Skip to main content
  • Research article
  • Open access
  • Published:

Association of single nucleotide polymorphic sites in candidate genes with aggressiveness and deoxynivalenol production in Fusarium graminearum causing wheat head blight

Abstract

Background

Fusarium graminearum sensu stricto (s.s.) is an ubiquitous pathogen of cereals. The economic impact of Fusarium head blight (FHB) is characterized by crop losses and mycotoxin contamination. Our objective was to associate SNP diversity within candidate genes with phenotypic traits. A total of 77 F. graminearum s.s. isolates was tested for severity of fungal infection (= aggressiveness) and deoxynivalenol (DON) production in an inoculated field experiment at two locations in each of two years. For seven genes known to control fungal growth (MetAP1, Erf2) or DON production (TRI1, TRI5, TRI6 TRI10 and TRI14) single nucleotides polymorphic sites (SNPs) were determined and evaluated for the extent of linkage disequilibrium (LD). Associations of SNPs with both phenotypic traits were tested using linear mixed models.

Results

Decay of LD was in most instances fast. Two neighboring SNPs in MetAP1 and one SNP in Erf2 were significantly (P < 0.05) associated with aggressiveness explaining proportions of genotypic variance (p G ) of 25.6%, 0.5%, and 13.1%, respectively. One SNP in TRI1 was significantly associated with DON production (p G = 4.4).

Conclusions

We argue that using the published sequence information of Fusarium graminearum as a template to amplify comparative sequence parts of candidate genes is an effective method to detect quantitative trait loci. Our findings underline the potential of candidate gene association mapping approaches to identify functional SNPs underlying aggressiveness and DON production for F. graminearum s.s populations.

Background

Fusarium head blight (FHB) is a destructive disease to cereals including wheat and barley. Fusarium graminearum (teleomorph Gibberella zeae) is considered to be the main causal agent of this disease in addition to other species such as F. culmorum [1]. F. graminearum complex has been subdivided into several cryptic species [2], the main member in Germany is F. graminearum sensu stricto (s.s.) [3]. FHB disease leads to prematurely bleached spikes in infected plants [1] with considerable yield losses and contamination by mycotoxins such as deoxynivalenol (DON) [4]. DON is the most common food and feed contaminant in Europe and the maximum permissible level in unprocessed wheat used for food is 1.25 mg kg-1 [5].

The quantitative ability of an isolate to cause disease on a susceptible host plant in a non-race specific pathosystem is defined as aggressiveness [6]. Aggressiveness is an important factor determining the potential ability of an isolate to cause yield losses. Large genetic variation of aggressiveness, type of mycotoxin, and DON production was found among isolates sampled in the same country or even the same field [711]. The molecular causes of this tremendous variation in aggressiveness and DON production in F. graminearum are still unclear. To date, only one QTL (quantitative trait locus) study from a single cross was reported detecting at least two QTL for aggressiveness [12]. It was shown before, that a large amount of molecular variation was found within field populations of F. graminearum s.s. using SSR (simple sequence repeat) markers. Comparing 12 individual field populations each consisting of 30 isolates, Talas et al. [3] reported that 71% of the molecular variation was assigned within populations and 29% between populations. A cause for this high diversity might be the mating system. F. graminearum is a haploid, homothallic fungus and propagates asexually by conidia and sexually by ascospores allowing both selfing and outcrossing [7]. With the availability of the full genomic sequence of F. graminearum [13], it is now possible to deeper analyze this variation by a candidate gene approach. An array of candidate genes, including components of transcription, signal transduction, host-specific nutrition, host infection/colonization and trichothecene biosynthesis is available in the version 3.2 of the Pathogen-Host Interactions database [14]. The candidate genes we were aiming for are known to control DON biosynthesis and/or aggressiveness (TRI1, TRI5, TRI6, TRI10, TRI14) [15, 16] or are expected to have a link to aggressiveness but are yet uncharacterized for Fusarium spp. (MetAP1, Erf2) [17, 18]. Candidate gene association mapping is a sensitive tool if the mapping resolution is high enough [19]. The resolution of association mapping depends on the extent of linkage disequilibrium (LD), i.e. the non-random association of alleles present in a species [20]. Low LD was reported in F. graminearum estimated by different types of genetic makers such as VNTR, AFLP, RFLP [7, 11, 21], but no data on LD within and among genes is available. Because LD is species and population specific, it should consequently be determined before conducting an association mapping study.

The specific objectives of our study were to: (i) investigate the nucleotide diversity on gene level, (ii) investigate the extent of LD between single nucleotide polymorphic sites (SNPs) within and among candidate genes, (iii) identify SNPs of candidate genes Erf2 and MetAP1 for testing whether they affect the quantitative variation for aggressiveness, and (iv) evaluate associations of SNPs in TRI1 and other four genes of the TRI5 cluster with variation in DON production.

Results

Phenotypic analysis

Phenotypic data were analyzed in detail in a companion study [9]. All 77 F. graminearum s.s. isolates produced symptoms of FHB disease in all environments. Briefly, genotypic proportion of phenotypic variance of the isolates was significant (P < 0.01) for both traits: Aggressiveness measured as FHB rating on scale from 0 to 100 and DON production measured as DON concentration in wheat kernels in mg kg-1. Isolate × environment interaction variance was significant (P < 0.01) only for mean FHB rating but not for DON production. Entry-mean heritability was moderate for mean FHB rating (0.55) and DON production (0.62).

