Genome-Wide Analysis of Corynespora cassiicola Leaf Fall Disease Putative Effectors

Abstract

Corynespora cassiicola is an Ascomycetes fungus with a broad host range and diverse life styles. Mostly known as a necrotrophic plant pathogen, it has also been associated with rare cases of human infection. In the rubber tree, this fungus causes the Corynespora leaf fall (CLF) disease, which increasingly affects natural rubber production in Asia and Africa. It has also been found as an endophyte in South American rubber plantations where no CLF outbreak has yet occurred. The C. cassiicola species is genetically highly diverse, but no clear relationship has been evidenced between phylogenetic lineage and pathogenicity. Cassiicolin, a small glycosylated secreted protein effector, is thought to be involved in the necrotrophic interaction with the rubber tree but some virulent C. cassiicola isolates do not have a cassiicolin gene. This study set out to identify other putative effectors involved in CLF. The genome of a highly virulent C. cassiicola isolate from the rubber tree (CCP) was sequenced and assembled. In silico prediction revealed 2870 putative effectors, comprising CAZymes, lipases, peptidases, secreted proteins and enzymes associated with secondary metabolism. Comparison with the genomes of 44 other fungal species, focusing on effector content, revealed a striking proximity with phylogenetically unrelated species (Colletotrichum acutatum, Colletotrichum gloesporioides, Fusarium oxysporum, nectria hematococca, and Botrosphaeria dothidea) sharing life style plasticity and broad host range. Candidate effectors involved in the compatible interaction with the rubber tree were identified by transcriptomic analysis. Differentially expressed genes included 92 putative effectors, among which cassiicolin and two other secreted singleton proteins. Finally, the genomes of 35 C. cassiicola isolates representing the genetic diversity of the species were sequenced and assembled, and putative effectors identified. At the intraspecific level, effector-based classification was found to be highly consistent with the phylogenomic trees. Identification of lineage-specific effectors is a key step toward understanding C. cassiicola virulence and host specialization mechanisms.

Keywords: Corynespora cassiicola; Hevea brasiliensis; cassiicolin; effectors; gene expression; genomics; plant-pathogens interaction.

Figure 1

Corynespora cassiicola strain CCP and the Corynespora Leaf Fall (CLF) disease symptoms. (A) CCP mycelium colony on PDA medium 7 days after subculture. (B) Optical microscopy view of CCP conidia in water. (C) Hevea brasiliensis clone PB217 leaflet inoculated with CCP spores (7 days post-inoculation, 500 spores per spot). (D) CLF symptoms on a susceptible rubber tree clone (left, RRIC103) next to a tolerant clone (right, GT1), in Nigeria.

Figure 2

Interspecific phylogenomic tree, protein clusters and putative effectors of 45 fungal species. Details of the species and their genome access links are listed in Table S1. Life styles are indicated by a letter: S, saprotrophic; N, necrotrophic; H, hemibiotrophic; B, biotrophic; E, ectomycorrhizal. The maximum likelihood phylogenetic tree was based on 651 concatenated core protein sequences. Branch lengths are indicated by the bar (substitutions/site); 1,000 bootstrap values are shown as percentages. Clustered and non-clustered (species-specific) protein numbers were predicted based on OrthoMCL clustering. Putative effectors: number of gene models encoding CAZymes, peptidases, lipases and other secreted proteins (left diagram), or involved in the secondary-metabolism (NRPS, PKS, terpenes synthases, right diagram).

Figure 3

Interspecific principal component analysis (PCA) of 45 fungal species based on their putative effector composition. Species are represented by their alias name (Table S1). PCA was computed from the counts in each effector category as represented in Figure 2 (14 dimensions). Different colors represent different clusters.

Figure 4

RNA-Seq heatmaps of CCP putative effectors differentially expressed 24 and 48 h after spore inoculation on detached rubber tree leaves (susceptible clone PB260). Differential expression was calculated as the Log2 fold change at each time point using the spore suspension as reference. The code numbers refer to JGI transcript IDs (http://genome.jgi.doe.gov/Corca1) except for cassiicolin Cas1 (888888) which was manually annotated. Genes with predicted secretion signals are noted with an asterisk. Colored bars on the left side define the three clusters obtained by hierarchical classification of the differential expression values. (A) CAZymes; (B) other effectors.

Figure 5

Intraspecific phylogenomic tree and putative accessory effectors of C. cassiicola isolates. The maximum likelihood phylogenomic tree was based on 12,420 conserved protein sequences. Branch lengths are indicated by the bar (substitutions/site); 1,000 bootstrap values are shown as percentages. The isolates names are followed by their country and host codes, and by the toxin class, as indicated in Table 1. The diagram on the right represents the composition in putative accessory effectors (i.e. either absent or varying in copy number in at least one isolate).

Figure 6

Intraspecific principal component analysis (PCA) of 36 plant-associated C. cassiicola isolates based on their composition in putative accessory effectors (i.e. either absent or varying in copy number in at least one isolate). PCA was computed from the counts in each effector subcategory (611 dimensions). Different colors represent different clusters.