1. Introduction
As an important source of edible oil and protein for humans and livestock, the peanut (
Arachis hypogaea L.) is widely grown in more than 100 countries, with China, India, and the United States being the largest producers [
1]. Cultivated peanuts (
A. hypogaea, AABB, 2n = 4x = 40) are believed to be derived from the diploid wild species
A. duranensis (AA genome) and
A. ipaensis (BB genome), originating in South America through heterologous hybridization and long-term domestication [
2]. Peanuts are widely planted in China as an important oil crop. In recent years, China has witnessed continuous enhanced peanut productivity, which reached 17.33 million tons in 2018 [
3]. The peanut has become the largest oilseed crop in terms of annual production for more than one decade; however, peanut farmers and industries in China, as well as in many developing countries, are facing serious challenges of an increased risk of aflatoxin contamination.
Aflatoxin B1, B2, G1, and G2 (AFB
1, AFB
2, AFG
1, and AFG
2) are the most toxic and carcinogenic naturally occurring mycotoxins produced by
Aspergillus in peanuts.
Aspergillus flavus is able to produce AFB
1 and AFB
2, while
Aspergillus parasiticus can produce all four aflatoxins [
4,
5,
6]. Restrictions on levels of aflatoxins in agricultural and food products have been set up in many countries in the world. China and the U.S. have set limits on levels of aflatoxin at 20 μg/g in food and feed, while the European Union has imposed stringent regulations on levels of aflatoxin at 2–4 μg/g [
7]. Aflatoxin contamination can directly affect food safety, international trade, and market competitiveness of peanut products, resulting in enormous economic losses. Various measures, such as bio-control agents, good agricultural practices, and genetic improvement in host plants have been used to prevent and control aflatoxin contamination in peanuts [
8,
9,
10,
11]. It is well-known that risk of aflatoxin contamination can be effectively controlled by planting resistant peanut varieties combined with necessary crop management [
12].
However, efforts at breeding for aflatoxin resistance in peanuts are highly limited by many factors, such as complex interactions among the peanut varieties, Aspergillus fungal strains, and environmental factors. Lack of desirable parental genotypes with reliable and high-level resistance has been a common constraint in peanut breeding programs. Besides, a lack of rapid and accurate methods for phenotypic identification of aflatoxin contamination is also a common technical drawback worldwide. Therefore, more precise identification of peanut materials resistant to aflatoxin contamination and development of molecular markers applicable for marker-assisted selection (MAS) in breeding are of great significance to reduce the risk of aflatoxin.
According to the flowering and branching characteristics, peanut germplasms have been divided into six botanical varieties or types. Peanut germplasm accessions and breeding lines, including GFA-1, GFA-2, AR-3, AR-4, Yueyou 9, and Zhonghua 6 were identified to be resistant to aflatoxins [
13,
14]. However, the improved peanut varieties with aflatoxin resistance that are available have only been used in relatively limited regions, due to their relatively low yield or other undesirable agronomic traits. In order to develop better aflatoxin-resistant peanut varieties, more systematic identification is needed to screen germplasms for resistance. Although over 8000 peanut germplasm accessions have been assembled in China, it is difficult to screen all of them for complex traits like aflatoxin resistance, which can only be accessed using complicated and dear cost method. Under such circumstances, phenotyping for aflatoxin resistance in the manageable peanut germplasm set, like the Chinese mini-mini core collection consisting of 99 diverse accessions, would be feasible [
15].
The molecular mechanisms of the biosynthesis of aflatoxins have been well-investigated. The initial stage of aflatoxin biosynthesis is similar to that of fatty acid biosynthesis, where acetyl CoA acts as the start unit, and malonate monoacyl CoA acts as the extension unit to form aflatoxin’s polyketone skeleton, catalyzed by polyketide synthase (PksA) [
16,
17]. More than 18 enzymes, such as Nor-1 (oxidoreductase), AvnA (monooxygenase), AdhA (dehydrogenase), FAD (Flavin adenine dinucleotide)-containing monooxygenase, EstA (esterase), and VerB (versicolorin B desaturase) have been identified to be involved in the biosynthesis of aflatoxins [
18]. Advances in plant proteomics and fungal genomics partly reveal the resistance mechanisms of aflatoxin contamination in host plants [
19]. Three types of plant factors that may influence fungal growth and aflatoxin contamination have been involved in the processes of resistance mechanisms: The first type is seed proteins that act as fungal cell wall degrading enzymes; the second type is proteins or secondary metabolites from the host seed that could directly influence fungal growth and/or aflatoxin synthesis; and the third type is plant stress responsive proteins that are synthesized by the host in large amounts after infection by
A. flavus [
20]. Research on these resistance mechanisms may promote genetic improvement of aflatoxin contamination resistance in peanuts.
