Genome, transcriptome and secretome analyses of the antagonistic, yeast-like fungus Aureobasidium pullulans to identify potential biocontrol genes

Abstract

Aureobasidium pullulans is an extremotolerant, cosmopolitan yeast-like fungus that successfully colonises vastly different ecological niches. The species is widely used in biotechnology and successfully applied as a commercial biocontrol agent against postharvest diseases and fireblight. However, the exact mechanisms that are responsible for its antagonistic activity against diverse plant pathogens are not known at the molecular level. Thus, it is difficult to optimise and improve the biocontrol applications of this species. As a foundation for elucidating biocontrol mechanisms, we have de novo assembled a high-quality reference genome of a strongly antagonistic A. pullulans strain, performed dual RNA-seq experiments, and analysed proteins secreted during the interaction with the plant pathogen Fusarium oxysporum. Based on the genome annotation, potential biocontrol genes were predicted to encode secreted hydrolases or to be part of secondary metabolite clusters (e.g., NRPS-like, NRPS, T1PKS, terpene, and β-lactone clusters). Transcriptome and secretome analyses defined a subset of 79 A. pullulans genes (among the 10,925 annotated genes) that were transcriptionally upregulated or exclusively detected at the protein level during the competition with F. oxysporum. These potential biocontrol genes comprised predicted secreted hydrolases such as glycosylases, esterases, and proteases, as well as genes encoding enzymes, which are predicted to be involved in the synthesis of secondary metabolites. This study highlights the value of a sequential approach starting with genome mining and consecutive transcriptome and secretome analyses in order to identify a limited number of potential target genes for detailed, functional analyses.

Keywords: Aureobasidium; Fusarium; antagonism; biocontrol; genome; proteome; secretome; transcriptome; yeast.

Figure 1. FIGURE 1: Comparison of five Aureobasidium species with respect to growth on agar plates and antagonistic activity.

The different Aureobasidium species grow at comparable rates in spot assays, but show differences in antagonistic activity against the plant pathogenic fungus F. oxysporum NRRL 26381/CL57. Overnight cultures of the five strains were diluted to OD₆₀₀=0.1 and three serial 1:10 dilutions were spotted on PDB agar (left). Competition assays were performed with highly diluted Aureobasidium samples (OD₆₀₀ = 0.001) and the area of F. oxysporum growth was measured to reveal the differences in their antagonistic activity. The relative growth of F. oxysporum was reduced in competition with both A. pullulans strains (NBB 7.2.1 and EXF-150), while the other Aureobasidium species were less antagonistic.

Figure 2. FIGURE 2: Whole genome comparison of five Aureobasidium species.

Gene annotations reveal that the A. pullulans NBB 7.2.1 genome contains more unique KEGG terms than the genomes of four other Aureobasidium strains. (A) KEGG term distribution among annotated genes of five Aureobasidium species. 548 terms were commonly found in all species, but a substantial number of terms (169) was unique for the A. pullulans NBB 7.2.1 genome or specifically lacking in this genome (129 terms; all bold). (B) The relative number of genes annotated to one of the six main enzyme classes (EC 1-6; oxidoreductases, transferases, hydrolases, lyases, isomerases, ligases) for all KEGG terms in the five Aureobasidium species and the subsets described in A. The enzyme classes highlighted with an asterisk (*) are significantly overrepresented (adj. p-value≤0.05) in the respective group (as compared to their frequency among all KEGG terms found in any of the five Aureobasidium genomes (①)). Results are shown for those genes with terms shared among all genomes except A. pullulans NBB 7.2.1 (②), shared among all five genomes (③), or only present in A. pullulans NBB 7.2.1 (④). (C) Pan-genome clustering of five Aureobasidium genomes visualized using an Up Set bar diagram [96]. The A. pullulans NBB 7.2.1 genome contained fewer (480) unique gene models than the four other Aureobasidium genomes.

Figure 3. FIGURE 3: Genome mining of Aureobasidium genomes to identify potential biocontrol genes.

Functional annotations of the five Aureobasidium genomes (obtained from the DOE-JGI MycoCosm) identified a plethora of genes and gene clusters that may contribute to biocontrol activity. (A) Relative percentage of A. pullulans NBB 7.2.1 genes (for the 4019 genes with a KEGG annotation or only the subset containing a predicted signal peptide) assigned to the different KEGG pathways. (B) KEGG term distribution among the five Aureobasidium genomes for all annotated genes containing a predicted signal peptide. 53 terms were commonly found in all species (③), while 47 terms were unique for the A. pullulans NBB 7.2.1 genome (④) and 21 were exclusively found in the other genomes (②) (bold numbers in Venn diagram). (C) The relative percentage of genes with a predicted signal peptide annotated to belong to one of the six main enzyme classes (EC 1-6; oxidoreductases, transferases, hydrolases, lyases, isomerases, ligases). Results are shown for genes with terms that were found in any of the five genomes (①) and for those shared in all genomes except A. pullulans NBB 7.2.1 (②), shared among all five genomes (③), or only present in A. pullulans NBB 7.2.1 (④). The enzyme classes highlighted with an asterisk were significantly overrepresented (adj. p-value≤0.05) in the respective group. Among the predicted hydrolase genes with a signal peptide, esterases (EC 3.1), glycosylases (EC 3.2), and peptidases (EC 3.4) were by far the most frequent and significantly overrepresented in genes with terms shared by all five Aureobasidium genomes.

