Genome-Wide Association Mapping Uncovers Fw1, a Dominant Gene Conferring Resistance to Fusarium Wilt in Strawberry

Abstract

Fusarium wilt, a soil-borne disease caused by the fungal pathogen Fusarium oxysporum f. sp. fragariae, threatens strawberry (Fragaria × ananassa) production worldwide. The spread of the pathogen, coupled with disruptive changes in soil fumigation practices, have greatly increased disease pressure and the importance of developing resistant cultivars. While resistant and susceptible cultivars have been reported, a limited number of germplasm accessions have been analyzed, and contradictory conclusions have been reached in earlier studies to elucidate the underlying genetic basis of resistance. Here, we report the discovery of Fw1, a dominant gene conferring resistance to Fusarium wilt in strawberry. The Fw1 locus was uncovered in a genome-wide association study of 565 historically and commercially important strawberry accessions genotyped with 14,408 SNP markers. Fourteen SNPs in linkage disequilibrium with Fw1 physically mapped to a 2.3 Mb segment on chromosome 2 in a diploid F. vesca reference genome. Fw1 and 11 tightly linked GWAS-significant SNPs mapped to linkage group 2C in octoploid segregating populations. The most significant SNP explained 85% of the phenotypic variability and predicted resistance in 97% of the accessions tested-broad-sense heritability was 0.96. Several disease resistance and defense-related gene homologs, including a small cluster of genes encoding nucleotide-binding leucine-rich-repeat proteins, were identified in the 0.7 Mb genomic segment predicted to harbor Fw1 DNA variants and candidate genes identified in the present study should facilitate the development of high-throughput genotyping assays for accurately predicting Fusarium wilt phenotypes and applying marker-assisted selection.

Keywords: Fragaria; Fusarium wilt; innate immunity; polyploid; strawberry.

Figure 1

Phenotypic distributions for resistance to Fusarium wilt in a genome-wide association study (GWAS) in strawberry. Histograms are shown for phenotypes observed in (A) 2016 and (B) 2017 field experiments in Davis, California among 565 strawberry germplasm accessions artificially inoculated with isolate AMP132 of Fusarium oxysporum f. sp. fragariae. Phenotypes were observed four to nine weeks post-inoculation in 2016 and 26 to 36 weeks post-inoculation in 2017. Least square means were estimated from four clonal replicates per entry with entries arranged in a square lattice experiment design. Disease scores ranged from 1 to 5, where 1 = healthy and 5 = dead.

Figure 2

Phenotypic correlation (Pearson’s r = 0.84, P < 0.001) between years for Fusarium wilt resistance phenotypes in strawberry. A fitted linear regression is shown in blue. Fusarium wilt resistance was phenotyped in 2016 and 2017 field experiments in Davis, California among 565 strawberry germplasm accessions artificially inoculated with isolate AMP132 of Fusarium oxysporum f. sp. fragariae. The phenotypes shown were observed nine weeks post-inoculation in 2016 (x-axis) and 36 weeks post-inoculation in 2017 (y-axis).

Figure 3

Genome-wide association study for resistance to Fusarium wilt in octoploid strawberry using chromosome positions from the diploid (x = 7) F. vesca reference genome (Edger et al. 2018). Manhattan plots are for phenotypes observed in 2016 (A) and 2017 (B) experiments. The horizontal dashed line identifies a 0.01 Bonferroni-corrected significance threshold.

Figure 4

SNPs in linkage disequilibrium with a Fusarium wilt resistance gene (Fw1) that were genetically mapped to chromosome 2C in octoploid strawberry. Manhattan plots are shown for phenotypes observed in 2016 (A) and 2017 (B) GWAS experiments with -log₁₀ p-values for nine SNPs plotted against chromosome positions in a diploid (x = 7) F. vesca reference genome (Edger et al. 2018). Pairwise marker linkage disequilibrium statistics are shown for the GWAS panel (C).

Figure 5

Distributions for Fusarium wilt resistance phenotypes observed in octoploid segregating populations. Histograms are shown for Fusarium wilt resistance phenotypes in segregating populations developed by self-pollinating the F. × ananassa cultivars (A) Fronteras and (B) Portola. Parents (Fronteras and Portola) and grandparents (04C018P004/05C165P001 and 97C093P007/97C209P001) of the S₁ populations and 93 S₁ individuals from each population were artificially inoculated with Fusarium oxysporum f. sp. fragariae isolate AMP132. Phenotypes were observed 36 weeks post-inoculation in a 2017 field experiment in Davis, California.

Figure 6

Genetic mapping of a Fusarium wilt resistance gene (Fw1) in octoploid segregating populations. Likelihood-odds (LOD) statistics and linkage group positions (cM) are shown for a quantitative trait locus (QTL) for Fusarium wilt resistance identified by interval mapping in (A) Portola and (B) Fronteras S₁ mapping populations. The parents (Fronteras and Portola) and grandparents (04C018P004/05C165P001 and 97C093P007/97C209P001) of the S₁ populations and 93 S₁ individuals from each population were genotyped with the iStraw35 SNP array and artificially inoculated with isolate AMP132 of Fusarium oxysporum f. sp. fragariae at planting. Phenotypes were observed 36 weeks post-inoculation in a 2017 field experiment in Davis, California. The Fw1 QTL mapped to identical locations on the upper arm of chromosome 2C in both populations. One- and two-LOD confidence intervals are shown. Highlighted SNPs (bold red) were significant in genome-wide association studies.

Figure 7

Pedigree network for 1,663 F. × ananassa germplasm accessions with birth years ranging from 1814 to 2012. The north to south orientation of the network is approximately chronological. The pedigree records are supplied in Supplemental File 5. Nodes represent germplasm accessions, whereas connecting lines represent first-degree relatives (parent-offspring). The color of the node signifies a combination of the Fusarium wilt resistance phenotype and the AX-166521396 SNP marker genotype for 565 germplasm accessions. The other 1,098 germplasm accessions in the pedigree network were untested (small light gray nodes). The AX-166521396 SNP marker was in linkage disequilibrium with the Fw1 gene conferring resistance to Fusarium wilt. The A allele was associated with the resistant allele (Fw1), whereas the G allele was associated with the susceptible allele (fw1). AX-166521396 SNP marker genotypes predicted Fusarium wilt resistance phenotypes in 97% of the germplasm accessions tested: most A/A and A/G genotypes were resistant (blue and cyan filled circles, respectively), whereas most G/G genotypes were susceptible (salmon filled circles). Seven G/G genotypes were resistant and predicted to carry novel Fusarium wilt resistance genes (black filled circles).