Effects of abiotic stress and crop management on cereal grain composition: implications for food quality and safety

Highlights

A comprehensive review of the topic, covering phytochemicals, vitamins, fibre, protein, amino acids, sugars, and oils, with an emphasis on the implications of changes in grain composition for nutritional value, formation of processing contaminants, and food safety.

Abstract

The effects of abiotic stresses and crop management on cereal grain composition are reviewed, focusing on phytochemicals, vitamins, fibre, protein, free amino acids, sugars, and oils. These effects are discussed in the context of nutritional and processing quality and the potential for formation of processing contaminants, such as acrylamide, furan, hydroxymethylfurfuryl, and trans fatty acids. The implications of climate change for cereal grain quality and food safety are considered. It is concluded that the identification of specific environmental stresses that affect grain composition in ways that have implications for food quality and safety and how these stresses interact with genetic factors and will be affected by climate change needs more investigation. Plant researchers and breeders are encouraged to address the issue of processing contaminants or risk appearing out of touch with major end-users in the food industry, and not to overlook the effects of environmental stresses and crop management on crop composition, quality, and safety as they strive to increase yield.

Introduction

The ability of crops to tolerate abiotic stresses such as an excessive or inadequate supply of water, high winds, extreme temperatures, frost, salt, and other osmotic stresses is a key aspect of yield resilience, and its improvement has long been a target for plant breeders. The issue now has more resonance than ever because of the anticipated effects of climate change, which is predicted by the International Panel for Climate Change to bring about a rise in temperature of up to 5 °C by the end of this century and an increase in the frequency and severity of extreme weather events (Stocker et al., 2013), potentially affecting crop yields, farmer earnings, reliability of the food supply, food quality, and food safety (Vermeulen et al., 2012; Curtis and Halford, 2014).

The severe droughts experienced in Australia in 2006–2007 and Russia in 2010 may be portents of such events. Both resulted in spikes in crop commodity prices (Curtis and Halford, 2014) because both Australia and Russia are major wheat exporters in normal years. The hike in food prices in 2008 was a tipping point in the west, with policy-makers finally accepting that the food security that had been enjoyed for several decades could not be taken for granted. There has also been concern over the potential impact of climate change on food security in the world’s most populous country, China (see, for example, Xiong et al., 2007 2010; Wang et al., 2009; Zhang et al., 2010). This concern has increased recently following a period of extreme heat and drought in the south of the country in 2013. Figure 1 shows yellowing of rice plants brought about by severe drought and heat in a field trial at Shanghai Academy of Agricultural Sciences in 2013, and the leaf rolling and white spike phenotype caused by drought stress in Hainan province in 2010.

Fig. 1.

Top: upland rice growing in a field trial at Shanghai Academy of Agricultural Sciences, Shanghai, China during a period of extreme temperature and drought in August 2013. Bottom: rice grown in a farm of Hainan province (China) in 2010, showing the leaf rolling and white spike phenotypes brought about by drought stress (kindly provided by Professor Hanwei Mei, Shanghai Agrobiological Gene Centre).

Clearly, improving crop yield per se and making it more resilient to stress will be crucial to ensuring food security in the coming decades. However, yield is not the only crop parameter affected by abiotic stress, and the impact of stress and, by implication, climate change on crop composition is also important. The composition (quality) of a crop product affects its processing properties and the nutritional value, flavour, colour, and aroma of the food that is produced from it. Crucially, it also affects food safety and regulatory compliance, with the potential for formation of undesirable contaminants such as acrylamide, furan and related products, and trans fatty acids being determined by the composition of the raw crop product. Composition is also affected by crop management, notably plant nutrition, with management factors interacting with the effects of the environment.

Here we focus on the world’s major grain crops and review how cereal grain composition is affected by abiotic stress and crop management, and we consider the implications this has for food quality, processing, safety, and regulatory compliance.

Phytochemicals, vitamins, and dietary fibre

Cereal products are deeply embedded in the dietary culture of many countries. They are a rich source of energy, as well as fibre, protein, B vitamins, iron, calcium, phosphoric acid, zinc, potassium, and magnesium (Nyström et al., 2008; Shewry, 2009). While all cereal grains contain valuable minerals, vitamins, and phytochemicals (defined as non-nutritive plant chemicals that are believed to have protective or disease-preventive properties), there is considerable variation between species and genotypes. The European Union FP6 HEALTHGRAIN diversity programme, for example, analysed 150 bread wheat (Triticum aestivum) lines and 50 other lines of spelt (Triticum spelta), emmer wheat (Triticum dicoccum), einkorn wheat (Triticum monococcum), oats (Avena sativa), rye (Secale cereale), and barley (Hordeum vulgare), and found significant differences in levels of dietary fibre and phytochemicals that are considered to have health benefits (Ward et al., 2008). The study identified a number of wheat lines that combined high levels of phytochemicals and fibre with good yield and processing quality, leading the authors to conclude that it would be possible for plant breeders to adopt phytochemical and fibre content as breeding targets. There is little indication that this has occurred yet, possibly because breeders and the food industry are unconvinced that increasing the levels of beneficial components that are already present in a food would add value to their products. However, commercial decisions such as this one are outside of the control of scientists.

