1. Introduction
Common bean (
Phaseolus vulgaris L.) originated in Mexico 4 to 6 million years ago [
1] and was independently domesticated in Mesoamerica and the Andes 8000 years ago, constituting two main known gene pools [
2]. Beans refers to legumes of the genus
Phaseolus, family Fabaceae, subfamily Papilionoideae, tribe Phaseoleae, and subtribe Phaseolinae [
3]. Common bean (
Phaseolus vulgaris L.) is a diploid (2n = 2× = 22), annual, predominantly self-pollinating species and one of the most important pulses worldwide [
4,
5].
Based on nucleotide sequences of chloroplasts, patterns of phaseolins, and genetic signatures in domesticated and wild accessions, it appears that the greatest genetic variation occurs among genotypes of the Mesoamerican gene pool, the most preferred type of bean for consumption in Brazil [
6,
7,
8,
9]. According to the FAO (Food and Agriculture Organization of the United Nations), global production of dry bean in 2018 was approximately 31.5 million tons. Brazil is considered the third largest producer in the world, with production of approximately 3.1 million tons [
10]. The favorable edaphic and climatic conditions for growing common bean in Brazil allow wide distribution in every Brazilian state, with different harvest seasons, which is key for an annual supply [
11].
An increase in planted area, especially under irrigation, combined with multiple crop seasons, has created conditions for high incidence of soil diseases, which are among the main causes of low crop yield and considerable losses [
12]. One of the main fungal diseases, Fusarium wilt, is a severe vascular disease in common bean whose causal agent is
Fusarium oxysporum Schlecht. f. sp.
phaseoli Kendrick & Snyder (
Fop) [
13,
14]. The infection process begins in the roots, colonizes the xylem, and causes leaf wilt, vascular discoloration, chlorosis, dwarfism, and premature plant death [
15,
16].
Therefore, it is necessary to identify potential bean sources of resistance to effectively control the pathogen. The development of resistant cultivars is a promising alternative for control of this disease as resistant cultivars are easily adopted by producers and do not cause environmental risks [
17,
18]. Pathogenicity testing through inoculation methods can be used to characterize the degree of pathogenicity of
Fop strains [
19]. Pathogenicity testing also provides an alternative for assessing the diversity of physiological races of the pathogen, the main cause of breakdown in genetic resistance to
Fop in bean cultivars [
20,
21]. Currently, several definitions describe the complexity of genetic resistance of bean to
Fop; some studies report it to be monogenic [
22,
23], some as oligogenic [
24,
25,
26], and another as polygenic [
17].
In common bean, only a limited number of studies have been conducted with the goal of clarifying the comprehension of the molecular mechanisms and pathways involved in bean response to
Fop’s infection [
27]. Recent results demonstrate by Chen et al. [
27] using whole transcriptome and metabolome of common bean infected by
Fop shows the response to
Fop uses different and effective defense pathways comprising of a complex resistance network of structural, signaling, hormonal and chemical responses.
An alternative for understanding genetic control of bean resistance to
Fop, how resistance loci are distributed in the bean genome, and the intensity of their effects is to study them indirectly, through their association with genetic markers [
28,
29]. Among the preferred genetic markers, single nucleotide polymorphisms (SNPs) are noteworthy, since these markers can be integrated with the QTN (Quantitative Trait Nucleotide) responsible for phenotypic variation in the trait of interest [
30].
The GWAS using natural populations have higher mapping resolution than linkage mapping and greater cost-effectiveness [
31]. In GWAS, the recombination events accumulated over innumerable generations reduce linkage disequilibrium (LD), allowing more precise estimates of the location of genes of interest to be obtained [
32]. In recent years, GWAS has been widely used to investigate the genetic architecture of complex characteristics in model plants such as Arabidopsis thaliana [
33], soybean [
34,
35], and common bean [
36,
37].
In the current study, a total of 2001 high-quality single-nucleotide polymorphisms (SNPs) distributed over the 11 bean chromosomes were genotyped using SNP Assay technology (Illumina BARCBean6K_3 BeadChip) [
38]. The BARCBean6K_3 BeadChip has successfully contributed to the study of several traits in beans [
39,
40,
41,
42,
43,
44]. This technology has also been used in the identification of genomic regions associated with disease resistance, such as common mosaic virus [
45], anthracnose [
41], root rot [
46], rust [
42], angular leaf spot [
47], and to fusarium wilt [
25].
The aim of the present study was to identify new genomic regions associated with Fop resistance in a Mesoamerican Diversity Panel (MDP) and, taking into account the phenotypical evaluations, identify potential genotypes of common bean as sources for Fop’s resistance.