Histograms based on best linear unbiased estimators (BLUEs) followed a normal distribution for both traits (Figure 1). The mean of FHB infection among all isolates was 30.6% ranging from 18.4 to 38.9%. Average DON production was 12.6 mg kg-1 ranging from 3.5 to 21.6 mg kg-1. One isolate (WET24) produced a very low concentration of DON (0.3 mg kg-1) and was found to be a nivalenol producer by chemotype-specific primers.

Figure 1
figure 1

Histograms for 77 F. graminearum isolates. Best linear unbiased estimates (BLUEs) for mean FHB rating (A) and DON content (B) calculated across four environments (location × year combinations)

Analysis of population structure and diversity

Principal coordinate analysis (PCoA) based on modified Rogers' distances between all isolates did not show a distinct separation of the isolates sampled from different locations (Figure 2A). Explained variance gradually decreased according to the first ten principal coordinates (Figure 2B). The violin plot (Figure 2C) had a continuous density of distribution over all ten principal coordinates without any division within the principal coordinates. Genetic similarity ranged from 0.057 to 1.0 with a mean value of 0.31 (Figure 2D).

Figure 2
figure 2

Population structure and familial relatedness. (A) Principal coordinate analysis of the 77 isolates, based on modified Rogers' distance estimates whereas location A = Hohenheim, location B = Schickelsheim, and location C = Wetze, (B) Explained genotypic variance of the first ten principal coordinates, (C) Violin plot showing the density distribution of the first ten principal coordinates, (D) Histogram of the genetic similarities among the 77 isolates.

Estimation of nucleotide diversity and linkage disequilibrium

Percentage of polymorphic sites per total sequenced region of each gene (without singletons) varied from 0.9% (5/513) on TRI10b to 8.8% (65/734) on MetAP1 indicating a high nucleotide diversity in most tested genes (Table 1). LD of SNPs within the gene MetAP1 decayed rapidly within 200 bp of physical distance, i.e., the robust locally fitted regression of r2 values has a trend to decay from r2 = 0.35 to r2 < 0.1, whereas LD within the gene Erf2 had r2 values ranging from 0.8 to 0.2 and the regression of r2 trends to decay already after 150 bp (Figure 3). LD with r2 values higher than 0.1 were detected between all allele combinations within the tested genes followed by a rapid decay negatively correlated with the physical distance in base pair (Figure 4). In line with the rapid decay, LD between genes located on the same chromosome (i.e., TRI1 and MetAP1) is low (r2 < 0.1). Interestingly, 48%, 19%, and 45% of the SNP pairs between the genes TRI10/MetAP1, MetAP1/TRI5 and MetAP1/Erf2, respectively, have higher values of r2 than 0.1, although they are located on different chromosomes. Low r2 values (< 0.1) were observed for SNPs of gene pairs TRI10/Erf2 and TRI5/Erf2, whereas r2 values of 0.2 were detected between SNP pairs of TRI10b and TRI5, followed by a rapid decay of LD.

Table 1 Sequenced regions of candidate genes and its related nucleotide polymorphism
Figure 3
figure 3

Linkage disequilibrium (LD) decay. Robust locally fitted regression of linkage disequilibrium over base-pair distance within the two candidate genes, MetAP1 (25 SNPs) and Erf2 (11 SNPs).

Figure 4
figure 4

Pairwaise linkage disequilibrium (LD). The structure of LD within and among five candidate genes for 77 Fusarium isolates. Significant LD (P < 0.05, above diagonal) and LD measured as r2 between all pairs of selected SNP loci (below diagonal). The horizontal and vertical lines separate the candidate genes, red coloring indicates significant LD and the higher the r2 values (r2 ≥ 0.1), the darker is the used color, white indicates non-significant LD or r2 = 0.

Association analysis for aggressiveness and mycotoxin

Two adjacent SNPs of the gene MetAP1 were significantly associated (P < 0.05 using Bonferroni-Holm correction) [22] with mean FHB rating with an explained genotypic proportion of variance p G = 25.6 and 0.5, respectively (Table 2, Figure 5A). Additionally, one SNP in the gene Erf2 was significantly associated with this trait showing a p G of 13.1% (Table 2, Figure 5A). A single SNP significantly associated with DON production was identified in the TRI1 gene explaining 4.4% of the genotypic variance of DON production (Table 2, Figure 5B). All detected SNPs that were associated to the mentioned phenotypic traits were non-synonymously substituted (Table 2). Applying a haplotype analysis across genes, we could confirm a significant association between mean FHB rating and MetAP1 and between DON production and TRI1 (Additional file1).

Table 2 SNPs in the candidate genes significantly associated with mean FHB rating or DON content
Figure 5
figure 5

Significant associations in the candidate genes. (A) Erf2 and MetAp1 for mean FHB rating and (B) TRI1 for DON content. The dashed horizontal line indicates the threshold (Bonferroni-corrected P < 0.05).

Discussion

Association mapping based on candidate genes is a promising tool for high-resolution mapping of genes contributing to quantitative traits [23]. Nevertheless, it has not yet applied to investigate the basis of quantitative variation in aggressiveness and/or DON production in F. graminearum. This is a totally different approach than using knock-out mutants [24], because we are aiming for the analysis of quantitative differences produced by single nucleotide changes of the respective genes in a set of 77 isolates. The high number of singletons may, besides sequencing noise, be caused by frequent sexual recombination of the fungus. Therefore, we used a strict threshold and were analyzing only SNPs with allele frequencies > 0.1.