Subsequent analysis of 26 of the wheat cultivars grown in six site×year combinations spread across Hungary, the UK, Poland, and France in 2006 and 2007 showed important effects of the environment on nutritional value, with all groups of phytochemicals apart from alkylresorcinols and bound phenolic acids showing strong positive correlations with the mean temperature between heading and harvest (Shewry et al., 2010). Folates and free phenolic acids also showed negative correlations with total rainfall during the same period, while stanols expressed as a percentage of total sterols also showed positive correlations with temperature and negative correlations with rainfall. The amount of water-extractable arabinoxylan, a major component of dietary fibre in wheat grain, on the other hand, correlated negatively with temperature and positively with rainfall in both the bran and white flour fractions.

Li et al. (2009) also examined the relative contribution of genotype and environment to variation in arabinoxylan content, in this case of 25 hard winter and 25 hard spring wheat varieties grown in three different environments across the north-west of the USA. They concluded that both the total arabinoxylan content of the grain and the distribution of arabinoxylans between water-soluble and -insoluble fractions could be highly influenced by the environment, although there were also clear genetic effects and gene×environment interactions.

Rakszegi et al. (2014) studied the effects of both heat and drought stress during grain development on the dietary fibre content and composition (arabinoxylan and β-glucan) in three winter wheat varieties. Both stresses reduced β-glucan content of the grain but increased protein and arabinoxylan content, except that drought stress decreased arabinoxylan content in a drought-tolerant variety, Plainsman V.

Barley grain is also a rich source of dietary fibre, but β-glucan rather than arabinoxylan is the predominant component, accounting for ~70% of the dietary fibre in the barley starchy endosperm (white flour fraction). β-Glucan content has been shown to be affected by nitrogen application and water availability in barley in a study conducted in Turkey (Güler, 2003). Generally, high nitrogen levels increased grain β-glucan content, while higher levels of irrigation tended to decrease it, in contrast to its effects on β-glucan in wheat grain reported by Rakszegi et al. (2014). Narasimhalu et al. (1995) also studied the effect of environment on β-glucan content of barley grain, this time in Canada, using 75 different cultivars and six different sites across the country. Environment, represented by the different sites, significantly affected total β-glucan content, with grain from barley grown in the drier east containing more β-glucan than grain produced in the west. Similarly, Swanston et al. (1997) found more β-glucan in grain from barley grown in Spain compared with barley from the considerably cooler and wetter environment of Scotland, UK. These results were at odds with those of a previous study that suggested that low glucan content occurred in grain from drier environments (Coles et al., 1991). Other factors may have been at play to explain the contrasting results, but at least it can be concluded that β-glucan, and therefore dietary fibre content of barley grain, is sensitive to environmental conditions.

Brown rice is also a valuable source of phytochemicals and vitamins. Indeed, it was the work of doctors in Asia in the late 19th century, investigating the effect of incorporating brown rice into the diets of prisoners and livestock for preventing the disease beriberi, that led to the discovery that foods contain complex nutrients that are essential for health. Prisoners and animals fed a diet of white rice were prone to developing the disease, which affects the peripheral nervous system, while those fed brown rice were not (Carpeneter, 2000). This eventually led to the coining of the word vitamin and the discovery of thiamine (vitamin B1). Brown rice also provides lipid-soluble antioxidants, including ferulated phytosterols such as γ-oryzanol, tocopherols, and tocotrienols. Elevated temperatures during cultivation have been shown to change the profile of some of these phytochemicals, for example increasing α-tocotrienol and/or α-tocopherol while decreasing γ-tocopherol and γ-tocotrienol (Britz et al., 2007).

This section of the review brings home the contribution that cereal products, particularly wholegrain cereal products, make to our diet and food security, in terms of providing nutritious as well as sufficient food. It is particularly important to make this point at the moment because some clinicians and dieticians have suggested that cereal, and particularly wheat, consumption contributes to overeating and obesity (see, for example, Davis, 2011). This notion has been debunked by Brouns et al. (2013), who instead found associations between the consumption of the recommended amounts of wholegrain wheat products and significant reductions in risk for type 2 diabetes, heart disease, and weight gain. The other conclusion from this section is that if we value the vitamins, phytochemicals, and dietary fibre present in wholegrain cereals then more should be done to understand the effect of environmental and crop management factors on their concentration.

It is a fact, however, that persuading consumers to eat more wholegrain cereal products in preference to foods produced from white flour has not proved easy, and while the consumption of brown rice may have significant health benefits compared with white rice, the husk is discarded not just because of consumer preference but also because it rapidly goes rancid during storage, especially in tropical countries.

Protein

Protein content and composition is an important determinant of cereal grain quality and it is sensitive to drought and heat stress as well as atmospheric CO₂ concentration. Heat stress, for example, reduces starch deposition in wheat grain, resulting in an increase in grain protein content (Wardlaw et al., 2002; Gooding et al., 2003). Wrigley et al. (1994) showed that it also affects protein composition and that, as a consequence, wheat dough strength is adversely affected by even a short period of high temperature (> 35 °C) during grain filling. The reduction in dough strength is associated with a decrease in the proportion of insoluble glutenin to soluble gliadin proteins in the grain (a higher proportion of glutenins promotes cross-linking between proteins in gluten, the proteinaceous matrix that gives wheat dough its viscoelasticity) (Blumenthal et al., 1993). The effect of the high temperature is greater if it occurs during mid to late grain filling than if it occurs early (Corbellini et al., 1997), while slightly lower temperatures, ~30 °C, are actually beneficial, giving better dough strength than lower temperatures (Wrigley et al., 1994). Balla et al. (2011) showed that drought stress also causes a disproportionate reduction in the glutenin fraction compared with the gliadin fraction in wheat grain, with an adverse effect on dough strength. Drought stress has also been shown to cause a small reduction in total protein content in two maize cultivars grown in Pakistan (Ali et al., 2010) but to increase it by up to a fifth in rice (Crusciol et al., 2008; Fofana et al., 2010). Temperature stress, on the other hand, causes a reduction in protein content in rice (Ziska et al., 1997).