4. Discussion
The success of association mapping in identifying markers effectively associated with the trait depends on how well the population structure is corrected in the association model and on the existing levels of LD [
76]. In a bean population, using a kinship matrix containing the population structure has been widely used in genome-wide association studies, successfully correcting the genetic relatedness between individuals using linear mixed models [
48,
77]. For association mapping in common bean, gene pools should be considered separately, because LD decays more rapidly within the Andean gene pool and is stronger within the Mesoamerican gene pool [
5,
78].
Regarding the Mesoamerican panel, the parameters observed in the current study agreed with those presented by [
37], who evaluated a Mesoamerican carioca (cream-colored seed coat with brown stripes) panel, which is, in fact, part of the MDP used. The BIC test was performed for the first five components, and no PCs were required for any of the traits. The formation of haplotypic blocks within the LD markers ranging from 0.03 Mb (Pv05) to 1.01 Mb (Pv01) indicated that the markers evaluated represent the possible constituent haplotypes in the Mesoamerican panel [
79].
Fop is genetically variable and often found in common bean growing in different countries and regions; up to now, seven pathogenic races related to geographical regions are cited in the literature [
20,
80,
81], and new races like UFV01 and IAC18001 occur, supporting
Fop pathogenic evolution [
82]. However, mutations and recombination between avirulence genes (avr) in sexually reproducing pathogens are postulated as the mechanisms responsible for variation in races [
83].
Resistance genes can be overcome by new or more virulent races; hence, broad-spectrum, durable resistance is needed [
84]. In the current study, only 75 accessions (36.58%) showed resistance to both strains of
Fop, demonstrating the difficulty of obtaining genotypes with resistance to different races of the fungus. Sala et al. (2006) evaluated 104 bean genotypes, of which 33% were resistant to the
Fop 1, 2, 3, and 4 races, indicating the difficulty of finding cultivars with multiple resistance to the pathogen. Leitão et al. [
25] evaluated a panel containing predominantly Andean accessions and the
Fop race 06 and observed only 14 accessions (9.27%) with resistance to the fungus, with heritability values from 40.8% to 71.5% considering the DSR and AUDPC parameters (49% and 63%).
Important SNPs associated with QTL (Quantitative Trait
loci) in the current study were associated with
Fop resistance (represented by the parameters DSR and AUDPC) for the two strains tested (UFV01 and IAC18001). The differences reflect the varied resistance spectra exhibited by these accessions. Despite the experiments with both strains being conducted in few experiments under controlled conditions, some of the QTL identified in this study are confirmed by the literature, evidencing the robustness of results. However, the successful establishment of disease by the
Fop pathogen demands a response in the plant defense system, and the entire molecular mechanism of pathogenesis remains to be elucidated to improve selective accuracy with additional experiments involving high-throughput phenotyping [
85,
86].
In bean,
Fop penetrates the epidermis of the plant roots, invades the cortex, and colonizes the vascular tissue of the host plant, causing obstruction and wilting [
14,
15]. Pathogens other than
Fusarium spp. can cause wilting in legumes; pathogens such as
Rhizoctonia spp.,
Verticillium spp., and
Aphanomyces euteiches [
87]. Gupta et al. [
88] confirmed that genes associated with the secondary cell wall are involved in the combined response of the plant to infection from wilt pathogens and to drought in
Arabidopsis thaliana.Furthermore, since we are likely dealing with polygenic inheritance with small additive genetic effects, increasing the sample size, thus maximizing the phenotypic diversity among the MDP, would enhance the power to recover meaningful associations [
23,
25]. Most of the SNPs associated by GWAS revealed that the genomic regions linked to
Fop traits were located inside or near the candidate genes on Pv01, Pv03, Pv04, Pv05, Pv07, Pv10, and Pv11 (
Table 2 and
Table 3).
The Pv01 chromosome also showed a significant SNP, ss715649713, associated with DSR for the IAC18001 strain at the 1.01 Mb LD haplotype block, positioned within the Phvul.001G074800 (Appr-1-p processing enzyme family protein) gene. Appr-1-pase is an important and ubiquitous cellular processing [
89]. Ubiquitination is a known mechanism in the regulation of plant defense against pathogens [
90]. Recent evidence shows that ubiquitination plays a critical role in regulating plant responses to abiotic stresses and plant tolerance of adverse environmental conditions [
91]. The ubiquitination mechanism may also be associated with actions on specific components for stress signaling [
92].