Choice of candidate genes

The genes for deoxynivalenol (DON) biosynthesis reside primarily in a 25 kb cluster (TRI5 cluster), TRI1 gene belong to a second smaller cluster that has a different chromosomal localization [16]. TRI1, TRI5, TRI14 have a direct effect on the production of DON or acetylated DON, TRI6 and TRI10 are regulating this pathway. TRI14 export the deoxynivalenol outside the mycelia in addition to its major role in aggressiveness [15].

The function of MetAP1 is still not described in Fusarium spp., a deletion, however, reduces growth in Saccharomyces cerevisiae [18, 25, 26]. In both, prokaryotes and eukaryotes the N-terminal methionine is often cleaved by methionine aminopeptidase encoded by the gene MetAP1 [18]. Erf2 gene is a component of the RAS protein subcellular localization pathway in yeast [17]. RAS2 gene is known to affect the pathogenesis of F. graminearum by regulation of hyphal growth and expression of hydrolytic enzymes [27]. Failing of palmitoylation mainly controlled by the Erf2 gene in S. cerevisiae reduces amount of RAS protein at the plasma membrane and lead to poor growth of yeast [17]. Both genes are not indispensible for fungal growth because their deletion does not prevent any growth at least in S. cerevisiae. And the 77 isolates analyzed in this study, all had a similar growth in vitro, produced enough spores for inoculation and were able to infect the host in the field. Significant, quantitative differences for aggressiveness and DON production, however, were found illustrating those genes with regulatory functions might be involved. Implementing associations between new candidate genes and important phenotypic traits may provide a useful tool to rapidly check a possible contribution of these genes to QTL of DON production or aggressiveness in F. graminearum as shown previously in mammal research [23]. If an association is found, a more detailed analysis of the function of the respective genes should, of course, follow.

Molecular diversity of the 77 isolates

All isolates were checked by respective primers for their designation to F. graminearum s.s. Nucleotide diversity within candidate genes of this study ranged from 1 to 9 per 100 bp being high compared to 0 to 2 SNPs per 100 bp reported by Cuomo et al. [13]. This was not unexpected since Cuomo et al. [13] used just two isolates (PH-1 and GZ3639) while in this study 77 isolates were analyzed. Many genes in F. graminearum are localized in regions of high SNP density, i.e. highly variable regions, especially those expressed during plant infection. Moreover, it is known that the density of SNPs is biased in F. graminearum in a way that 50% of the SNPs are within 13% of the genome sequence, the highest SNP density was reported in some regions of chromosome II, where most of the TRI genes are located [13]. The high nucleotide diversity of these groups of genes suggests that the fungus has a great capacity for adaptability and genetic change during its interaction with the host plant [13]. Despite this, we found much more SNPs with allele frequencies > 0.1 in the MetAP1 and Erf2 genes than in the TRI genes analyzed. This was not the case, when just the number of detected SNPs is regarded, indicating that nucleotide diversity may be similar, but more rare haplotypes occur in the TRI genes. Moreover, high SNPs density was detected in one or two large interstitial regions on chomosomes I, II, and IV [13].

Population structure and consequences for association mapping

Correcting for population and/or family structure is essential for association mapping to decrease the number of false positive QTL [28]. An appropriate statistical model should provide an excellent compromise between correcting for population stratification to decrease the probability of detecting false positive SNP-trait associations but still retaining enough information within the SNPs for QTL detection [29]. Although some grouping according to the sampling location might be seen in the PCoA, the violin plots did not show any separated knots on the first ten principle coordinates (Figure 2). Thus, no distinct subpopulations were identified and correction for familial relatedness between isolates should be sufficient. Therefore, trait-SNP association was investigated in detail with the K model, which incorporated estimates of kinship coefficients based on SSR data.

Extent of linkage disequilibrium and resolution of association mapping

Rapid decay of regression parameter r2 in a short physical distance of 200 bp within two genes was presented in this study (Figure 3), providing a high resolution of association mapping. Generally, low LD was reported in F. graminearum populations using selection-neutral markers (VNTR, AFLP, or RFLP) especially in the region that includes the TRI5 gene cluster [11]. Moreover, the weak correlation between r2 and physical distance in addition to the large proportion of unlinked SNP with significant LD, such as SNPs between gene pairs TRI5/MetAP1, TRI10/MetAP1, or MetAP1/Erf2 (Figure 4), suggest the presence of other forces generating LD between unlinked SNPs. The values of r2 between most pairs of SNPs in TRI5/TRI10b are > 0.1, presenting these two genes in LD, that might refer to the controlling role of TRI10 on TRI5 gene cluster [24, 30, 31]. That TRI10 and TRI5 are located close to each other on the same chromosome play a minor role hence the detected LD decay is occurring within 200 bp. Selection acting on oligo- or polygenic traits such as aggressiveness and DON production might be responsible for this [32]; an alternative explanation would be the involvement in the same trait network with other physiologically important genes.

Association mapping of genes underlying aggressiveness and DON production

We identified three SNPs related to aggressiveness and one SNP related to DON production using the fairly conservative Bonferroni-Holm correction. Setting the threshold for detection of trait associations to an allele frequency of 0.1 might underestimate the number of QTL detected but is due to the restricted population size. Association between SNPs of candidate gene Erf2 and phenotypic data revealed a significant association to mean FHB rating, but not to DON content. This might be explained by the fact that the Erf2 gene is involved in RAS2 processing or trafficking that precedes palmitoylation of RAS2 genes [17]. RAS2 is known to regulate the aggressiveness through affecting fungal growth and regulating other pathogenicity genes, e.g., Gmpk1, which controls the induction of extracellular enzymes required for pathogenesis [27]. Further study is needed to understand how Erf2 affects differences in aggressiveness of F. graminearum isolates.