Elevated levels of atmospheric CO₂ (700 ppm compared with 350 ppm) have been shown to have a negative effect on wheat grain quality, largely through reducing the overall protein content, although increased CO₂ does increase grain yield (Blumenthal et al., 1996). Grain grown under elevated CO₂ produces poorer dough and decreased loaf volume, which is important given that the atmospheric CO₂ concentration has already risen from a pre-industrial level of 270 ppm to almost 400 ppm and is rising at 1–2 ppm per year, taking it to levels not seen for 20 million years.

Wheat protein content and quality are also affected by plant nutrition and therefore crop management. The importance of a plentiful supply of nitrogen to produce a high yield of grain with not only a high protein content but also protein quality that is acceptable to bread-makers has been known since the 19th century (see Hawkesford, 2014, for a recent review). Also important for protein quality is an adequate supply of sulphur. Zhao et al. (1999), for example, showed that sulphur deficiency limits storage protein accumulation, with wheat grown in sulphur-deficient soil favourably accumulating sulphur-poor proteins at the expense of sulphur-rich glutenins. Indeed, in that study, sulphur was shown to be more important for bread-making quality than nitrogen. This confirmed a much earlier study that found barley plants to have reduced ability to synthesize sulphur-rich storage proteins (called hordeins in barley) when starved of sulphur (Shewry et al., 1983). The total hordein fraction was decreased from 51% to 27% of the seed nitrogen in one variety and from 46% to 26% in another. Different combinations of nitrogen and sulphur fertilizer have also been shown to affect grain quality in durum wheat (Lerner et al., 2006; Rogers et al., 2006; Godfrey et al., 2010).

The changes in accumulation of different types of protein in cereal grain in response to nutrition can be attributed to changes in gene expression. Seed storage protein gene expression is controlled primarily at the transcriptional level (Bartels and Thompson, 1986; Sørensen et al., 1989) and responds sensitively to the availability of nitrogen and sulphur in the grain (Giese and Hopp, 1984; Duffus and Cochrane, 1992). Many wheat, barley, and rye storage protein genes contain a so-called GCN4-like motif (GLM), nitrogen element, or N motif (reviewed by Shewry et al., 2003), nucleotide sequence G(A/G)TGAGTCAT, in the promoter region. This motif exerts a negative effect on gene expression at low nitrogen levels but a positive one when nitrogen levels are adequate (Müller and Knudsen, 1993), and binds at least two different transcription factors (Hammond-Kosack et al., 1993; Albani et al., 1997). However, little is known about the mechanisms by which storage protein gene expression responds to sulphur.

While considering the importance of grain proteins for quality, it should be remembered that for small subsets of the population the proteins of wheat, rye, and barley have an adverse effect on health (reviewed by Shewry, 2009). Some are allergenic, for example, with the respiratory allergy, bakers’ asthma, being an important occupational allergy affecting bakers. Many proteins appear to be involved (reviewed by Tatham and Shewry, 2008), but the most important are a class of α-amylase inhibitors known as CM proteins because of their solubility in chloroform/methanol mixtures. Wheat also causes food allergies, including wheat-dependent exercise-induced anaphylaxis (WDEIA), in which consumption of wheat followed by physical exercise can induce an anaphylactic response. The proteins responsible are a group of seed storage proteins known as ω5-gliadins (Tatham and Shewry, 2008; Shewry, 2009). Dietary intolerance to wheat is also an important problem, with coeliac disease, a chronic inflammation of the bowel that reduces the bowel’s ability to absorb nutrients, affecting ~1% of the population of Western Europe, and dermatitis herpetiformis, which causes skin eruptions, affecting between 0.2% and 0.5% of the population (Shewry, 2009). These conditions are triggered by a wide range of gluten proteins, so there appears to be little scope for addressing the problem through either breeding or crop management.

Free amino acid and sugar accumulation

Free amino acids are sometimes overlooked when considering grain concentration because most of the nitrogen in the grain is incorporated into proteins. This is unfortunate because, as we describe below, free amino acids, together with sugars, are major determinants of processing quality and in some cases food safety. Free amino acids are always present in plant tissues to enable protein synthesis to proceed. They accumulate to high concentrations during some developmental processes such as germination and senescence, but even higher concentrations can occur in almost any tissue in response to a range of both biotic and abiotic stresses, including nutrient deficiency, pathogen attack, toxic metals, drought, and salt stress (reviewed by Lea and Azevedo, 2007).

Examples of this in cereals include the accumulation of free amino acids in the leaves of Hordeum species in response to salt stress, which was demonstrated in a study by Garthwaite et al. (2005). This study showed that concentrations of glycinebetaine (an N-trimethylated amino acid), asparagine, and proline all increased in leaves with increased external NaCl concentration. The proline increase differed between species; H. vulgare, for example, had a 17-fold increase in the expanding leaf blade, while H. muratum had an 8-fold increase. A similar increase in free amino acids was also observed during leaf folding for drought-stressed pearl millet plants (Kusaka et al., 2005). The levels of free asparagine and proline were increased 9- and 18-fold, respectively, 6 d after stopping irrigation, and 15- and 28-fold after 9 d. The excess of asparagine in the leaves was considered to be an indication of protein degradation under stress conditions, but could have arisen in part through de novo synthesis.

