Common protein sequence signatures associate with Sclerotinia borealis lifestyle and secretion in fungal pathogens of the Sclerotiniaceae

Abstract

Fungal plant pathogens produce secreted proteins adapted to function outside fungal cells to facilitate colonization of their hosts. In many cases such as for fungi from the Sclerotiniaceae family the repertoire and function of secreted proteins remains elusive. In the Sclerotiniaceae, whereas Sclerotinia sclerotiorum and Botrytis cinerea are cosmopolitan broad host-range plant pathogens, Sclerotinia borealis has a psychrophilic lifestyle with a low optimal growth temperature, a narrow host range and geographic distribution. To spread successfully, S. borealis must synthesize proteins adapted to function in its specific environment. The search for signatures of adaptation to S. borealis lifestyle may therefore help revealing proteins critical for colonization of the environment by Sclerotiniaceae fungi. Here, we analyzed amino acids usage and intrinsic protein disorder in alignments of groups of orthologous proteins from the three Sclerotiniaceae species. We found that enrichment in Thr, depletion in Glu and Lys, and low disorder frequency in hot loops are significantly associated with S. borealis proteins. We designed an index to report bias in these properties and found that high index proteins were enriched among secreted proteins in the three Sclerotiniaceae fungi. High index proteins were also enriched in function associated with plant colonization in S. borealis, and in in planta-induced genes in S. sclerotiorum. We highlight a novel putative antifreeze protein and a novel putative lytic polysaccharide monooxygenase identified through our pipeline as candidate proteins involved in colonization of the environment. Our findings suggest that similar protein signatures associate with S. borealis lifestyle and with secretion in the Sclerotiniaceae. These signatures may be useful for identifying proteins of interest as targets for the management of plant diseases.

Keywords: Sclerotinia; amino acid usage; antifreeze protein; effector candidates; intrinsic disorder; lytic polysaccharide monooxygenase; psychrophily; secretome.

Figure 1

Sclerotinia borealis colonizes different niches than its close relatives S. sclerotiorum and Botrytis cinerea. Number of host plant genera (A) and geographic distribution (B) of the three fungal species according to the USDA Systematic Mycology and Microbiology Laboratory Fungus-Host Database (Farr and Rossman, 2015).

Figure 2

Bioinformatics pipeline for the identification of S. borealis protein sequence signatures in multiple ortholog alignments. Our pipeline uses complete predicted proteomes of S. borealis, S. sclerotiorum, and B. cinerea as inputs. It identifies orthologous protein pairs in S. borealis and S. sclerotiorum; and in S. borealis and B. cinerea using Inparanoid. Using S. sclerotiorum proteins as a reference, it identifies non-redundant core ortholog groups (COG) and overlapping regions (1). A second Inparanoid run is then used to define the longest aligned region in all three genomes (“consensus”) for each COG (2). Next, protein sequence metrics (disorder probability and amino acid frequencies) are calculated for consensus regions of all proteins (3). Finally, Wilcoxon sum rank tests are performed to identify metrics significantly different in S. borealis proteins.

Figure 3

Adaptation to S. borealis lifestyle associates with specific amino acid usage and protein disorder patterns. Distribution of the p-values of Wilcoxon sum rank tests performed to identify intrinsic disorder probabilities (A) and amino acid frequencies (B) that are significantly different in S. borealis core orthologs. For each amino acid frequency and intrinsic disorder probability, three pairwise tests were performed to compare (i) values in B. cinerea and S. sclerotiorum orthologs (p-values shown along the X-axis), (ii) values in S. borealis and B. cinerea orthologs (p-values shown along the Y-axis in green), and (iii) values in S. borealis and S. sclerotiorum orthologs (p-values shown along the Y-axis in red). Amino acid frequencies and intrinsic disorder probabilities that fell in the shaded areas were considered significantly different between S. borealis and the other fungi (p < 0.05) but not between S. sclerotiorum and B. cinerea (p>0.05). These properties were considered as associated with S. borealis lifestyle.

