Haplotype-resolved powdery mildew resistance loci reveal the impact of heterozygous structural variation on NLR genes in Muscadinia rotundifolia

Abstract

Muscadinia rotundifolia cv. Trayshed is a valuable source of resistance to grape powdery mildew. It carries 2 powdery mildew resistance-associated genetic loci, Run1.2 on chromosome 12 and Run2.2 on chromosome 18. The purpose of this study was to identify candidate resistance genes associated with each haplotype of the 2 loci. Both haplotypes of each resistance-associated locus were identified, phased, and reconstructed. Haplotype phasing allowed the identification of several structural variation events between haplotypes of both loci. Combined with a manual refinement of the gene models, we found that the heterozygous structural variants affected the gene content, with some resulting in duplicated or hemizygous nucleotide-binding leucine-rich repeat genes. Heterozygous structural variations were also found to impact the domain composition of some nucleotide-binding leucine-rich repeat proteins. By comparing the nucleotide-binding leucine-rich repeat proteins at Run1.2 and Run2.2 loci, we discovered that the 2 loci include different numbers and classes of nucleotide-binding leucine-rich repeat genes. To identify powdery mildew resistance-associated genes, we performed a gene expression profiling of the nucleotide-binding leucine-rich repeat genes at Run1.2b and Run2.2 loci with or without powdery mildew present. Several nucleotide-binding leucine-rich repeat genes were constitutively expressed, suggesting a role in powdery mildew resistance. These first complete, haplotype-resolved resistance-associated loci and the candidate nucleotide-binding leucine-rich repeat genes identified by this study are new resources that can aid the development of powdery mildew-resistant grape cultivars.

Keywords: genetic resistance; genomic structural variation; haplotype phasing; nucleotide-binding leucine-rich repeat genes.

Fig. 1.

Detailed parentage of e6-23 (Run1.2b⁺) and 08391-029 (Run2.2⁺). F2-35 was produced by crossing V. vinifera Cabernet Sauvignon with V. vinifera Carignane. OP, open pollinator which is assumed V. vinifera.

Fig. 2.

Haplotype comparison and NLR content at Run1.2 and Run2.2 in M. rotundifolia Trayshed. Whole-sequence alignments of the reconstructed haplotypes of Run1.2 (a) and Run2.2 (b) loci. Normalized median DNA-seq coverage per 10 kb of e6-23 (Run1.2b⁺) and 08391-029 (Run2.2⁺) on the diploid genome of M. rotundifolia Trayshed was used to identify Run1.2b and Run2.2 on the Haplotype 2 of chromosome 12 and Haplotype 1 of chromosome 18, respectively. Only DNA-seq reads aligning perfectly on the diploid genome of M. rotundifolia Trayshed were used for base coverage analysis. Chromosomal position of the Run1.2- and Run2.2-associated genetic markers is indicated by black triangles and dashed lines. Chromosomal positions of the TIR-NBS-LRR genes whose predicted proteins cluster with G52’s MrRUN1 and MrRPV1 in the phylogenetic tree of Fig. 3a are indicated by gray and white triangles, respectively.

Fig. 3.

Comparison of the NLRs composing Run1.2 and Run2.2 in M. rotundifolia Trayshed. (a) Neighbor-joining clustering of the predicted protein sequences of the NLRs composing Run1.2 and Run2.2 haplotypes, and Run1/Rpv1 (Feechan et al. 2013). (b) Domain diagram of Trayshed’s TIR-NBS-LRRs clustering with G52’s MrRUN1 and MrRPV1. Proteins are reported using same number assigned in (a). LRR motifs were identified using the consensus sequence LxxLxLxx, with L indicating a leucine residue and x indicating any amino acid (Kajava and Kobe 2002).

Fig. 4.

Transcript abundance of NLR genes in Run1.2b (a) and Run2.2 (b). Gene expression was monitored in Run1.2b⁺ and Run2.2⁺ at 1 and 5 dpi with E. necator conidia (PM) or a mock solution (Mock). Transcript abundance is shown as the mean of TPM. n = 3. NLR genes differentially expressed in response to PM are indicated by an asterisk (P ≤ 0.05). White triangle indicates the chromosomal position of the TIR-NBS-LRR gene whose predicted protein clusters with G52’s MrRPV1 in Fig. 3a.