The increase in concentration of some amino acids and their derivatives in vegetative tissues in response to osmotic stress imposed by, for example, drought and salt, has led to the hypothesis that increases in the concentration of these metabolites are not just symptoms of stress but are an important part of the stress response, decreasing cell osmotic potential and thereby increasing turgor while decreasing plant water potential. It follows that increased tolerance to stress could be imparted by genetic interventions that increase the concentration of these metabolites, with proline and glycinebetaine receiving particular attention. This has been reviewed in detail by Lawlor (2013). Examples of this approach being applied to cereals include the overexpression of a Δ¹-pyrroline-5-carboxylate synthetase from mothbean (Vigna aconitifolia L.) under the control of a stress-inducible promoter in transgenic rice (Zhu et al., 1998), leading to increased proline accumulation. Second-generation transgenic plants expressing the transgene showed increased biomass compared with controls under both salt and drought stress. A mothbean Δ¹-pyrroline-5-carboxylate synthetase has also been overexpressed in wheat, again under the control of a stress-inducible promoter (Vendruscuolo et al., 2007), resulting in improved drought tolerance, although this was mainly due to reduced oxidative stress rather than osmotic adjustment. Lightfoot et al. (2007) targeted another area of amino acid metabolism by overexpressing a bacterial glutamate dehydrogenase gene in transgenic maize. The plants showed improved drought tolerance, leading the authors to conclude that transgenic plants overexpressing glutamate dehydrogenase could outperform their conventional counterparts in semi-arid conditions.

Lawlor (2013) expressed scepticism over whether transgenic plants such as these show genuinely improved drought tolerance, as opposed to changes in growth, development, or metabolism that give the appearance of improved drought tolerance but would not bring about better performance of the crop under drought conditions in the field. Nevertheless, it is an approach that continues to be adopted by many researchers.

Proline concentration in vegetative tissues may be the target of most genetic interventions like these, but, as we state above, proline is not the only free amino acid that accumulates in response to stress. The free asparagine concentration is also highly responsive to environmental conditions and, crucially for grain quality and food safety, while it usually accounts for <10% of the total free amino acids in cereal grains (Lea et al., 2007), it can become by far and away the predominant free amino acid in the grain under stress conditions. In general, free asparagine accumulates when the rates of protein synthesis are low and there is a plentiful supply of reduced nitrogen (Lea et al., 2007), either as a result of inhibition of protein synthesis or through direct effects on asparagine metabolism, or both. Not surprisingly, therefore, nitrogen availability correlates positively with free asparagine content and this has been shown in barley (Winkler and Schön, 1980), wheat (Martinek et al., 2009), and rye (Postles et al., 2013). When there is a plentiful supply of nitrogen, deficiencies in other minerals become important (reviewed by Lea et al., 2007). Sulphur deficiency in particular can cause a massive (up to 30-fold) increase in the accumulation of free asparagine (up to 50% of the total free amino acid pool) and to a lesser extent free glutamine in wheat, barley, and maize (Shewry et al., 1983; Baudet et al., 1986; Muttucumaru et al., 2006; Granvogl et al., 2007; Curtis et al., 2009), although rye is much less responsive, at least under field conditions (Postles et al., 2013).

As described in the section on proteins, sulphur deficiency brings about a reduction in the expression of sulphur-rich storage proteins in wheat and barley, and this may explain some of the increase in free amino acid accumulation in the grain. However, the fact that free asparagine and to a lesser extent free glutamine are disproportionately accumulated compared with other free amino acids suggests that these amino acids are used by the plant to store nitrogen under nutrient-limited conditions. Consistent with this hypothesis, asparagine synthetase gene expression in wheat has been shown to increase under sulphur-limited growth conditions. This response appears to involve the protein kinase, TaGCN2 (Byrne et al., 2012); TaGCN2 is related to General Control Non-derepressible 2 (GCN2), a master regulator of amino acid metabolism and protein synthesis in yeast (Saccharomyces cerevisiae) (Wek et al., 1989). GCN2 and TaGCN2 phosphorylate the α-subunit of translation initiation factor 2 (eIF2α). In yeast this results in translational up-regulation of a transcription factor, GCN4. There is no clear homologue of GCN4 in plants, but, as discussed in the section on proteins, some storage protein genes do contain a regulatory motif that matches the GCN4 binding site. It is therefore possible that TaGCN2 could link the regulation of storage protein and asparagine synthesis, although much more of the signalling systems involved would have to be elucidated before that could be confirmed.

Asparagine synthetase gene expression in wheat also increases in response to salt and osmotic stress (Wang et al., 2005), and there is evidence from several studies that the free asparagine concentration in the grain of both wheat and rye varies considerably in the same variety grown at different locations or in different years, showing that asparagine metabolism is responsive to multiple environmental and crop management factors (Taeymans et al., 2004; Baker et al., 2006; Claus et al., 2006a ; Curtis et al., 2009, 2010).

The same is true for sugar concentrations in cereal grain. While significant genotypic variation has been reported in studies on, for example, wheat (Claus et al., 2006a ; Muttucumaru et al., 2006; Hamlet et al., 2008) and rye (Curtis et al., 2010; Postles et al., 2013), environmental factors are also important. For example, high temperatures during grain filling in wheat cause an increase in sucrose, reducing sugars, and sugar phosphates, and a reduction in starch (Jenner, 1991). Gooding et al. (2003) also reported a decrease in Hagberg falling number, which is indicative of reduced starch and increased sugars, in heat-stressed wheat.