Advances in next-generation sequencing (NGS) have made genome-scale population genetic studies more straightforward and economical [
12,
21,
22,
23,
24,
25]. Restriction-site-associated DNA sequencing (RAD-Seq) is a fractional genome sequencing strategy that can identify large numbers of genetic variations, such as single nucleotide polymorphism (SNP), through sequencing genomes digested by restriction nuclease [
26]. In peanuts, SNP markers obtained by RAD-seq have been successfully used for genetic linkage map construction and bulk segregant analysis (BSA) [
27,
28,
29]. Resistance to aflatoxin accumulation in peanuts is a complex trait affected by several environmental factors. Strong interaction has been detected between environment factors and the genotype for aflatoxin contamination [
20]. In our previous study, four major and six minor QTLs (quantitative trait locus) were identified for aflatoxin resistance through genome-wide QTL analysis with a genetic linkage map constructed by SSR (simple sequence repeats) markers [
30]. Due to the limitation of the genetic linkage map’s resolution, the confidence interval’s genetic distance of these major QTLs was >7 cM (centimorgan). It is difficult to identify related candidate genes by such a large confidence interval. However, the genome-wide association study (GWAS) is an effective trait-mapping approach for identifying candidate genes that underlie complex phenotypic traits, and it has been applied in identifying associated markers and candidate genes for several important agronomic traits in peanuts [
15,
19,
31], but has not yet focused on aflatoxin resistance. Therefore, the present study was performed to conduct GWAS in 99 accessions of the Chinese peanut mini-mini core collection, using RAD-Seq-based, high-density genotyping and phenotyping data for AFB
1 and AFB
2 contents from three environments via inoculation with
A. flavus under laboratory conditions. This study successfully identified SNP markers and candidate genes associated with aflatoxin content in peanut seeds, which may open up further opportunities in developing a genomic solution for aflatoxin production resistance in peanuts.
3. Discussion
Occurrence of aflatoxin contamination in peanuts can have serious impacts on economies and human health worldwide. Development of peanut varieties with aflatoxin contamination resistance through genetic improvement is the most efficient solution to reduce risks of aflatoxin. This study is the first report to systematically identify associated molecular markers for aflatoxin accumulation resistance in peanuts. A total of 18 association peaks were identified for AFB
1 and AFB
2, with PVE ranging from 16.87% to 31.70%, including four peaks specific to AFB
1, and two peaks specific to AFB
2 (
Table S2). According to annotation information of two ancestor wild
Arachis species of cultivated peanuts (
A. duranensis and
A. ipaensis), 99 candidate genes were identified in 15 candidate genomic regions (
Table S3).
Significant effects were detected in the genotype, environment, and genotype × environment interaction by a two-way ANOVA (
Table 2). Several research reports indicated that the content of aflatoxin in peanut seeds was greatly affected by environmental factors during seed development. Fountain et al. (2017) pointed out that high temperature and drought stress had significant effects on the interaction between maize and
A. flavus and the production of aflatoxin [
32]. The same phenomenon was also observed in peanuts [
33,
34,
35]. The reduced capacity of seeds to produce phytoalexins as the seed moisture content decreases during drought environment is believed to be an important factor for aflatoxin contamination [
20]. In this study, as aflatoxin content is a trait that is highly affected by the environment, the phenotype of seeds harvested in the field varied greatly in different environments (
Table 1,
Table 2). Based on these reasons, no duplicate association SNP markers were found in different environments, neither for AFB
1 nor for AFB
2 (
Table S2). However, except for 2015AFB
2 vs. 2017AFB
2, the Pearson correlations between the same phenotype in different environments were significant (
Table S4). The selected set of peanut lines in our association panel showed stable performance for extreme phenotypes across three environments, with a 4- to 10-fold difference in aflatoxin content between resistant and susceptible genotypes (
Table 4). These results indicated that screening aflatoxin-resistant materials through rapid identification of the mini core germplasm method is effective and that the identified resistant lines can be used for breeding aflatoxin contamination-resistant varieties.