Two SNPs in MetAP1 gene significantly associated with aggressiveness were located adjacent to each other forming a collinearity pattern. The role of MetAP1 was reported in S. cerevisiae as reducing cell growth by N-terminal protein modification. A similar role might be expected for MetAP1 in F. graminearum, hence significant positive correlations between aggressiveness and fungal biomass (r = 0.7, P = 0.01), and fungal biomass and DON content (r = 0.8, P = 0.01) were reported among 50 F. graminearum isolates [33]. Nevertheless, analyzed isolates were sampled from visually diseased spikelets, thus all were able to infect wheat ears and induce symptoms, non-aggressive isolates were not included in our F. graminearum population sample.

A single SNP detected in TRI1 was associated with DON production in F. graminearum. TRI1 encodes a cytochrome P450 monooxygenase that catalyzes hydroxylation of C-8 position during trichothecene biosynthesis [34]. This confirmation change obviously does not affect aggressiveness, because the respective SNP is just associated with DON content but not with aggressiveness.

Conclusions

This is the first candidate gene association mapping study provided insights on some genes involved in aggressiveness and DON concentration of F. graminearum s.s. The described associations should be validated using a larger number of isolates and different environments. The validated genes are an important starting point for further functional analyses.

Methods

Fungal material

Ears of winter wheat (Triticum aestivum L.) showing symptoms of FHB were sampled from three commercial fields in Germany to establish a fungal population of F. graminearum s.s.: Stuttgart-Hohenheim in southwest Germany in 2008, Wetze and Schickelsheim in Lower Saxony in 2006 and 2007, respectively. From each infected head, one isolate was recovered, transferred onto a fresh SNA plate and placed under permanent UV light for induction of sporulation as described recently [3]. From each isolate, one single spore was picked out under the microscope and transferred onto a fresh SNA plate to establish a single-spore culture. The single-spore isolates were checked morphologically and analyzed for their species specificity and chemotype by different PCR-based assays as described recently in detail [3]. In total, 77 single-spore isolates of F. graminearum s.s. (former lineage 7) were inoculated on the moderately resistant German spring wheat cultivar Taifun (KWS LOCHOW GMBH, Bergen, Germany) for analyzing aggressiveness and DON production in the field.

Design of field studies

Field experiments were planted at each of two locations in 2009 and 2010: Hohenheim (HOH, longitude 9° 12' 58″, latitude 48° 42' 50″, altitude 400 m) and Oberer Lindenhof (OLI, longitude 9° 18' 12″, latitude 48° 28' 26″, altitude 700 m), resulting in four environments (location × year combinations). Mean annual temperatures at HOH and OLI were 10.1°C and 9.0°C, respectively, mean annual precipitations were 644 mm and 723 mm, respectively, across 2009 and 2010. Plants were grown in two-rowed micro-plots of 1 m length and 0.42 m width. Plots were arranged in a chessboard-like design, i.e., each plot with a wheat entry was bordered by four plots of similar size that were planted with a long-strawed spring triticale cultivar Nilex (NORDSAAT GmbH, Halberstadt, Germany) to reduce inter-plot interference caused by drifting of inoculum during spraying or secondary distribution of spores. The experiment was arranged in a split-plot design with two replications. The main plot factor was the F. graminearum s.s. population, the subplot factor the isolate. Both factors were assigned to a randomized complete block design. Eight plots per replication included in the subplot factor were not inoculated to estimate the degree of natural infection.

Inoculum was sprayed with a concentration of 2 × 105 spores ml-1 onto the wheat heads of each plot. Inoculation was performed at full flowering time of the wheat cultivar to ensure maximum susceptibility of wheat to F. graminearum. Fusarium head blight (FHB) aggressiveness was rated visually four times as the percentage of infected spikelets per plot (0-100%). This rating included both the number of infected spikes per plot and the number of infected spikelets per spike. To compare FHB reactions, the arithmetic means of four ratings were used and assigned as mean FHB rating throughout the paper. All plots were harvested by hand, threshed, and the grain analyzed to quantify the amount of DON by a commercially available immunotest (R-biopharm AG, Darmstadt, Germany) as previously described in detail [9]. This test cannot differentiate between DON and 3-ADON, so the results include both mycotoxins. Generally, however, the amount of 3-ADON is only 2-3% of total DON content [35]. In each of the four environments, the natural infection rate was very low ranging from 0 to 3% FHB rating and from 0 to 0.34 mg kg-1 DON concentration. Prediction of NIV chemotype was done using the primers N11, 15D11, 3D11, and 11R in a multiplex polymerase chain reaction (PCR) as designed by Zhang et al. [36].

Detection of population structure and gene sequencing

In order to analyze the population structure, all isolates were fingerprinted with 19 simple sequence repeat (SSR) markers dispersed throughout the whole genome following standard protocols [3].