Curtis et al. (2010) analysed a diverse selection of rye varieties, including some that are no longer used in cultivation, grown at sites in France, Hungary, Poland, and the UK over several growing seasons, and reported a wide range of sugar concentrations, particularly for sucrose (26.81–49.52 mmol kg^–1), across the different sites and harvest years, even within the same genotype in some cases. Postles et al. (2013) also measured sugar concentrations in rye, but in this case the grain came from elite commercial varieties grown in a field trial at a single location over one growing season, and a lower and much tighter range of concentrations was obtained (4.39–5.03 mmol kg^–1 for sucrose). The two studies also showed differences in reducing sugar concentration, with glucose being the most abundant reducing sugar in the 2010 study, again showing a considerable range, from 0.64 mmol kg^–1 to 33.43 mmol kg^–1, and fructose being the most abundant reducing sugar in the 2013 study, ranging from 2.88 mmol kg^–1 to 4.37 mmol kg^–1.

Postles et al. (2013) also looked at the effect of nitrogen and sulphur nutrition on sugar levels in rye grain. Nitrogen affected the fructose, glucose, and total reducing sugar concentration, but the effect was variety-dependent, with one variety, Agronom, for example, accumulating a higher concentration of reducing sugars at an intermediate nitrogen application rate (100kg ha^–1) than at very low or high application rates, while another, Askari, had a lower concentration of reducing sugars at the intermediate nitrogen than at lower or higher application rates. Sulphur had no effect on the reducing sugars, but its application at 15kg ha^–1 or 40kg ha^–1 resulted in a significant reduction in sucrose concentration in one variety, Rotari.

The results of these studies emphasize the effects of genotype, environmental conditions, including crop management, and genotype×environment interactions on sugar concentrations in cereal grain. An even wider range of sucrose concentrations has been reported for maize (12.91–89.60 mmol kg^–1; Harrigan et al., 2007) and rice (15.10–58.60 mmol kg^–1; Smyth et al., 1986). However, relatively few studies have investigated the effects of abiotic stress on sugar concentrations in cereal grain, while many have focused on the effect of stresses on the sugar content of leaves and seedlings, a theme that has characterized physiological studies of the effects of stress and crop management on sugars in cereals and, indeed, other crops. We have reviewed these studies in detail previously (Halford et al., 2011) and, since they are not directly relevant to this review, have not done so again here. In summary, cereals and other plants interconvert monosaccharides, disaccharides, and more complex carbohydrates such as fructan in order to cope with osmotic stresses, including those caused by salt, freezing, and drought, as well as other stresses such as hypoxia and early senescence (Halford et al., 2011). Indeed, rye’s superior stress tolerance compared with its near relatives is attributed to a greater capacity for carbohydrate storage and more rapid hydrolysis of fructan to increase concentrations of simple sugars when required. Interestingly, fructan has recently been shown to accumulate in the early stages of wheat grain development (Verspreet et al., 2013), although its role in that case is not known.

Colour, flavour, and aroma volatiles and processing contaminants produced from sugars and free amino acids

Understanding the role of sugars and free amino acids, and the interconversion of simple sugars and complex, insoluble carbohydrates in vegetative tissues in response to osmotic and other stresses is important for the development of genetic interventions that will make crops more resilient to and therefore higher yielding under abiotic stress conditions. However, it is also important to understand the effects that abiotic stress has on the concentration of these metabolites in the grain because free amino acids and sugars have profound effects on the processing properties of the grain. Indeed, free amino acids and sugars have been considered in the same section of this review because they combine during baking, frying, and high-temperature processing to produce a host of compounds, including some imparting colour, flavour, and aroma, and others that are potentially harmful to health.

The reaction in which free amino acids and sugars combine to form these compounds is the Maillard reaction, an umbrella term for a series of non-enzymic reactions that takes place only at high temperatures. The reaction requires a reducing sugar such as glucose, fructose, or maltose, and the products of the reaction include melanoidin pigments and a complex mixture of compounds that impart flavour and aroma, including pyrazines, pyrroles, furan (which can also be described as a contaminant), oxazoles, thiazoles, and thiophenes (Mottram, 2007). The most important bread flavour compound, for example, is 6-acetyl-1,2,3,4-tetrahydroxypyridine, while 2-acetyl-1-proline provides the main flavour in wheat and rice crackers. So, it is the Maillard reaction that gives bread crust, biscuits, rye crisp-breads, and a wide variety of other popular foods their characteristic flavour, aroma, and texture.

The Maillard reaction is very complex, and is described in detail elsewhere (Nursten, 2005; Mottram, 2007; Halford et al., 2011; Curtis et al., 2014b ). While many of its products are highly desirable, some are not (Friedman, 2005) and can be classed as processing contaminants; that is, substances that are produced in a food when it is cooked or processed, are not present or are present at much lower concentrations in the raw, unprocessed food, and are undesirable either because they have an adverse effect on product quality or because they are potentially harmful (Curtis et al., 2014b ). This definition distinguishes them from substances such as mycotoxins that are biotic in origin and, while important, are outside the scope of this review.

