Aflatoxins, are highly toxic and carcinogenic compounds produced by the fungi, Aspergillus flavus and A. parasiticus during growth on crops such as maize, peanut, cottonseed, and tree nuts [1]. A. flavus is most commonly associated with aflatoxin contamination of susceptible crops though A. parasiticus is often associated with contamination of peanut. The presence of aflatoxins in agricultural commodities poses a serious health threat to both humans and domesticated animals which is why the U.S. Food and Drug Administration (FDA) and their counterparts in many other countries have established strict regulations for aflatoxins in food and feeds [2]. The FDA has established an action level for total aflatoxins in human food at 20 parts per billion (ppb) and 0.5 ppb of aflatoxin M1 in milk. The European Union has enacted even stricter action levels for imported agricultural commodities. Action levels for aflatoxins have also been set for various categories of animal feed. Unfortunately, developing countries in many regions of the world, such as Sub-Saharan Africa, cannot afford the costs associated with monitoring and mitigation of aflatoxin in food and feed crops. This has led to an increased risk of exposure to aflatoxin resulting in outbreaks of acute aflatoxin poisoning (aflatoxicosis) [3] and increased morbidity in children suffering from stunted growth and malnutrition (kwashiorkor) [4,5]. In addition to the adverse effects that aflatoxin has on human and animal health worldwide there are also significant economic costs incurred trying to mitigate aflatoxin contamination of crops. Estimates reveal direct annual crop revenue losses in the U.S. in the tens of millions of dollars and in particularly severe years of Midwestern maize contamination losses can be in the hundreds of millions of dollars [6]. Of course total costs attributable to aflatoxin are much higher when you take into account crop losses in other countries in addition to other factors such as export market losses, sampling and testing, and human and animal health effects [7].

Pre-harvest aflatoxin contamination is a very complex problem affected by a multitude of biotic and abiotic factors. Therefore, a multi-pronged approach may need to be employed to control aflatoxin contamination when conditions in the field are favorable for fungal infection. An area of intense study for the control of aflatoxin contamination is the development of preharvest host plant-resistance [8,9]. This is because A. flavus infects affected crops prior to harvest and a host-resistance strategy may be the easiest for the grower to integrate into the various crop management systems to prevent preharvest contamination with aflatoxins. Several maize lines have been identified and developed with increased resistance to A. flavus infection and aflatoxin contamination and this has enabled the identification of natural resistance traits [10,11,12,13]. However, these investigations have indicated that resistance to aflatoxin contamination is polygenic. Therefore, attempts to move resistance from inbred lines into commercial varieties with desirable agronomic characteristics has been a slow process due to the lack of availability of biomarkers to facilitate the transfer of resistance genes [14]. Unlike maize, cotton has a limited diversity of germplasm and to date no varieties have been identified with natural resistance to A. flavus. For this reason, it is critical that a seed-based resistance be developed in cotton. A number of potential maize resistance-associated proteins (RAPs) and the genes encoding them have been identified and some of these have been introduced into cotton for evaluation [15,16,17]. However, more research is needed to elucidate the biochemical mechanisms that manifest the resistance phenotype in maize kernels or other sources so that they can be utilized to enhance resistance through marker-assisted breeding in maize or genetic engineering in cotton [12,18,19,20,21].

This review will focus on the following areas of research that will be critical for successful preharvest control of aflatoxin contamination in maize and cottonseed: (1) identification of new sources of maize germplasm resistant to fungal infection and aflatoxin contamination; (2) identification of RAPs and their corresponding genes in maize kernels through comparative proteomics and genomics of resistant and susceptible maize inbreds; (3) development of practical technology for use by maize breeders based upon identification of RAP-associated proteins (and genes) as resistance markers to aid in marker-assisted maize breeding; and (4) production of genetically engineered cotton with resistance alleles of RAP genes from maize and also genes from non-native sources to enhance resistance to aflatoxin contamination.

Studies demonstrating subpericarp (wounded-kernel) resistance in corn kernels have led to research for identification of subpericarp resistance mechanisms. The examination of kernel proteins of several genotypes revealed differences between resistant and susceptible genotypes [35]. The first proteins shown to be potentially involved in corn kernel resistance were germination-induced ribosome inactivating protein (RIP) and zeamatin [36], which were also shown to be involved in inhibition in vitro of A. flavus growth. It has also been shown that both constitutive and induced kernel proteins are necessary for the expression of kernel resistance to aflatoxin production [37], and that a high level of expression of constitutive antifungal proteins actually constitutes a major difference between resistant and susceptible kernels.

In a study of protein production in corn inbred Tex6, two kernel proteins were identified which may contribute to its resistance to aflatoxin production [38]. One of the proteins, 28 kDa in size, inhibited A. flavus growth, while a second with a molecular mass greater than 100 kDa, inhibited toxin formation with little effect on fungal growth. The 28 kDa protein was identified as a unique chitinase [39]. In another investigation, an examination of kernel protein profiles of 13 corn genotypes revealed that a 14 kDa trypsin inhibitor protein (TI) is present at relatively high concentrations in seven resistant corn lines, but is present only in low concentrations in six susceptible ones [40]. The mode of action of TI against fungal growth may be partially due to its inhibition of fungal α-amylase [41], limiting A. flavus access to potential simple sugars required for toxin production [42]. Comparisons of kernel protein profiles between susceptible and resistant genotypes may shorten the time it takes to identify RAPs.