Five candidate genes were chosen according to their role in trichothecene biosynthesis and/or aggressiveness (TRI1, TRI5, TRI6, TRI10, and TRI14). Two other, yet uncharacterized candidate genes in Fusarium spp. (Erf2, MetAP1) were tested for a possible association with aggressiveness. Nucleotide sequences of these genes were imported from Fusarium graminearum database FGDB [37]. The selected genes were (Table 3): TRI1 (FGSG_00071), TRI5 (FGSG_03538), TRI6 (FGSG_16251), TRI10 (FGSG_03538), TRI14 (FGSG_03543), MetAP1 (FGSG_01397), and Erf2 (FGSG_08531). Specific primers were designed (see Additional file 2) to amplify parts of these genes using the software Primer Premier 4.0 (Premier Biosoft International, CA, U.S.A.).

Table 3 Name of tested genes, chromosomal localization, primer sequence and expected amplified DNA products

Polymerase chain reaction (PCR) was performed using the designed primers related to each gene separately following standard protocols, however, with a different annealing temperature for each gene (Table 3). Purification of PCR products was performed by precipitating the DNA with 10% (v/v) of 3 M Sodium Acetate and 25% (v/v) of absolute cold ethanol overnight. Precipitated DNA was cleaned twice with 70% ethanol and finally diluted in 10 μl ddH2O. Purified PCR products were sequenced (QIAGEN® Sequencing Services, Hilden, Germany). The sequence was performed once for each isolate, hence the alignment was mostly identical (the sequence was repeated in case it fail to align to the other sequence or in case of high noise in the sequence). Expected sizes of PCR products were obtained for all tested genes. All sequences of TRI6 and TRI14 were located in the coding region of the genes (exon, see Additional file 1). Sequences of the genes TRI5, MetAP1, Erf2, and TRI10 (two parts) were stretched over two exons. Sequences of genes were obtained from right to left sides of PCR products, just genes MetAP1 and Erf2 were obtained from left to right side. Sequence of TRI1 was located over four exons according to the reference sequence. Number of identified SNPs was high over all sequenced regions of candidate genes (Table 1). The sequences were aligned (according to the initial matrix published in FGDB) using CLC sequence viewer 6.3 (CLC-bio, Aarhus, Denmark) to identify single nucleotide polymorphisms (SNPs) among the 77 isolates.

Phenotypic data analyses

The following linear mixed model was used to estimate variance components: y ijn = μ+Iso i +Env j +(Iso × Env) ij +Rep n +e ijn , where μ is the population mean, Iso i the genetic effect of the ith isolate, Env j the effect of the jth environment, Iso × Env the isolate times environment interaction effect, Rep n the effect of the nth replication, and e ijn the residual error. Variance components were determined by the restricted maximum likelihood (REML) method using the software ASReml 2.0 (VSN International Ltd, Hemel Hempstead, U.K.). Significance for variance component estimates was tested by model comparison with likelihood ratio tests where the halfed P values were used as an approximation [38]. Heritability (h2 ) on an entry-mean basis was estimated as the ratio of genotypic to phenotypic variance according to Melchinger et al. [39]. Furthermore, genotypes were regarded as fixed effects and best linear unbiased estimates (BLUEs) were determined for all isolates and traits.

Genetic relatedness among the 77 isolates was determined by applying principal coordinate analysis (PCoA) [40] based on the modified Rogers' distances of the isolates [41]. Linkage disequilibrium (LD) between the selected SNPs was assessed by the LD measure r2 [42] and significance of LD was tested with Fisher's exact tests [43]. LD and PCoA computations were performed with the software package Plabsoft [44].

Association analysis

A two-step association approach was applied in this study and the BLUEs per environment were used as input for the association analysis. The linear mixed model for the association approach was: y ijn = μ+a p +Iso i +Env j +e ijp , where a p is the effect of allele p. The allele effect a p was modeled as fixed effect whereas Iso i and Env j were regarded as random effects. We assumed that the variance of the random genetic effect was Var(g) = 2Kσ2 g , where σ2 g refers to the genetic variance estimated by REML and K was a 77 × 77 matrix of kinship coefficients that define the degree of genetic covariance between all pairs of entries. We followed the suggestion of Bernardo [45] and calculated the kinship coefficient K ij between isolates i and j on the basis of the SSR marker data as K ij = 1+(S ij -1)/(1-T ij ), where S ij is the proportion of marker loci with shared variants between isolates i and j, and T ij is the average probability that a variant from one isolate i and a variant from one isolate j are alike in state, given that they are not identical by descent. The coefficient T ij was estimated separately for each gene and trait using a REML method setting negative kinship values between isolates to zero. SNPs with allele frequencies < 0.1 were not considered in the association analysis or LD estimation. The obtained optimum T values were for DON content: TRI1 (0.025), TRI5 (0.575), TRI10 (0.375), MetAP1 (0.300), and Erf2 (0.400). For mean FHB rating: TRI1 (0.600), TRI5 (0.150), TRI10 (0.200), MetAP1 (0.075), and Erf2 (0.275). For the detection of main effects of the candidate gene SNPs, these were fitted as fixed effects in the mixed model and their significance was tested by a Wald F test. Based on the Wald F statistic, we performed tests for the presence of marker-phenotype associations with a significant (P < 0.05) effect on DON content and mean FHB rating applying the Bonferroni-Holm procedure [22] to correct for multiple testing.

