Development of a tool combining multiplex PCR and high-throughput sequencing to study the polymorphism of eight Leptosphaeria maculans avirulence genes and its application to field surveys in France

Abstract

Leptosphaeria maculans is a fungal pathogen causing stem canker of oilseed rape (Brassica napus). The disease is mainly controlled by the deployment of varieties with major resistance genes (Rlm). Rlm genes can rapidly become ineffective following the selection of virulent isolates of the fungus, i.e. with deletions or mutations in the corresponding avirulence genes (AvrLm). Reasoned and durable management of Rlm genes relies on the detection and monitoring of virulent isolates in field populations. Based on previous knowledge of AvrLm gene polymorphism, we developed a tool combining multiplex PCR and Illumina sequencing to characterise allelic variants for eight AvrLm genes in field L. maculans populations. We tested the method on DNA pools of 71 characterised L. maculans isolates and of leaf spots from 32 L. maculans isolates. After multiplex-PCR and sequencing with MiSeq technology, reads were mapped on an in-house AvrLm sequence database. Data were filtered using thresholds defined from control samples included in each run. Proportions of each allelic variant per gene, including deletions, perfectly correlated with expected ones. The method was then applied to around 1300 symptoms (42 pools of mainly 32 leaf spots) from nine B. napus fields. The proportions of virulent isolates estimated by sequencing leaf spot pools perfectly correlated with those estimated by pathotyping. In addition, the proportions of allelic variants determined at the national scale also correlated with those previously determined following individual sequencing of AvrLm genes in a representative collection of isolates. Finally, the method also allowed us to detect still undescribed and rare allelic variants. Despite the diversity of mechanisms generating virulent isolates and the gene-dependant diversity of AvrLm gene polymorphism, the method proved suitable for large-scale and regular monitoring of L. maculans populations, which will make it possible to choose effective Rlm genes and to detect resistance breakdowns at early stages.