Proteomics can also enhance the identification of RAPs. Through proteomic analysis and side-by-side comparisons of constitutive kernel embryo and endosperm proteins of resistant and susceptible genotypes, unique or elevated levels of stress-related proteins were discovered in aflatoxin-resistant lines [16,43]. These proteins can be grouped into three categories based on their peptide sequence homology: (1) storage proteins, such as globulins (GLB1, GLB2), and late embryogenesis abundant proteins (LEA3, LEA14); (2) stress-responsive proteins, such as aldose reductase (ALD), glyoxalase I (GLX1) and heat shock proteins, and (3) antifungal proteins, including TI. The discovery of stress-related RAPs may be very important given the observation of drought’s enhancement of aflatoxin accumulation.

Another valuable contribution to RAP identification has been the discovery of closely-related lines from the same backcross differing in aflatoxin accumulation for proteomic analysis in the IITA-SRRC collaborative breeding program [28,29]. Investigating corn lines sharing close genetic backgrounds should enhance the identification of RAPs without the confounding effects experienced with lines of diverse genetic backgrounds [12]. Using pairs of closely-related lines, a proteomic study was conducted which confirmed the earlier identification of three categories of proteins and added a fourth category, putative regulatory proteins [44]. Regulatory proteins, not always easily seen on 2-D gels, may have been more identifiable due to the use of closely-related lines. Proteomics has been used to study corn resistance using rachis and silk tissue, as well [45,46]. Results, as might be expected, support the findings of kernel studies regarding genotypic differences in resistance.

Chromosome regions associated with resistance to A. flavus and inhibition of aflatoxin production in maize have been identified through Restriction Fragment Length Polymorphism (RFLP) analysis in three “resistant” lines (R001, LB31, and Tex6) in an Illinois breeding program, after mapping populations were developed using B73 and/or Mo17 elite inbreds as the “susceptible” parents [50,51]. Chromosome regions associated with inhibition of aflatoxin included regions on chromosome arms 2L, 3L, 4S, and 8S and these may prove promising for improving resistance through marker-assisted breeding into commercial lines. In some cases, chromosomal regions were associated with resistance to Aspergillus ear rot and not aflatoxin inhibition, and vice versa, whereas other chromosomal regions were found to be associated with both traits. This suggests that these two traits may be at least partially under separate genetic control.

Quantitative trait locus (QTL) studies involving other populations have identified chromosome regions associated with low aflatoxin accumulation. In a study involving 2 populations from Tex6 × B73, promising QTLs for low aflatoxin were detected in bins 3.05–3.06, 4.07–4.08, 5.01–5.02, 5.05–5.05, and 10.05–10.07 [52]. Environment strongly influenced detection of QTLs for lower toxin in different years; QTLs for lower aflatoxin were attributed to both parental sources. In a study involving a cross between B73 and resistant inbred Oh516, QTL associated with reduced aflatoxin were identified on chromosomes 2, 3, and 7 (bins 2.01 to 2.03, 2.08–2.09, 3.08–3.09, and 7.06–7.07) [53]. QTLs contributing resistance to aflatoxin accumulation were also identified using a population created by B73 and resistant inbred Mp313E, on chromosome 4 of Mp313E [54]. Another QTL in this study, which has similar effects to that on chromosome 4, was identified on chromosome 2. In recent studies to identify aflatoxin-resistance QTL and linked markers for marker-assisted breeding (using a population developed from an aflatoxin-resistant maize inbred, Mp717 and a susceptible NC300), QTL were identified on all chromosomes, except 4, 6, and 9 [55], and on chromosomes 1,3,5, and 10 in multiple years (4 and 9 in one year) when using a population developed using resistant inbred Mp715 and susceptible T713 [56]. In all of the above-mapping projects, no QTL was identified that contributed more than ~20% to phenotypic variation [57].

A number of RAP genes identified through comparative proteomics have been mapped to chromosomal location [12] using the genetic sequence of B73 now available online [57]. Using the DNA sequence of the RAPs and running a BLAST similarity search against the B73 sequence allowed each gene to be placed into a virtual Bin, facilitating the mapping of their exact location on the chromosome. The chromosomes involved include the above-mentioned chromosomes 1, 2, 3, 7, 8, and 10, and some in bins closely located to those described above. This adds support to proteomic data and characterization results that suggest the involvement of the 14 kDa TI, water stress inducible protein, zeamatin, one of the heat shock proteins, a cold-regulated protein, glyoxalase I, PR-10 protein and beta-1-3-glucanase in aflatoxin-resistance. In fact, heat shock protein 16.9 mapped with a QTL (bin 3.04–3.05) that accounted for a relatively high level (16%) of phenotypic variation [45]. From the above QTL investigations, it is observed that variation can exist in the chromosomal regions associated with Aspergillus ear rot and aflatoxin inhibition in different mapping populations. This suggests the presence of different genes for resistance in the different identified resistant germplasm. It will be important to map resistant lines investigated through proteomics or to obtain data from associative mapping panels regarding gene location.