The proportion of genotypic variance (p G ) explained by the detected SNP was calculated by fitting each SNP in a linear model to obtain R2 adj . The ratio p G = R2 adj /h2 yielded the proportion of genotypic variance [46]. In the case of MetAP1, where two SNPs were detected for mean FHB rating, both were simultaneously fitted in the linear model in the order of their P values to correct for collinearity. Haplotype analyses were performed with the same procedure to test the association with both phenotypic traits (Additional file 2: Table S1). We applied the same threshold level like for SNP analyses. All mixed model calculations were performed using the software ASReml 2.0 (VSN International Ltd, Hemel Hempstead, U.K.).

References

  1. Agrios GN: Plant Pathology. 2004, San Diego: Elsevier Press, 5

    Google Scholar 

  2. O'Donnell K, Ward TJ, Geiser DM, Kistler HC, Aoki T: Genealogical concordance between the mating type locus and seven other nuclear genes supports formal recognition of nine phylogenetically distinct species within the Fusarium graminearum clade. Fungal Genet Biol. 2004, 41 (6): 600-623. 10.1016/j.fgb.2004.03.003.

    Article  PubMed  Google Scholar 

  3. Talas F, Parzies HK, Miedaner T: Diversity in genetic structure and chemotype composition of Fusarium graminearum sensu stricto populations causing wheat head blight in individual fields in Germany. Eur J Plant Pathol. 2011, 131 (1): 39-48. 10.1007/s10658-011-9785-3.

    Article  Google Scholar 

  4. Miedaner T, Cumagun CJR, Chakraborty S: Population genetics of three important head blight pathogens Fusarium graminearum, F. pseudograminearum and F. culmorum. J Phytopathol. 2008, 156 (3): 129-139. 10.1111/j.1439-0434.2007.01394.x.

    Article  Google Scholar 

  5. European Commission: Commission Regulation (EC) No. 1126/2007 of 28 September 2007 amending Regulation (EC) No. 1881/2006 setting maximum levels for certain contaminants in foodstuffs as regards Fusarium toxins in maize and maize products. 2007, Official Journal of the European Union L255, 14-17. Published online

    Google Scholar 

  6. Vanderplank J: Disease Resistance of Plants. 1968, New York: Academic Press

    Google Scholar 

  7. Gale LR, Chen LF, Hernick CA, Takamura K, Kistler HC: Population analysis of Fusarium graminearum from wheat fields in Eastern China. Phytopathology. 2002, 92 (12): 1315-1322. 10.1094/PHYTO.2002.92.12.1315.

    Article  CAS  PubMed  Google Scholar 

  8. Miedaner T, Schilling AG, Geiger HH: Molecular genetic diversity and variation for aggressiveness in populations of Fusarium graminearum and Fusarium culmorum sampled from wheat fields in different countries. J Phytopathol. 2001, 149: 641-648. 10.1046/j.1439-0434.2001.00687.x.

    Article  CAS  Google Scholar 

  9. Talas F, Kalih R, Miedaner T: Within-field variation of Fusarium graminearum sensu stricto isolates for aggressiveness and deoxynivalenol production in wheat head blight. Phytopathology. 2012, 102 (1): 128-134. 10.1094/PHYTO-06-11-0162.

    Article  CAS  PubMed  Google Scholar 

  10. Ward TJ, Clear RM, Rooney AP, O'Donnell K, Gaba D, Patrick S, Starkey DE, Gilbert J, Geiser DM, Nowicki TW: An adaptive evolutionary shift in Fusarium head blight pathogen populations is driving the rapid spread of more toxigenic Fusarium graminearum in North America. Fungal Genet Biol. 2008, 45 (4): 473-484. 10.1016/j.fgb.2007.10.003.

    Article  PubMed  Google Scholar 

  11. Zeller KA, Bowden RL, Leslie JF: Population differentiation and recombination in wheat scab populations of Gibberella zeae from the United States. Mol Ecol. 2004, 13 (3): 563-571. 10.1046/j.1365-294X.2004.02098.x.

    Article  PubMed  Google Scholar 

  12. Cumagun CJR, Bowden RL, Jurgenson JE, Leslie JF, Miedaner T: Genetic mapping of pathogenicity and aggressiveness of Gibberella zeae (Fusarium graminearum) towards wheat. Phytopathology. 2004, 94 (5): 520-526. 10.1094/PHYTO.2004.94.5.520.

    Article  CAS  PubMed  Google Scholar 

  13. Cuomo CA, Güldener U, Xu JR, Trail F, Turgeon BG, Di Pietro A, Walton JD, Ma LJ, Baker SE, Rep M, Adam G, Antoniw J, Baldwin T, Calvo S, Chang YL, DeCaprio D, Gale LR, Gnerre S, Goswami RS, Hammond-Kosack K, Harris LJ, Hilburn K, Kennell JC, Kroken S, Magnuson JK, Mannhaupt G, Mauceli E, Mewes HW, Mitterbauer R, Muehlbauer G, et al: The Fusarium graminearum genome reveals a link between localized polymorphism and pathogen specialization. Science. 2007, 317 (5843): 1400-1402. 10.1126/science.1143708.

    Article  CAS  PubMed  Google Scholar 

  14. Winnenburg R, Urban M, Beacham A, Baldwin TK, Holland S, Lindeberg M, Hansen H, Rawlings C, Hammond-Kosack KE, Kohler J: PHI-base update: additions to the pathogen host interaction database. Nucleic Acids Res. 2008, 36: 572-576.

    Article  Google Scholar 

  15. Dyer RB, Plattner RD, Kendra DF, Brown DW: Fusarium graminearum TRI1 is Required for high virulence and DON production on wheat but not for DON synthesis in vitr. J Agric Food Chem. 2005, 53: 9281-9287. 10.1021/jf051441a.