On Pv03, two significant SNPs associated with the
Fop reaction were found, the ss715647339 (IAC18001) and ss715648884 (UFV01) positioned at a distance of 1.01 Mb, and showed potential candidate genes involved in root development mechanisms (Phvul.003G258100) and in presumed disease-resistance proteins (Phvul.003G258700, Phvul.003G258800, and Phvul.003G260300). The Phvul.003G258400 gene is associated with the putative Cytochrome P450 superfamily protein also in this region family members can act in the control of abscisic acid (ABA) production that are involved in critical processes in plant growth and development. They can also act in biotic and abiotic stress responses [
93,
94] and the formation of secondary metabolites, such as terpenoids, flavonoids, steroids, alkaloids, phenylpropanoids, glucosinolate, and cyanogenic glycoside all of which are typically made as part of host defense [
95].
The SNP ss715648681 identified on Pv04 associated with AUDPC for the IAC18001 strain is positioned within the Phvul.004G001900 gene (MATE efflux family protein). In plants, MATE transporters have been directly or indirectly implicated in mechanisms of disease resistance [
96], in the transport of diverse types of secondary metabolites, such as alkaloids [
97], flavonoids [
98,
99], anthocyanidins [
100], and hormones, such as salicylic acid (SA) and ABA, and in drought tolerance [
101]. Mandal et al. [
102] demonstrated that the induced resistance observed in tomato against
Fusarium oxysporum f. sp.
lycopersici (Fol) might be a case of salicylic acid-dependent systemic acquired resistance.
Another significant SNP, ss715645397, was found in Pv05 associated with AUDPC for UFV01 at 0.004 Mb from the Phvul.005G152600 gene (ARM repeat superfamily protein). The Armadillo (ARM) domain has motifs with the structure of repeat proteins, such as Leucine-rich repeats (LRR), that have been extensively studied in plants, suggesting a critical role of these repeating peptides in plant cell physiology, plant stress, and plant development [
103]. In this region close to the marker, Nakedde et al. [
46] identified a QTL mapped in a recombinant inbred line (RIL) population that accounted for 9.20% to 10.06% of phenotypic variation associated with Fusarium Root Rot (FRR) and root architecture traits. This QTL was located at 39.22 Mb in a 0.31 Mb interval on Pv05.
Another candidate gene associated with the ss715646169 marker positioned at 1.99 Mb on Pv05 (between 0.0 Mb and 0.56 Mb) for DSR and AUDPC of the IAC18001 strain. This marker was positioned within the Phvul.005G022100 gene (Cellulose synthase family protein). The cellulose synthase (CesA) superfamily genes are among the most important agents involved in the biosynthesis of plant cell walls, which are mainly composed of biopolymers such as celluloses, hemicelluloses, pectins, and lignins [
104]. Among the several defense mechanisms in the plant–pathogen resistance interaction, structural changes must be highlighted. These structural changes lead to strengthening of the plant cell wall by the deposition of callose, followed by lignification, a phenomenon that can be determinant in a resistance or susceptibility reaction in interaction with
Fusarium oxysporum, with the possibility of quantitative differences in response [
105].
Our results showed a group of candidate gene associated with the ss715646169 marker are the genes related to the zinc finger domain (Phvul.005G016200; Phvul.005G019900; Phvul.005G020000 and Phvul.005G022000). Zinc finger proteins play a crucial role in many metabolic pathways, as well as in stress response and defense in plant-pathogen interactions to the defense of plants, and may be associated with a JA-dependent defense pathway [
106,
107]. The SNP ss715647730 identified on Pv07 and associated with AUDPC for IAC18001 was positioned at 0.01 Mb from the Phvul.007G199600 gene (drought-responsive family protein). Although drought-responsive proteins exhibit various patterns depending on plant species, genotypes, and stress intensity, proteomic analyses show that dominant changes occurred in sensing and signal transduction, reactive oxygen species scavenging, osmotic regulation, gene expression, protein synthesis/turnover, cell structure modulation, and carbohydrate and energy metabolism [
108].
Leitão et al. [
25] performed association mapping for
Fop race 06 using a panel of 133 common bean accessions from Portugal and observed significant associations detected for DSR and AUDPC on the Pv04, Pv05, Pv07, and Pv08 chromosomes. They noted that the DART03480 marker on Pv04 was at a small distance of approximately 0.1 Mb from the ss715648681 marker, which was also detected in our study.
The Pv10 chromosome showed a significant SNP, ss715645508, positioned at a distance of 0.001 Mb from the Phvul.010G137000 gene (SNARE-like superfamily protein). This gene may be considered a novel determinant of salinity/drought tolerance and a potential candidate to increase salinity and drought tolerance in crop plants [
109]. Erfatpour et al. [
110] identified a QTL in this same genomic region between 39.97 Mb and 40.29 Mb, with forty candidate genes associated with non-darkening (ND) in seed coat color at 1.6 Mb from the significant marker in our study.