Products of the Maillard reaction in cereal products that can be considered as processing contaminants include acrylamide and furan (Curtis et al., 2014b ) (Fig. 2). Of these, acrylamide is probably the one that is of most concern to the food industry at present. Acrylamide is formed within the Maillard reaction when free asparagine participates in the later stages (Mottram et al., 2002; Stadler et al., 2002; Zyzak et al., 2003). Indeed, the carbon skeleton of the acrylamide that forms is derived entirely from asparagine. It should be noted, however, that while this appears to be the major route for acrylamide formation, others have been proposed, for example with 3-aminopropionamide as a possible transient intermediate (Granvogl and Schieberle, 2006) or through pyrolysis of gluten (Claus et al., 2006b ). Nevertheless, asparagine concentration correlates closely with acrylamide formation in heated wheat and rye flour (Muttucumaru et al., 2006; Granvogl et al., 2007; Curtis et al., 2009 2010; Martinek et al., 2009; Postles et al., 2013).

Fig. 2.

Diagrams representing the structures of acrylamide, furan, and hydroxymethylfurfuryl

Acrylamide is familiar to biochemists because of its use in gel electrophoresis, and it also has applications in wastewater treatment, papermaking, and manufacture of fabrics. It has been classified as a Group 2A, ‘probably carcinogenic to humans’, chemical by the International Agency for Research on Cancer (International Agency for Research on Cancer, 1994) because of the carcinogenicity it has shown in rodent toxicology studies (Friedman, 2003; Taeymans et al., 2004). The concentrations of acrylamide in the human diet are much lower than those used in such studies (Friedman, 2003), and the results of epidemiological studies on the effect of dietary acrylamide on cancer rates in humans have been inconsistent. Nevertheless, the latest report on the issue from the European Food Safety Authority (EFSA)’s Expert Panel on Contaminants in the Food Chain (CONTAM) described acrylamide in food as potentially increasing the risk of developing cancer for consumers in all age groups (EFSA CONTAM Panel, 2014). The Food and Agriculture Organization of the United Nations and the World Health Organization (FAO/WHO) Joint Expert Committee on Food Additives (JECFA) has also concluded that the presence of acrylamide in the human diet is a concern (Joint FAO/WHO Expert Committee on Food Additives, 2011). Acrylamide also has neurological, reproductive, and developmental effects at high doses, but CONTAM considers these not to be a concern at current levels of dietary exposure (EFSA CONTAM Panel, 2014).

The European Commission issued ‘indicative’ levels for acrylamide in food in early 2011, then revised them in 2013, with levels for many cereal products being lowered (European Commission, 2013). These values are intended to indicate the levels that the Commission considers the food industry should be able to achieve, based on the results of its acrylamide monitoring programme. In the USA, the Food and Drug Administration (FDA) has not issued advice or restrictions, but has developed an ‘action plan’ with a number of goals, including identifying means to reduce exposure. However, in 2005, the Attorney General of the State of California filed a lawsuit against four food companies for not putting a warning label on their products to make consumers aware that the products contained acrylamide. The lawsuit was settled in 2008 when the companies committed to cutting the level of acrylamide in their products to <275 μg kg^–1 and paid US$3 million in fines.

Fried potato products and coffee are major contributors to dietary acrylamide, but many cereal-based foods are also affected, including bread, biscuits, breakfast cereals, crisp-breads, cakes, pastries, and tortilla chips. More acrylamide tends to form in products that are baked to a crisp texture and dark colour, such as biscuits, crisp-breads, and breakfast cereals: the most recent EFSA report puts the mean acrylamide content for biscuits and crisp-breads sampled in the European Union at 265 μg kg^–1 (parts per billion) (EFSA CONTAM Panel, 2014) and for breakfast cereals at 161 μg kg^–1. However, dietary exposure is a function not only of the concentration of acrylamide in the product but also of the amount of the product that is consumed, and although the acrylamide content of soft bread is comparatively low (mean 42 μg kg^–1), soft bread is the major cereal-based contributor to dietary intake, accounting for between 15% of total intake in the UK and 42% in Germany (European Food Safety Authority, 2011a ).

The food industry reacted quickly to the discovery of acrylamide in some of its products and has devised many strategies for reducing acrylamide formation by modifying food processing (compiled in a ‘Toolbox’ produced by Food Drink Europe, 2011). However, the cereal breeding industry generally did not react with the same urgency. So, while several studies have shown that there are significant differences between varieties with respect to free asparagine concentration and therefore acrylamide-forming potential in the grain (reviewed by Halford et al., 2012), there is no sign of low free asparagine concentration being adopted as a major target by most breeders or that varieties with high acrylamide-forming potential are being excluded from breeding programmes. What is more, with most work on the issue being carried out in Europe, where rice and maize are not major crops, there is little information available on the acrylamide content of rice and maize products that may contain acrylamide, such as fried rice, rice crackers, and cakes (such as Chinese nian gao) and maize tortillas and tortilla chips, and screening of maize and rice varieties appears to be lacking altogether. In this respect, cereal breeders appear to be out of touch with the needs and priorities of the food industry that uses its products. In contrast, some potato breeders have begun to use acrylamide-forming potential in their advertising, which at least indicates that they are taking the issue seriously.