    Article  CAS  PubMed  Google Scholar 

  16. Kimura M, Tokai T, Takahashi-Ando N, Ohsato S, Fujimura M: Molecular and genetic studies of Fusarium trichothecene biosynthesis: pathway genes and evolution. Biosci Biotechnol Biochem. 2007, 71 (9): 2105-2123. 10.1271/bbb.70183.

    Article  CAS  PubMed  Google Scholar 

  17. Bartels DJ, Mitchell DA, Dong X, Deschenes RJ: Erf a novel gene product that affects the localization and palmitoylation of RAS in Saccharomyces cerevisiae. Mol Cell Biol. 1999, 19 (10): 6775-6787.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  18. Chen S, Vetro JA, Chang YH: The specificity in vivo of two distinct methionine aminopeptidases in Saccharomyces cerevisiae. Arch Biochem Biophys. 2002, 408 (1): 87-93. 10.1016/S0003-9861(02)00447-2.

    Article  Google Scholar 

  19. Yu J, Buckler ES: Genetic association mapping and genome organization of maize. Curr Opin Biotechnol. 2006, 17 (2): 155-160. 10.1016/j.copbio.2006.02.003.

    Article  CAS  PubMed  Google Scholar 

  20. Wilson LM, Whitt SR, Ibanez AM, Rocheford TR, Goodman MM, Buckler ES: Dissection of maize kernel composition and starch production by candidate gene association. Plant Cell. 2004, 16 (10): 2719-2733. 10.1105/tpc.104.025700.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  21. Karugia GW, Suga H, Gale LR, Nakajima T, Tomimura K, Hyakumachi M: Population structure of Fusarium graminearum species complex from a single Japanese wheat field sampled in two consecutive years. Plant Dis. 2009, 93 (2): 170-174. 10.1094/PDIS-93-2-0170.

    Article  CAS  Google Scholar 

  22. Holm S: A simple sequentially rejective multiple test procedure. Scand J Stat. 1979, 6: 65-70.

    Google Scholar 

  23. Magee DA, Sikora KM, Berkowicz EW, Berry DP, Howard DJ, Mullen MP, Evans RD, Spillane C, MacHugh DE: DNA sequence polymorphism in a panel of eight candidate bovine imprinted genes and their association with performance traits in Irish Holstein-Friesian cattle. BMC Genet. 2010, 11: 93-

    Article  PubMed Central  PubMed  Google Scholar 

  24. Proctor RH, Hohn TM, McCormick SP, Desjardins AE: TRI encodes an unusual zinc finger protein involved in regulation of trichothecene biosynthesis in Fusarium sporotrichioides. Appl Environ Microbiol. 1995, 61 (5): 1923-1930.

    PubMed Central  CAS  PubMed  Google Scholar 

  25. Dummitt B, Micka WS, Chang YH: Yeast glutamine fructose 6-phosphate aminotransferase (Gfa) requires methonine aminopeptidase activity for proper function. J Biol Chem. 2005, 280 (14): 14356-14360. 10.1074/jbc.M501059200.

    Article  CAS  PubMed  Google Scholar 

  26. Li X, Chang YH: Amino terminal protein processing in Saccharomyces cerevisiae is an essential function that requires two distinct methionine aminopeptidases. Proc Natl Acad Sci. 1995, 92 (26): 12357-12361. 10.1073/pnas.92.26.12357.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  27. Bluhm BH, Zhao X, Flaherty JE, Xu JR, Dunkle LD: RAS regulates growth and pathogenesis in Fusarium graminearum. Mol Plant Microbe Interact. 2007, 20 (6): 627-636. 10.1094/MPMI-20-6-0627.

    Article  CAS  PubMed  Google Scholar 

  28. Reif JC, Maurer HP, Korzun V, Ebmeyer E, Miedaner T, Würschum T: Genetic architecture of grain yield and heading time in soft winter wheat. Theor Appl Genet. 2011, 123 (2): 283-292. 10.1007/s00122-011-1583-y.

    Article  PubMed  Google Scholar 

  29. Würschum T, Maurer HP, Schulz B, Möhring J, Reif JC: Genome-wide association mapping reveals epistasis and genetic interaction networks in sugar beet. Theor Appl Genet. 2011, 123 (1): 109-118. 10.1007/s00122-011-1570-3.

    Article  PubMed  Google Scholar 

  30. Peplow AW, Tag AG, Garifullina GF, Beremand MN: Identification of new genes positively regulated by TRI1 and a regulatory network for trichothecene mycotoxin production. Appl Environ Microbiol. 2003, 69 (5): 2731-2736. 10.1128/AEM.69.5.2731-2736.2003.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  31. Seong KY, Pasquali M, Zhou X, Song J, Hilburn K, McCormick S, Don Y, Xu JR, Kistler HC: Global gene regulation by Fusarium transcription factors TRI6 and TRI10 reveals adaptations for toxin biosynthesis. Mol Microbiol. 2009, 72 (2): 354-367. 10.1111/j.1365-2958.2009.06649.x.

    Article  CAS  PubMed  Google Scholar 

  32. Stich B, Melchinger AE, Frisch M, Maurer HP, Heckenberger M, Reif JC: Linkage disequilibrium in European elite maize germplasm investigated with SSRs. Theor Appl Genet. 2005, 111 (4): 723-730. 10.1007/s00122-005-2057-x.

