, diabetes, cancer, obesity and gut health 60 . Observational studies have shown that legumes can reduce cardiovascular disease risk, and intervention studies suggest that this is mediated via improvements in blood pressure, lipid profile, inflammation, blood sugar metabolism and body weight 56,61 . Legumes also offer a food-based solution to decreasing risk of prediabetes 62 and diabetes management as well as diabetes-associated complications, especially cardiovascular disease 63 . A meta-analysis of 11 studies showed that daily consumption of legumes for more than 4 weeks resulted in a significant reduction in fasting blood glucose and insulin 64 . The effects are more pronounced when legumes are consumed as part of a low-GI diet. Since diabetes is a major risk factor for several cancers and neurodegeneration, the future health of ageing populations may be dependent on a food system that provides legumes in an affordable, palatable and sustainable way. A number of meta-analyses of observational studies have associated eating legumes with lower risk of several cancers, including bowel cancer 56 . There is growing evidence that the human microbiome plays an important role in health outcomes with several conditions, such as cardiovascular disease, obesity and colorectal cancer. Increasing evidence suggests that legumes can act as prebiotics that potentially alter bowel flora, affecting production of gut hormones and consequently appetite 56 .
Human studies using lupin-enriched foods provide insights into possible mechanisms by which legumes contribute to cardio-metabolic health. Flour made from lupin seeds contains about 45% protein and 30% fibre. Lupin-enriched foods, such as bread, pasta and biscuits, are palatable and acceptable to consumers. In clinical trials, lupin-flour-enriched bread reduced appetite and energy intake, suppressed plasma ghrelin (an orexigenic hormone that stimulates appetite) and reduced post-meal glucose and insulin responses 65 . Thus, bread made from lupin flour may influence appetite, improve glucose control, and reduce energy intake. And although regular consumption of lupin-enriched bread by overweight men and women did not alter body weight or body composition, it did reduce blood pressure and improve measures of insulin resistance 66,67 .   The positive findings on insulin resistance in normoglycaemic overweight individuals suggest that lupin foods could benefit people with type 2 diabetes. The acute effects of a lupin-based beverage on glucose and insulin responses in type 2 diabetic subjects were determined in a controlled, randomized, crossover trial, in which participants consumed a beverage containing glucose (control), or glucose plus lupin flour, or glucose plus fibre and protein from soya isolates 68 . Post-beverage glycaemic responses were significantly lower in participants following intake of the lupin beverage than the control beverage over a 4 h period after consumption. These studies on lupin-enriched foods provide evidence that legumes have cardiovascular benefits, particularly in patients with diabetes who are at a significantly increased risk of cardiovascular disease. The legume intake needed for health benefits is not clear, but observational and intervention studies indicate that intakes averaging 15 to 30 g per day are likely to be beneficial 56 .
Grain legumes such as cowpea also have potential uses in the cosmetic, food, textile and pharmaceutical sectors because of their therapeutic properties 69 . Cowpea is a source of vitamins and minerals such as folic acid, vitamins A and B, thiamine, niacin and the water-soluble vitamins riboflavin, pyridoxine and folic acid, as well as minerals such as calcium, zinc, potassium, iron and phosphorous and other trace elements 70 . Cowpea proteins have high lysine contents and are potentially an excellent supplement for cerealbased diets 71 . And the marama bean serine protease inhibitor that prevents elastase activity provides a safe and natural tryptase and elastase inhibitor (United States Patent 5869063) -elastase is part of the chymotrypsin-like clan, and human elastase has been suggested to have an important role in various inflammatory disorders, including pulmonary emphysema, sepsis, arthritis, nephritis and certain skin diseases 72 .

Conclusions and perspectives
The current use of nitrogen fertilizers in agriculture is ~110 Mt per annum, with the majority directed to cereal production. However, only 30-50% of applied nitrogen is used by the crop, and excess nitrogen fertilization has negative impacts on climate change and biodiversity. Recent increasing nitrogen fertilizer costs have focused attention on improving efficiency in cropping systems and also created a notional 'legume-envy' culture. But, while attempts to create nitrogen-fixing cereals are underway, grain legumes currently receive less research and development attention. Addition of legumes to existing cropping systems also increases the diversity of such systems. The legume-rhizobia symbiosis has enormous but largely untapped potential for sustainable agriculture, plant diversity and enhancement of primary production with reduced fertilizer use -benefits that may also extend to phosphorus-poor soils 73 .
Grain legumes lag behind cereals in terms of area expansion and productivity gains, despite increasing global demand. This lag may be due in part to unstable grain legume prices because of high variability in their yields and competition from high-yielding cereal crops. In addition, government price support policies often exist for cereals, particularly in developing countries, but inputs into grain legume breeding and agronomic technologies have been insufficient to improve yields. Our current overreliance on a handful of major staple cereal crops has inherent agronomic, ecological, nutritional and economic risks, and restricts the contributions made by underused future crops such as grain legumes. The static or declining production trends especially in developing countries, despite increasing global demand, threatens current and future food security. 2016 as the UN FAO international year of pulses (grain legumes) provides an excellent opportunity to reflect on the status of global grain legume production, consumption and potential opportunities for future expansion. Legume crops will, however, only achieve a competitive advantage if their profitability to the farmer exceeds that of the dominant cereals. To hasten the adoption of grain legume production technology by resource-poor farmers in developing countries, on-farm, farmer-participatory adaptive research and developmental approaches are required to a much greater extent than currently being implemented. The potential socio-economic gains through a boost in grain legume production and consumption are enormous. Hence, the increased public perception of the health and wellbeing advantages of a grain-legume-rich diet may be an important driver of culture change in considering grain legumes as key to food security.  Large-scale germplasm stored in different gene banks can be characterized into smaller sets such as a core/mini-core germplasm set. Such small sets of germplasm can be characterized extensively for traits of interest. Subsequently, specialized genetic stocks such as a mini-core collection (diverse set), bi-parental and multi-parent mapping populations and mutant populations can be developed. Subjecting the populations to whole-genome sequencing (WGS), genotyping by sequencing (GBS) and/or array-based genotyping, together with phenotyping for traits of interest, can provide quantitative trait loci (QTLs), quantitative trait nucleotides (QTNs), superior alleles and haplotypes. In the end, modern breeding approaches such as marker-assisted selection (MAS), markerassisted backcrossing (MABC) and genomic selection can be deployed for integrating/accumulating superior alleles. QTNs can be edited through a genome-editing approach called PAGE (promotion of alleles through genome editing). Candidate genes identified using 'omics' approaches can be deployed in genetic engineering approaches. By using one or more of the approaches mentioned above, cultivars with improved yield and nutritional quality can be developed.