Farmers also have a part to play. The effect of sulphur deficiency on the concentration of free asparagine in wheat and barley grain means that sulphur availability is the most important factor affecting the acrylamide-forming potential of these grains. In a recent study in the UK, for example, acrylamide formation was analysed in flour from wheat grain samples produced from six field trials in which different levels of sulphur had been applied (Curtis et al., 2014a ). The trials comprised four different varieties of winter wheat, grown at three different locations over three harvest years, with five different levels of sulphur fertilization. The level of application that gave a significant benefit compared with no sulphur application, with no further significant benefit with higher levels of application, ranged from 5kg to 20kg of sulphur ha^–1. However, application at the 20kg sulphur ha^–1 rate led to significantly lower grain asparagine concentration and less acrylamide formation compared with lower rates in two of the trials, the results for one of which are shown in Fig. 3. Given the necessity of preventing free asparagine accumulation in all conditions, it was concluded that sulphur-containing fertilizer should be applied at a rate of 20kg sulphur ha^–1 wherever sulphur deficiency is thought to be a risk and wheat is being grown for use in products for human consumption.

Fig. 3.

Free asparagine concentration in the grain and acrylamide formation in wholemeal flour (heated for 20min at 170 °C) for wheat cv. Alchemy grown at Brockhampton, Herefordshire, UK in 2009–2010 with different rates of sulphur application (data from Curtis et al., 2014a ).

This rate of sulphur application is at the top of the range of rates of sulphur fertilizer already recommended in the Fertiliser Manual (RB209) for wheat in the UK (Fertiliser Manual, 2011). However, previous recommendations were based on the rate of application required to ensure that the protein quality of bread-making wheat was sufficient to meet the requirements of bakers. Some food products with a relatively high risk of acrylamide formation, such as breakfast cereals and biscuits, do not require wheat with the high quality protein content that is preferred for bread-making. The results of the study showed that it is important that sulphur be applied at a rate of 20kg ha^–1 to wheat destined for these products as well as wheat being grown for bread production, regardless of yield and other quality issues (Curtis et al., 2014a ).

As we have described above, nitrogen has the opposite effect on free asparagine concentration to sulphur, with increasing nitrogen application bringing about an increase in free asparagine concentration. Nitrogen fertilizer is, of course, extremely important for obtaining high yields of good quality grain, in any cereal. The current best advice is that the crop should be supplied with other nutrients as well, particularly sulphur, to ensure that as much of the nitrogen as possible is incorporated into protein rather than accumulating as free asparagine (Halford et al., 2012).

There are some things, of course, that farmers cannot control and, as we have stated earlier in this section, asparagine metabolism is responsive to multiple environmental factors. It is important that the exact environmental triggers for asparagine accumulation in cereal grain are identified in the interests of quality control and, in the longer term, the development of strategies for producing varieties that are consistently low in grain asparagine concentration across a range of environments.

Acrylamide is not the only contaminant formed in the Maillard reaction: furan (Fig. 2) and related compounds are also Maillard reaction products and are attracting increasing attention because of their potential effects on human health. Furan consists of a five-membered aromatic ring comprising four carbon atoms and one oxygen (Fig. 2). It is classed by the International Agency for Research on Cancer as ‘possibly carcinogenic’ to humans (Group 2B) (International Agency for Research on Cancer, 1995), because it causes liver cancer in rodents (Leopardi et al., 2010). Like acrylamide, the actual risk posed by the presence of very low levels of furan in food is not known. The Maillard reaction involving sugars and free amino acids is not the only route for its production: it can also form directly from the degradation of sugars, polyunsaturated fatty acids (PUFAs; which also feature in the next section), or ascorbic acid (reviewed by Curtis et al., 2014b ).

Cereal-based foods make a significant contribution to dietary exposure to furan prior to adulthood. Cereal-based baby foods, for example, contribute 3% of the dietary intake in infants and 12% in toddlers, while other cereal products contribute 7.7% of the dietary intake of older children and 10% of the intake of adolescents (European Food Safety Authority, 2011b ). In adults, however, dietary intake from cereal products is dwarfed by that from coffee.

A related compound that can form in cereal products is hydroxymethylfurfuryl (HMF) (Ramírez-Jiménez et al., 2000), which consists of a furan ring with both aldehyde and alcohol functional groups (Fig. 2). HMF arises via the dehydration of fructose (Román-Leshkov et al., 2006) and is a common contaminant of dark beers (Husøy et al. 2008) and over-cooked biscuits. Indeed, it is sometimes used as an indicator of excessive heat treatment in biscuit manufacture. This is mainly because products containing high levels of HMF may also contain a lot of acrylamide, so measuring HMF levels is a quality control procedure that is used to ensure that acrylamide levels are lower than the indicative value. However, HMF itself may have safety implications because one of its metabolic products is 5-sulphoxymethylfurfural, which shows potential toxicity and carcinogenicity in rodent studies (Husøy et al., 2008).

As with acrylamide, modifying food processing conditions can reduce furan formation in foods (Fan et al., 2008), but there appears to have been little work on reducing the furan-forming potential of crop products, possibly because of the multiple routes through which furan can form. Clearly, lowering fructose levels would be a possible strategy for reducing HMF formation, but so far we are not aware of this being investigated in a scientific study and it would almost certainly also affect the production of beneficial Maillard reaction products. Environmental factors are clearly also potentially extremely important, given the changes in sugar concentration that can occur in response to abiotic stresses.

Oils

Most cereal grains do not contain enough oil to be considered suitable for commercial oil production, but maize and oat grain contain ~5% and 7% oil, respectively. Oat oil is not widely used in food or industrial processes, except as an ingredient in some cosmetics and skin moisturisers, but maize oil is an important commodity, with 1.3 billion litres of it being used in the USA alone in 2012, mostly for margarine and cooking oil but more recently also for biodiesel production.