Linkage mapping reported genomic regions associated with
Fop resistance to race 04 [
23]. The authors identified significant markers positioned on Pv03, Pv10, and Pv11, and a QTL of greater effect that explained 63.5% of the phenotypic variance on Pv10. A SCAR marker (U20.750) linked to this QTL was developed, with evaluation in Andean and Mesoamerican germplasm, and the marker had high accuracy in Mesoamerican accessions [
111].
Gene annotation allowed the identification of candidate genes associated with putative effects in disease-resistance mechanisms (R), such as a cluster of 20 candidate genes annotated as “leucine-rich repeat-containing protein” (LRR), with distances from 0.03 Mb from the Phvul.011G200300 gene up to 0.39 Mb from the Phvul.011G203100 gene positioned close to the ss715648096 marker on Pv11 associated with DSR and AUDPC for UFV01 (
Table 4). The region of 51.50 Mb associated with the significant ss715648096 marker on Pv11 corroborates previous studies, and the region being associated with other important fungal diseases of common bean, such as anthracnose, by the association of marker S11_51790295 to race 73 of
Colletotrichum lindemuthianum (the anthracnose pathogen), positioned at a distance of approximately 0.20 Mb [
112]. The identification of LRR receptor-like protein kinases (PK) and their role in adaptive selection supports prior literature indicating a co-evolution of common bean and the anthracnose fungus [
44,
113].
The GWAS of the Mesoamerican panel also revealed the S11_50585184 marker at 0.91 Mb from the ss715648096 marker associated with
Fop that is related to the Phvul.011G192400 (NBS-LRR with typical NB-ARC domain) gene associated with
Rhizoctonia solani resistance on Pv11 [
36]. The response to different soil diseases may be because the NB-ARC domain contains a functional ATPase region that regulates the resistance, and this domain interacts with the nucleotide-binding domain in order to exchange the nucleotides that are associated with activating ATPase change, which, in turn, reshapes to NB-ARC ATPase and alters resistance specificity and the possibility that the LRR interacts with similar elicitors from both pathogens [
114,
115].
Hoyos-Villegas et al. [
116] used the GWAS procedure for wilting score associated with drought-tolerant genotypes and reported one significant association at the SNP ss715639678, which is located at the end of Pv11, in a region that was found to be in high LD, with 1131 genes. In addition, gene ontology enrichment analysis revealed 19 biological processes and 30 molecular functions that were significantly associated. Myers et al. [
117], using GWAS for finding markers associated with total phenolic content (TPC), identified 11 QTNs linked with TPC, especially the SNP ss715650328 at 52.96 Mb on Pv11. Various biological functions may be related, including the production of compounds such as phenolic acids, flavonoids, and proanthocyanidins, which are the main polyphenols associated with plant defense and postharvest darkening in common bean [
118,
119].
The physical barriers that act at different levels in defending plants inhibit the penetration and colonization of plant tissues by the pathogen, associated with biochemical reactions in the host cells that produce toxic substances and/or create adverse conditions for growth of the pathogen inside the plant. Therefore, substances produced in the host cells, before or after infection, contribute significantly to resistance [
120].
Some signaling components, such as phytohormones, combined with functional gene transcription factors and their regulators, are involved in responses to combined abiotic and biotic stresses in plants, factors that can be modulated according to environmental conditions [
121]. The effect of water can modulate the response of the plant to pathogens, in which several pathogens translocate virulence proteins (effectors) into host cells to target different components of the plant [
122].
Chen et al. [
27], using whole transcriptome and metabolome, showed bean-
Fop pathosystem includes different and effective defense pathways comprising of a complex resistance network of structural, signaling, and chemical responses. The authors demonstrated the validation of differentially expressed genes located in Pv03, Pv04, Pv07, Pv08 and Pv11 by qRT-PC showing strong roles in signaling routes such as salicylic acid (SA), jasmonate, and ethylene.
Fop also induced the flavonoid biosynthesis pathway which was the most significantly enriched one in response to
Fop’s infection.
Xue et al. [
123] using the cDNA amplified fragment length polymorphisms (cDNA-AFLPs), found five transcript-derived fragments involved in the mechanism of plant hormone regulation. These five genes belonged to the jasmonate, auxin, Abscisic acid (ABA), and SA-dependent pathways can be implicated to play a role in the plant’s defense responses.
After exposure to the pathogen, the plant starts a signaling network mediated by protein kinases, such as mitogen-activated protein kinases (MAPK) and begins a process of recognition of pathogen-associated molecular patterns (PAMPs) through their PAMP-recognition receptors (PRRs), known as pattern-triggered immunity (PTI) and pathogen effector-triggered immunity (ETI), two important mechanisms for averting disease attacks [
124].