Conventional breeding in maize has delivered much-improved cultivated crops through enhancement of agronomic traits such as disease and stress resistance. These successes have been possible because of the availability of desired resistance genes in the maize germplasm [8,12,58]. A genetic engineering approach in cotton is warranted as natural resistance to mycotoxin-producing, saprophytic fungal pathogens such as A. flavus has not been identified in the germplasm base. For this reason, transgenic cotton varieties with antifungal traits that confer resistance to A. flavus will be extremely valuable. Available literature on transgenic crops exhibiting microbial resistance is mostly on bacterial or viral resistance. In fact, no fungal resistant crops have yet been deregulated in spite of several field tests, underlying the complexity of host-pathogen relationships. This gets more confounded by the fact that A. flavus is a saprophyte, and not a typical plant pathogen that demonstrates a gene for gene relationship. Success of a genetic engineering approach in developing cotton with increased resistance to A.flavus infection and aflatoxin contamination will depend in large part on identification of resistance genes, native or foreign, with target specificity that express inhibitory activity against aflatoxigenic fungi [59,60,61]. Table 1 summarizes information on a number of candidate proteins/peptides from maize and other sources that have been tested for efficacy against A. flavus as well as other fungal plant pathogens.

Outbreaks of severe aflatoxin contamination of maize and cottonseed are inevitable and appear to be caused in large part by stress on the host plant usually in the form of drought and/or insect pressure. It is likely that significant control of aflatoxin contamination will require a multipronged approach that utilizes biological control and improved agronomic practices as well increased resistance by the host plant arising from either marker-assisted breeding in maize or genetic engineering of cotton. To this end, the identification of natural resistance traits to aflatoxin accumulation in maize genotypes has provided an inroad to the development of a host resistance strategy in which genes encoding resistance associated proteins can be utilized as molecular markers for transfer of aflatoxin resistance traits into elite maize varieties. Much work has been accomplished on the identification of maize genotypes demonstrating increased resistance to aflatoxin contamination and now technologies such as proteomics and genomics are being utilized to identify the proteins/genes that contribute to the observed resistance. Of equal importance in selecting candidate resistance-associated genes will be to determine what effects stress has on their expression and at what developmental stage and in what tissues they are being expressed so as to maximize their efficacy against A. flavus.

Development of cotton with enhanced resistance to aflatoxin contamination will be more problematic. The lack of genetic diversity in the germplasm renders conventional breeding approaches challenging as no natural resistance to aflatoxin accumulation has been identified in cotton. This is why it is imperative that genes encoding resistance-associated proteins from maize be identified and introduced into cotton by transgenic approaches in an effort to control A. flavus growth and aflatoxin production. Knowing that resistance in maize is multigenic it is very likely that improved resistance in cotton will require engineering of cotton with multiple resistance genes. However, sources of resistance for development of transgenic cottons do not have to originate solely from maize and this review has presented information on the efficacy of a number of potential antifungal proteins/peptides from sources other than maize against A. flavus. Of particular interest are the small, lytic, antifungal peptides, both natural and synthetic. Utilizing automated peptide synthesis chemistry it is relatively simple to have a peptide synthesized for use in in vitro assays against A. flavus. If necessary, promising natural antifungal peptides can be modified to increase their potency and specificity toward A. flavus utilizing rational design approaches. Because many of these antifungal peptides act via membrane permeabilization, there is less likelihood that A. flavus or other pathogens will be able to develop resistance to these peptides. Bioassays in our labs have indicated that A. flavus usually requires the highest level of antifungal protein/peptide concentrations to inhibit its growth compared to other fungal pathogens. Therefore, identification of a single or combination of genes encoding protein/peptides that control A. flavus growth when expressed in maize or cotton will probably be effective against a number of fungal pathogens that infect food and feed crops.

Another, as of yet unexplored, transgenic approach for aflatoxin resistance in maize and cotton is the use of RNA interference (RNAi) as a means of downregulating expression of genes vital to fungal growth and toxin formation. In this scenario, the transgenic plant would be engineered with vectors for the expression of self-complementary hairpin RNAs of antifungal/anti-aflatoxin genes that will result in production of small interfering RNAs (siRNAs) by the host plant’s DICER-like enzymes. This technique has been successfully demonstrated in planta for silencing of gus gene expression in Fusarium verticillioides interacting with transgenic tobacco generating gus siRNAs [99]. Critical to the success of this approach will be the ability of the invading fungus to efficiently take up the host-plant generated siRNAs in planta to activate the machinery for silencing of the targeted gene.