SNP markers tightly linked to root knot nematode resistance in grapevine (Vitis cinerea) identified by a genotyping-by-sequencing approach followed by Sequenom MassARRAY validation

Abstract

Plant parasitic nematodes, including root knot nematode Meloidogyne species, cause extensive damage to agriculture and horticultural crops. As Vitis vinifera cultivars are susceptible to root knot nematode parasitism, rootstocks resistant to these soil pests provide a sustainable approach to maintain grapevine production. Currently, most of the commercially available root knot nematode resistant rootstocks are highly vigorous and take up excess potassium, which reduces wine quality. As a result, there is a pressing need to breed new root knot nematode resistant rootstocks, which have no impact on wine quality. To develop molecular markers that predict root knot nematode resistance for marker assisted breeding, a genetic approach was employed to identify a root knot nematode resistance locus in grapevine. To this end, a Meloidogyne javanica resistant Vitis cinerea accession was crossed to a susceptible Vitis vinifera cultivar Riesling and results from screening the F1 individuals support a model that root knot nematode resistance, is conferred by a single dominant allele, referred as MELOIDOGYNE JAVANICA RESISTANCE1 (MJR1). Further, MJR1 resistance appears to be mediated by a hypersensitive response that occurs in the root apical meristem. Single nucleotide polymorphisms (SNPs) were identified using genotyping-by-sequencing and results from association and genetic mapping identified the MJR1 locus, which is located on chromosome 18 in the Vitis cinerea accession. Validation of the SNPs linked to the MJR1 locus using a Sequenom MassARRAY platform found that only 50% could be validated. The validated SNPs that flank and co-segregate with the MJR1 locus can be used for marker-assisted selection for Meloidogyne javanica resistance in grapevine.

Fig 1. Flow chart for GBS and filtering of the C2-50 and Riesling SNP sets.

Genomic DNA was isolated from the parents and F₁ progeny genotypes. Genome complexity was reduced by digesting the genomic DNA with the ApeKI methylation sensitive restriction endonuclease. Libraries were sequenced using Illumina HiSeq 2000/2500 and sequence reads were aligned to the PN40024 reference genome [40,41] using BWA [42]. After using the TASSEL V3.0.166 SNP calling pipeline [39], a 509,293 SNP set was generated. The called SNPs were filtered using VCFtools [43] with a depth of read (DP) > 10, minor allele frequency (MAF) > 0.2, missing data (MD) = 1, and a genotype quality score (GQ) > 98. This filtering step reduced the SNP set to 18,124. SNPs were parsed using a pseudo test cross strategy [44]. The C2-50 and Riesling SNP sets contained 3974 and 2973 SNPs, respectively.

Fig 2. The number of sequence reads per vine.

The number of sequence reads produced for the 90 F₁ individuals and two parents are displayed. Vine and number of reads (millions) are on the x- and y-axis, respectively. The number of reads ranged from 1,348,872 to 6,021,751 per vine. The number of sequence reads obtained for C2-50 and Riesling were 7,657,007 and 6,618,727, respectively. The horizontal line indicates the average number of sequence reads.

Fig 3. Linkage mapping of MJR1 maps on LG18 at 102.6 in C2-50.

The order of the markers on LG18 was determine using the 367 SNP set plus the MJR1 marker followed by genetic mapping using R/OneMap. SNPs designated on the right side of linkage group. Genetic distance in cM is displayed on the left side of the linkage group. Note: all SNPs on LG18 were located on chromosome 18 in the PN40024 reference genome, as indicated by the position number provided in the marker name.

Fig 4. Interval mapping using the binary model for MJR1.

(A) A single LOD peak on LG18 reached a maximum of 21.1 in C2-50. Linkage group number and LOD scores are displayed on the x-axis and y-axis, respectively. (B) A diagram of LG18 displayed the LOD maximum peak at 106 cM. The SNPs and map position (cM) are shown on the x-axis and LOD scores on the y-axis. The threshold, as determined by 1000 permutations, was 3.02.

Fig 5. Linkage mapping of MJR1 using validated SNPs.

SNPs at the MJR1 locus were validated using the Sequenom MassARRAY [51]. Non-polymorphic GBS-predicted SNPs and incorrectly genotyped SNPs were removed before mapping MJR1 using the 380 SNP set. SNP ID is shown on the right and distance in cM on displayed on the left side of the linkage group.

Fig 6. Binary model of interval mapping using validated SNPs at MJR1 locus.

Interval mapping was performed with the 372 SNP set, which contains eight validated markers (S18_26580875, S18_26558715, S18_27884817, S18_30104225, S18_30236024, S18_32680428, S18_33954011). (A) A single LOD maximum of 26.7 was detected on LG18. Linkage group number and LOD scores are displayed on the x-axis and y-axis, respectively. (B) A diagram of LG18 displayed the LOD maximum at 100.3 cM with markers. Markers and map position (cM) are displayed on the x-axis and LOD scores are shown on the y-axis. The threshold, as determined by 1000 permutations, was 4.81.

Fig 7. Meloidogyne javanica ‘pt 1103P’ induced cellular necrosis in the root meristem cells of C2-50.

In vitro grown roots for (A-D) C2-50 and (E and F) Riesling. (A) C2-50 roots treated with sterile water. (B) C2-50 roots inoculated with M. javanica ‘pt. 1103P’. (B) White arrow points at M. javanica ‘pt 1103P’ induced cellular necrosis in root meristem. (C and D) C2-50 roots inoculated with M. javanica ‘pt 1103P’ and stained for nematodes using 10% bleach followed by acid fuchsin staining. (C) C2-50 roots with no cellular necrosis were not penetrated with a nematode(s). (D) C2-50 roots that displayed cell necrosis in root meristem had nematodes. Black arrow points at M. javanica ‘pt 1103P’ embedded in necrotic cells (Note: the 10% bleach treatment reduced the brown coloration of the necrotic cells in the root meristem). Riesling roots treated with (E) sterile water or (F) inoculated with M. javanica ‘pt 1103P’.