In this study, only AFB
1 and AFB
2 were detected in peanut seeds, after artificial inoculation of AF2202 (
A. flavus). The mean content of AFB
1 in each sample was about ten folds of AFB
2 (
Table 1). The value of phenotypic correlations between AFB
1 and AFB
2 ranged from 0.78 to 0.99, with there being a significant level (
p < 0.01) (
Table 3). As a result, of the 19 SNP peaks detected in GWAS, 12 peaks were detected to be associated with both AFB
1 and AFB
2 (
Table S2). It is noteworthy that both AFB
1 and AFB
2 are downstream products of versicolorin B (VB) in the aflatoxin metabolic pathway of
A. flavus [
36]. These results indicated that the mechanism of aflatoxin resistance in peanuts occurs mainly in the upstream metabolic pathway of VB production. Besides, in the AFB
1 synthesis pathway, versicolorin A (VA) is the downstream product of VB, which contains the 2, 3 double bond in the dihydrobisfuran ring. This double bound can be oxidized in the host organism [
18]. These studies could explain why the content of AFB
1 is much higher than that of AFB
2 in each sample of this study.
GWAS is considered to be an efficient genetic analysis method for complex traits. There are several elements that affect the precision of GWAS, such as population size, marker density, and statistical methods. In theory, an association population with large number accessions that encompasses the genetic diversity as much as possible is an optimal choice for association analysis. However, the operation of this experiment is relatively tedious, and systematic errors are easily introduced in seed sample selection, pre-treatment, artificial inoculation, and liquid chromatography measurements during the use of large populations. Hence, a mini core collection is a useful strategy [
15]. The population structure and relative kinship were calculated by the genotype of association panel, and then used in five GWAS statistical methods to control false-positive results. Similar to other studies using peanuts [
37,
38], the 99 accessions of the Chinese mini-mini core collection in this study were classified into two subgroups (
Figure 3).
Breeding for peanut varieties for resistance to aflatoxin contamination is the most effective approach for reducing hazards by aflatoxin. In a previous study, major QTLs for aflatoxin contamination resistance in peanuts were identified in chromosome A07 and B06 by linkage analysis with a genetic linkage map developed by a SSR marker [
30]. However, because of the limited density of SSR markers in the genetic map, the confidence intervals of these QTLs are very large (7–16 cM). It is difficult to identify related resistance candidate genes by such a large confidence interval. In this study, a total of 18 association peaks distributed in 11 chromosomes were identified as associated with aflatoxin content in peanut seeds across multiple environments, 12 of which were associated with both AFB
1 and AFB
2, four were associated with AFB
1, and two were associated with AFB
2 (
Table S2). These results imply that the resistances to AFB
1 and AFB
2 in peanut seeds largely share the same mechanism controlled by multiple genes. In plants, proteins with a leucine-rich repeat (LRR) domain have been confirmed to be involved in processes of plant disease resistance. These proteins work as the first point of pathogen defense, where the innate immune response is initiated through the sensing of pathogen-associated molecular patterns (PAMPs). Two genes encoding a LRR domain were identified in chromosome B01 in a 50 kb candidate genomic region of the SNP marker SNP19994; besides, a gene encoding WRKY transcription factor was also identified in this genomic region. The SNP marker SNP02686, which has shown the largest PVE for AFB
2, was also associated with AFB
1 in the 2017 environment. The Aradu.WAPPM gene was located 33.18 kb from SNP02686 in chromosome A02, which was predicated to encode ATP-citrate lyase (ACLY), responsible for generating cytosolic acetyl-CoA and oxaloacetate (
Table S3). In
A. flavus, acetyl-CoA is the key substrate of the aflatoxin biosynthesis pathway. Although there is no direct evidence that acetyl-CoA produced in peanuts is involved in the biosynthesis of aflatoxin in
A. flavus, the findings of this study can provide ideas for further research. Peanut accessions with “AG” (contained in Zh.h0551 and Zh.h2150) and “GG” genotypes of either SNP19994 or SNP02686 possessed significantly lower aflatoxin content than that of the “AA” genotype. Further study of these markers may contribute to the development of aflatoxin production diagnostic molecular markers.