The principle components of plant oils are, of course, triacylglycerols, which are made up of three fatty acid molecules esterified with glycerol. Different fatty acids are distinguished by the length of their hydrocarbon chain and by the number and position of double bonds between the carbon atoms in the chain, with saturated fatty acids containing no double bonds, monounsaturates a single double bond, and PUFAs multiple double bonds. Maize oil is typically made up of ~12% palmitic acid (saturated, 16-carbon chain, 0 double bonds, represented as 16:0), 2% stearic acid (18:0), 30% oleic acid (18:1), 54% linoleic acid (18:2), and 1% α-linolenic acid (18:3), but Ali et al (2010) showed that this changes dramatically in response to drought stress, which not only causes a reduction in grain oil content by 40% but also causes oleic acid to increase to >25% of the total at the expense of linoleic acid.

PUFAs are relatively unstable because they are subject to thermal oxidation during cooking and high-temperature processing, giving rise to lipid peroxides. Lipid peroxides form polymers that give a dark coloration and may be toxic, and can break down to form products that cause a rancid, ‘off’ flavour and odour, as well as furan, which was discussed in the previous section. Oxidation also occurs during long-term storage, so a high PUFA content shortens the shelf-life of the oil. Food processors prevent PUFA oxidation by chemical hydrogenation of the double bonds, producing a more stable, saturated fatty acid. This may also be done to solidify the oil, making it suitable for the production of margarines (saturated fatty acids have a higher melting point that unsaturated fatty acids). The problem with chemical hydrogenation of PUFAs is that some of the double bonds remain unsaturated but change from the cis form, with the two hydrogen atoms attached to the carbon atoms involved in the double bond on the same side, to the trans form, with the two hydrogen atoms on opposite sides. Trans fatty acids arising from partial hydrogenation of plant oils are now regarded as being as harmful as saturated fatty acids when consumed (reviewed by Brouwer et al., 2010), and can therefore be considered to be another important class of processing contaminants.

The accumulation of more oleic acid and less linoleic acid in maize under drought stress may produce a more stable oil that does not require chemical hydrogenation. However, the fact that the environmental conditions can have such a profound effect on the fatty acid profile is important and requires further study. This also has implications for biodiesel production, which is now a major use for many plant oils. Biodiesel is manufactured from plant oils by transesterification of triacylglycerols with methanol, producing fatty acid methyl esters (FAMEs) with glycerol as a by-product. The properties of a particular biodiesel are therefore dependent on the composition of the plant oil from which it is made.

Conclusions

We have reviewed the effects of abiotic stress and crop management on the composition of cereal grain and the implications this has for food quality and safety. The need to ensure that cereal crops have a sufficient supply of nutrients so that the grain contains sufficient, good quality protein for high-end uses such as bread-making is widely understood. What has received less attention from cereal scientists and breeders is the importance of grain composition for food safety. One aspect in particular that needs much more investigation is the identification of specific environmental stresses that affect composition in ways that have implications for food safety and how these stresses interact with genetic factors and will be affected by climate change. This is a key conclusion of this review.

It is important that this is addressed because the problem of processing contaminants in cereal and other food products does not appear to be one that is going to go away any time soon, and the food industry faces a difficult and evolving regulatory system that has to be complied with regardless of the inconsistencies in the raw materials that the industry has to use. Indeed, the acrylamide problem is one of the most pressing currently facing large sectors of the food industry. Processors have put a lot of effort and money into reducing the levels of acrylamide in their products to as low as is reasonably achievable without fundamentally changing the characteristics that define those products and are demanded by consumers (the ALARA principle). There is increasing frustration within the food industry at the lack of engagement of plant scientists and breeders when it is clear that the development of varieties that gave a more consistent raw material would make it much easier for food producers to comply with regulations. Plant researchers and breeders are therefore in danger of appearing to be out of touch with major end-users.

It is also conceivable that plant breeders who continue to ignore the issue will suffer in the marketplace. One of the simplest and cheapest ways for a food company to address a contaminants problem in one of its products is to switch to a raw material with a lower risk of that contaminant forming, whether it be a different variety or even in some cases a different crop species. The company can then continue to make its product without costly changes to the manufacturing process.

It is in the interests of researchers in the plant sciences to recognize the importance of the food industry as a potential beneficiary of plant research. As a community, plant scientists argue that their role in underpinning a sustainable and competitive agriculture industry justifies the support they receive from taxpayers, but would do well to point out just as strongly the relevance and importance of their work to the food industry. The economic case for this in developed countries is simple: the UK, for example, which is classed by the World Bank as a high income country, has an efficient and highly productive agriculture sector, but the sector’s gross value added (GVA) figure (i.e. the total value of the goods that it produces) of £7.7 billion per annum (Department for Environment, Food and Rural Affairs, 2013) represents only 0.6% of the UK’s total, while the GVA of the food industry is more than three times that at £24 billion (Food and Drink Federation, 2014). Indeed, on the basis of its turnover of £92 billion and workforce of 400 000 people, the food industry is the largest sector within UK manufacturing. Agriculture continues to be a major contributor to the economy of developing economies, but its contribution tends to decline as the economy develops. In China, for example, which is classed as an upper middle income country, agriculture accounted for 15.1% of economic activity in 2000 but for only 10.1% in 2012 (OECD, 2014).

We conclude by encouraging researchers, breeders, and farmers not to overlook the effects of environmental stresses and crop management on crop composition in the drive to increase yield, with a reminder that the FAO’s definition of food security refers to safe and nutritious as well as sufficient food (Food and Agriculture Organization of the United Nations, 2003).