Figure 4. FIGURE 4: Transcriptome analysis of A. pullulans NBB 7.2.1 and F. oxysporum NRRL 26381/CL57 in pure culture and during competition with each other.

A. pullulans NBB 7.2.1 strongly responds to co-cultivation with F. oxysporum NRRL 26381/CL57 at the transcriptome level. (A) Volcano plot showing expression of the A. pullulans NBB 7.2.1 genes with a signal peptide. In the co-culture, 178 genes exhibited significantly changed expression by at least a factor of four (92 and 86 up- and downregulated genes, respectively) compared to pure culture. (B) Proportion of upregulated DEGs (log2FoldChange > 2) for genes annotated to different KEGG categories. Biosynthesis of secondary metabolites (11.2%), biodegradation of xenobiotics (9.1%), metabolism of other amino acids (7.1 %), and metabolism of cofactors and vitamins (6.9 %) comprised the categories with the highest frequency of upregulated genes.

Figure 5. FIGURE 5: Identification of secondary metabolite clusters in the A. pullulans NBB 7.2.1 genome with the fungal antiSMASH v.5.1.2 online tool [41, 42].

Core biosynthetic genes of three A. pullulans NBB 7.2.1 secondary metabolite clusters were upregulated in response to co-cultuvation with F. oxysporum NRRL 26381/CL57. The core biosynthetic genes (dark red colour) of cluster 5 (NRPS and polyketide synthase cluster) (A), 7 (terpene cluster) (B), and 18 (NRPS cluster) (C) were upregulated after one day of interaction with F. oxysporum NRRL 26381/CL57.

Figure 6. FIGURE 6: Transcriptional regulation of A. pullulans NBB 7.2.1 genes with a predicted signal peptide upon co-cultivation with F. oxysporum NRRL 26381/CL57.

Many A. pullulans NBB 7.2.1 hydrolase genes were strongly up- or downregulated during the interaction with F. oxysporum NRRL 26381/CL57. (A) Percentage of upregulated DEGs in each KEGG pathway class for those A. pullulans NBB 7.2.1 genes containing a predicted signal peptide. (B) Variation of the Log2 FC values within each hydrolase enzyme category. Variation of the Log2 FC values of all A. pullulans NBB 7.2.1 DEGs is also shown on a lower level of categorisation for hydrolases acting on ester bonds (C), glycosylases (D), and peptidases (E).

Figure 7. FIGURE 7: Secretome analysis of A. pullulans NBB 7.2.1 and F. oxysporum NRRL 26381/CL57 pure cultures and a co-culture.

Supernatant from pure and co-cultures of A. pullulans NBB 7.2.1 and F. oxysporum NRRL 26381/CL57 were filtered with a 0.2 µm membrane and concentrated with > 50 kDa ultrafiltration tubes. These extracts of secreted proteins were analysed with a proteomics pipeline. Median abundance of all proteins detected in pure culture (x-axis) and during the competition (y-axis) for A. pullulans NBB 7.2.1 (A) and F. oxysporum NRRL 26381/CL57 (B). In the upper panel insert, proteins with or without a predicted signal peptide are highlighted in blue and orange color, respectively. The lower panel shows proteins of low abundance and highlights those only detected during the competition (orange) or in pure culture (blue). (C) For A. pullulans NBB 7.2.1, only few proteins were uniquely found during the competition with F. oxysporum NRRL 26381/CL57, while for F. oxysporum NRRL 26381/CL57 almost equal numbers of unique proteins were detected in the pure culture and the competition. KEGG term annotation of the proteins in the different categories is indicated.

Figure 8. FIGURE 8: Genome, transcriptome, and secretome analyses were used to reduce the number of target biocontrol genes for functional analyses from 10,925 (total number of A. pullulans NBB 7.2.1 genes) to 79.

As primary candidates for mediating biocontrol activity, genes upregulated during competition, encoding proteins detected in the secretome during the competition, and predicted to encode hydrolase or secondary metabolite biosynthesis genes were defined. The scale of the x-axis is logarithmic.