A genome-wide association study reveals that epistasis underlies the pathogenicity of Pectobacterium

Abstract

ABSTRACT Pectobacterium spp. are important bacterial pathogens that cause soft rot symptoms in various crops. However, their mechanism of pathogenicity requires clarity to help control their infections. Here, genome-wide association studies (GWAS) were conducted by integrating genomic data and measurements of two phenotypes (virulence and cellulase activity) for 120 various Pectobacterium strains in order to identify the genetic basis of their pathogenicity. An artificial intelligence-based software program was developed to automatically measure lesion areas on Chinese cabbage, thereby facilitating accurate and rapid data collection for virulence phenotypes for use in GWAS analysis. The analysis discovered 428 and 158 loci significantly associated with Pectobacterium virulence (lesion area) and cellulase activity, respectively. In addition, 1,229 and 586 epistasis loci pairs were identified for the virulence and cellulase activity phenotypes, respectively. Among them, the AraC transcriptional regulator exerted epistasis effects with another three nutrient transport-related genes in pairs contributing to the virulence phenotype, and their epistatic effects were experimentally confirmed for one pair with knockout mutants of each single gene and double gene. This study consequently provides valuable insights into the genetic mechanism underlying Pectobacterium spp. pathogenicity. Importance Plant diseases and pests are responsible for the loss of up to 40% of food crops, and annual economic losses caused by plant diseases reach more than $220 billion. Fighting against plant diseases requires an understanding of the pathogenic mechanisms of pathogens. This study adopted an advanced approach using population genomics integrated with virulence-related phenotype data to investigate the genetic basis of Pectobacterium spp., which causes serious crop losses worldwide. An automated software program based on artificial intelligence was developed to measure the virulence phenotype (lesion area), which greatly facilitated this research. The analysis predicted key genomic loci that were highly associated with virulence phenotypes, exhibited epistasis effects, and were further confirmed as critical for virulence with mutant gene deletion experiments. The present study provides new insights into the genetic determinants associated with Pectobacterium pathogenicity and provides a valuable new software resource that can be adapted to improve plant infection measurements.

Keywords

• Pectobacterium spp.
• genome-wide association studies (GWAS)
• pathogenicity
• epistasis
• artificial intelligence