Figure 4

The sTEKhot index discriminates S. borealis proteins in core ortholog groups and whole predicted proteomes. (A) Overall distribution of sTEKhot values from the three fungal species within COGs. Each bubble represents a COG positioned according to the contribution of each ortholog (sTEKhot%) to the total sTEKhot of the COG. Therefore, orthologs that have similar sTEKhot values in all three species appear at the center of the plot, while COGs appear near the corner of the species harboring the ortholog with the highest sTEKhot otherwise. The size of bubbles is proportional to the sTEKhot value of S. borealis orthologs. Data points are frequent above the 40% line for S. borealis sTEKhot, and less so for S. sclerotiorum and B. cinerea sTEKhot indicating frequent higher sTEKhot values in S. borealis orthologs. (B) Species distribution of orthologs having the highest sTEKhot value in COGs. (C) Distribution of the sTEKhot index in the whole predicted proteome of the three fungi.

Figure 5

Network representation of gene ontologies (GOs) of proteins with sTEKhot >1 in S. borealis proteome. Nodes correspond to GOs are sized according to the number of proteins with sTEKhot >1. They are colored from yellow to orange according to the p-value of a hypergeometric test for enrichment in proteins with sTEKhot >1 compared to whole proteomes. White nodes are GOs not significantly enriched among proteins with sTEKhot > 1 (p>0.05). GOs labeled in bold font correspond to functions possibly associated with host interaction.

Figure 6

Predicted secreted proteins have high sTEKhot values. (A) Distribution of sTEKhot values in the proteome and the secretome of S. borealis, S. sclerotiorum and B. cinerea. (B) Proportion of predicted secreted proteins according to sTEKhot cutoff values. In complete proteomes (sTEKhot ≥ 0), the proportion of secreted proteins is ~5% in all three fungal proteomes, whereas among proteins with sTEKhot ≥ 1.5 (dotted line) it reaches an average ~70%. (C) Proportion of whole proteomes and proteins with sTEKhot > 1.5 that are secreted, contain GPI-anchors, are N-glycosylated or contain transmembrane (TM) domains. Enrich., enrichment fold among sTEKhot > 1.5 as compared to whole proteomes.

Figure 7

Proportion of S. sclerotiorum proteins encoded by genes differentially expressed in planta according to sTEKhot cutoff values. In S. sclerotiorum complete genome (sTEKhot ≥ 0), the proportion of genes induced ≥2-fold in planta is ~4.31%, whereas it reaches ~27.1%.among proteins with sTEKhot ≥ 2 (dotted line).

Figure 8

Candidate proteins associated with colonization of the environment identified based on high sTEKhot values. (A) Multiple protein sequence alignment of B. cinerea BC1G_03854 (sTEKhot = 4.29), S. borealis SBOR_9046 (sTEKhot = 10.01), S. sclerotiorum SS1G_10836 (sTEKhot = 7.34) and the hyperactive Type I antifreeze protein “Maxi” from Pseudopleuronectes americanus (4KE2_A). (B) Superimposition of Maxi antifreeze protein structure (tan) and SS1G_10836 model structure (rainbow). (C) Surface hydrophobicity of SS1G_10836 model dimer. Dotted line corresponds to the position of the section shown on the right, illustrating the characteristic hydrophilic inner core of the dimer. (D) Multiple protein sequence alignment of B. cinerea BC1G_07573 (sTEKhot = 7.07), S. borealis SBOR_1255 (sTEKhot = 3.79), S. sclerotiorum SS1G_03146 (sTEKhot = 1.58) and the AA11 Lytic Polysaccharide Monooxygenase from Aspergillus oryzae (4MAH_A). (E) Superimposition of A. oryzae AA11 structure (tan) and SS1G_03146 model structure (rainbow). (F) SS1G_10836 and SS1G_03146 gene expression in vitro (PDB, Potato Dextrose Broth), during colonization of Arabidopsis thaliana (lesion periphery and lesion center) and in sclerotia. Error bars show standard error of the mean from two independent biological replicates.