    Article  PubMed  Google Scholar 

  33. Cumagun CJR, Rabenstein F, Miedaner T: Genetic variation and covariation for aggressiveness deoxynivalenol production and fungal colonization among progeny of Gibberella zeae in wheat. Plant Pathol. 2004, 53 (4): 446-453. 10.1111/j.1365-3059.2004.01046.x.

    Article  CAS  Google Scholar 

  34. Meek IB, Peplow AW, Ake CJR, Phillips TD, Beremand MN: TRI encodes the cytochrome P450 monooxygenase for C-8 hydroxylation during trichothecene biosynthesis in Fusarium sporotrichioides and resides upstream of another new TR gene. Appl Environ Microbiol. 2003, 69 (3): 1607-1613. 10.1128/AEM.69.3.1607-1613.2003.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  35. Von der Ohe C, Gauthier V, Tamburic-Ilincic L, Brule-Babel A, Fernando WGD, Clear R, Ward TJ, Miedaner T: A comparison of aggressiveness and deoxynivalenol production between Canadian Fusarium graminearum isolates with 3-acetyl and 15-acetyldeoxynivalenol chemotypes in field-grown spring wheat. Eur J Plant Pathol. 2010, 127: 407-417. 10.1007/s10658-010-9607-z.

    Article  Google Scholar 

  36. Zhang H, Zhang Z, Van der Lee T, Xu J, Yang L, Yu D, Waalwijk C, Feng J: Population genetic analyses of Fusarium asiaticum populations from barley suggest a recent shift favoring 3ADON producers in southern China. Phytopathology. 2010, 100 (4): 328-336. 10.1094/PHYTO-100-4-0328.

    Article  CAS  PubMed  Google Scholar 

  37. Wong P, Walter M, Lee W, Mannhaupt G, Münsterkötter M, Mewes HW, Adam G, Güldener U: FGDB: Revisiting the genome annotation of the plant pathogen Fusarium graminearum. Nucleic Acids Res. 2011, 39 (4): 637-639.

    Article  Google Scholar 

  38. Stram DO, Lee JW: Variance components testing in longitudinal mixed effects model. Biometrics. 1994, 50 (4): 1171-1177. 10.2307/2533455.

    Article  CAS  PubMed  Google Scholar 

  39. Melchinger AE, Utz HF, Schön CC: Quantitative trait locus (QTL) mapping using different testers and independent population samples in maize reveals low power of QTL detection and larger bias in estimates of QTL effects. Genetics. 1998, 149 (1): 383-403.

    PubMed Central  CAS  PubMed  Google Scholar 

  40. Gower JC: Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika. 1966, 53 (3): 325-338.

    Article  Google Scholar 

  41. Wright S: Evolution and Genetics of Populations, Variability within and among Natural Populations. 1978, Chicago: The University of Chicago Press, 4

    Google Scholar 

  42. Weir BS: Genetic data analysis II. 1996, Sunderland: Sinauer Associates, 2

    Google Scholar 

  43. Hill WG, Robertson A: Linkage disequilibrium in finite populations. Theor Appl Genet. 1968, 38 (2): 226-231.

    Article  CAS  PubMed  Google Scholar 

  44. Maurer HP, Melchinger AE, Frisch M: Population genetic simulation and data analysis with Plabsoft. Euphytica. 2008, 161 (1): 133-139. 10.1007/s10681-007-9493-4.

    Article  Google Scholar 

  45. Bernardo R: Estimation of coefficient of coancestry using molecular markers in maize. Theor Appl Genet. 1993, 85 (8): 1055-1062.

    Article  CAS  PubMed  Google Scholar 

  46. Utz HF, Melchinger AE, Schön CC: Bias and sampling error of the estimated proportion of genotypic variance explained by quantitative trait loci determined from experimental data in maize using cross validation and validation with independent samples. Genetics. 2000, 154 (4): 1839-1849.

    PubMed Central  PubMed  Google Scholar 

Download references

Acknowledgements

This work was funded by the German Academic Exchange Service (DAAD), Bonn, Germany and by the State Plant Breeding Institute (720), Universitaet Hohenheim, Germany. The authors like to thank Mrs. B. Lieberherr for excellent assistance in field work.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Thomas Miedaner.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

FT carried out the phenotypic and molecular analyses, performed parts of the statistical analyses and drafted the manuscript. TW performed parts of the statistical analyses and helped to draft the manuscript, JCR edited the manuscript, HKP† participated in the design of the study and supported the technical realization. TM participated in the design of the field study and helped to draft the manuscript. All authors read and approved the final manuscript.

Electronic supplementary material

12863_2011_988_MOESM1_ESM.EPS

Additional file 1: Genomic organizations of the seven candidate genes with exons in black color. The sequenced region is shown as a green bar above each gene. (EPS 269 KB)

Additional file 2: Table S1. Haplotype analysis of the genes MetAP1 and TRI1. (DOC 37 KB)

Authors’ original submitted files for images

Rights and permissions

Open Access This article is published under license to BioMed Central Ltd. This is an Open Access article is distributed under the terms of the Creative Commons Attribution License ( https://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article

Talas, F., Würschum, T., Reif, J.C. et al. Association of single nucleotide polymorphic sites in candidate genes with aggressiveness and deoxynivalenol production in Fusarium graminearum causing wheat head blight. BMC Genet 13, 14 (2012). https://doi.org/10.1186/1471-2156-13-14

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1471-2156-13-14

Keywords