Secretome analysis reveals effector candidates associated with broad host range necrotrophy in the fungal plant pathogen Sclerotinia sclerotiorum

Abstract

Background: The white mold fungus Sclerotinia sclerotiorum is a devastating necrotrophic plant pathogen with a remarkably broad host range. The interaction of necrotrophs with their hosts is more complex than initially thought, and still poorly understood.

Results: We combined bioinformatics approaches to determine the repertoire of S. sclerotiorum effector candidates and conducted detailed sequence and expression analyses on selected candidates. We identified 486 S. sclerotiorum secreted protein genes expressed in planta, many of which have no predicted enzymatic activity and may be involved in the interaction between the fungus and its hosts. We focused on those showing (i) protein domains and motifs found in known fungal effectors, (ii) signatures of positive selection, (iii) recent gene duplication, or (iv) being S. sclerotiorum-specific. We identified 78 effector candidates based on these properties. We analyzed the expression pattern of 16 representative effector candidate genes on four host plants and revealed diverse expression patterns.

Conclusions: These results reveal diverse predicted functions and expression patterns in the repertoire of S. sclerotiorum effector candidates. They will facilitate the functional analysis of fungal pathogenicity determinants and should prove useful in the search for plant quantitative disease resistance components active against the white mold.

Figure 1

S . sclerotiorum secretome prediction and analysis pipeline. (a) secretome analysis pipeline. We identified 745 predicted secreted proteins (yellow box) among which 486 showing experimental evidence for expression in planta (S. sclerotiorum secreted proteins expressed in p lanta, SPEPs). The number of proteins filtered out is indicated in grey with dotted arrows, the number of selected proteins is given within boxes, bioinformatics tools and resources used are indicated in blue. (b) Identification of effector candidates (ECs) based on sequence, motifs or protein domains conserved in fungal effectors. (c) Identification of ECs belonging to duplicated gene families and showing signatures of positive selection. (d) Identification of S. sclerotiorum-specific ECs of unknown function analogous to known protein folds. The results of analyses b, c and d are reported in tables and figures as indicated.

Figure 2

Sclerotinia effector candidates showing conserved domains: example of a novel class of protease inhibitors. (a) The domain organization of SS1G_01593, a S. sclerotiorum effector candidate with a peptidase inhibitor I9 domain (PF05922), and the distribution of its homologs across fungal taxonomy shown on a tree assembled based on published phylogenies, with branches color-coded from blue to red based on the percentage of species in a given order harboring homologs. (b) The predicted 3D structure of SS1G_01593 with residues color-coded based on conservation in fungi.

Figure 3

Sclerotinia effector candidates selected based on Ka/Ks ratio. (a) Distribution of Ka/Ks ratio for the 197 SPEP genes with orthologs in B. cinerea, calculated with Yn00 method on pairwise ortholog alignments. (b) Predicted 3D protein structure of SS1G_07749, a member of the glycoside hydrolase 11 xylanase family with global Ka/Ks = 2 in comparisons with B. cinerea orthologs. Residues of the 3D model are color-coded according to site-specific Ka/Ks ratios calculated using Bayesian inference with M8 model [28]. Residues with Ka/Ks > 1 are labeled on the structure. A putative Beta-D-Xylopyranose substrate molecule present in the 3b5l_B model, best structural analog of SS1G_07749, is shown as balls and sticks. The side chains of residues forming the predicted substrate binding site predicted by COFACTOR are show as sticks. The interface with plant xylanase inhibitors shown as a yellow dotted line was inferred from [53, 54] and the necrotizing peptide region shown as a grey dotted line was inferred from [9].

Figure 4

Expansion of effector candidate families in S . sclerotiorum genome. (a) Composition of clusters determined by Markov clustering of S. sclerotiorum and B. cinerea complete proteomes (MCL clusters) containing putative duplicated S. sclerotiorum SPEP genes. (b) Distribution of distances to the closest repetitive genomic element on 5’ and 3’ side for all genes (heatmap) and the genes encoding the 29 SPEPs in MCL clusters (dots). (c) Parsimony phylogenetic tree of SS1G_13371 SPEP and its 19 closest homologs. Bootstrap support calculated over 100 replicates is shown for the major branches, S. sclerotiorum clade is shown in red.

Figure 5

Taxonomic distribution of S . sclerotiorum SPEP genes across 14 fully sequenced fungal pathogen genomes. (a) Bar chart showing the number of S. sclerotiorum SPEP genes conserved along a phylogeny of fungal pathogens. Conservation was determined based on BlastP searches as described in the methods. (b) Distribution of S. sclerotiorum SPEP genes according to the number of species in which they are not conserved. S. sclerotiorum SPEP genes conserved in a given species but not in B. cinerea are shown in red.

Figure 6

Three S . sclerotiorum -specific effector candidates identified by pattern and fold recognition. (a) Superimposed 3D protein structure of SS1G_09512 model (rainbow) and the lectin domain of Streptococcus mitis lectinolysin (tan). RMSD calculated by TM-align was 3.50 Å. (b) Superimposed 3D protein structure of SS1G_12769 model (rainbow) and Necator americanus Saposin-like protein Na-SLP-1 (tan). RMSD calculated by TM-align was 2.43 Å. (c) Superimposed 3D protein structure of SS1G_13235 model (rainbow) and the C-terminal domain of Homo sapiens Death Associated Protein 5 (tan). RMSD calculated by TM-align was 2.98 Å. RMSD, root mean square deviation.

Figure 7

In planta expression analysis for selected S . sclerotiorum effector candidates on four different hosts. (a) Transcriptional profiles of 16 S. sclerotiorum effector candidate genes. Overrepresented (yellow) and underrepresented transcripts (blue) in planta are shown as log2-fold changes relative to expression in vitro, normalized using Actin expression. Hierarchical clustering based on Pearson correlation coefficients delimited five clusters. The SS1G_11173 ubiquitin 16 gene was used as a non-induced control. (b) Sequential transcriptional activation of effector gene candidates during the infection of N. benthamiana (left) and A. thaliana Sha. accession (right). (c) In planta expression pattern of candidate effector genes showing host-independent expression (SS1G_06213, left) and host-dependent expression (SS1G_08858, right). (d) Differential expression patterns of two candidate effector genes on susceptible (Sha.) and resistant (Rub.) A. thaliana accessions. Relative gene expression shown as log2-fold changes relative to expression in vitro, normalized using Actin expression. Error bars show standard deviation calculated from two technical replicates on each of three independent biological experiments. Rub., Rubezhnoe; Sha., Shahdara.