Network analyses predict major regulators of resistance to early blight disease complex in tomato

Abstract

Background: Early blight and brown leaf spot are often cited as the most problematic pathogens of tomato in many agricultural regions. Their causal agents are Alternaria spp., a genus of Ascomycota containing numerous necrotrophic pathogens. Breeding programs have yielded quantitatively resistant commercial cultivars, but fungicide application remains necessary to mitigate the yield losses. A major hindrance to resistance breeding is the complexity of the genetic determinants of resistance and susceptibility. In the absence of sufficiently resistant germplasm, we sequenced the transcriptomes of Heinz 1706 tomatoes treated with strongly virulent and weakly virulent isolates of Alternaria spp. 3 h post infection. We expanded existing functional gene annotations in tomato and using network statistics, we analyzed the transcriptional modules associated with defense and susceptibility.

Results: The induced responses are very distinct. The weakly virulent isolate induced a defense response of calcium-signaling, hormone responses, and transcription factors. These defense-associated processes were found in a single transcriptional module alongside secondary metabolite biosynthesis genes, and other defense responses. Co-expression and gene regulatory networks independently predicted several D clade ethylene response factors to be early regulators of the defense transcriptional module, as well as other transcription factors both known and novel in pathogen defense, including several JA-associated genes. In contrast, the strongly virulent isolate elicited a much weaker response, and a separate transcriptional module bereft of hormone signaling.

Conclusions: Our findings have predicted major defense regulators and several targets for downstream functional analyses. Combined with our improved gene functional annotation, they suggest that defense is achieved through induction of Alternaria-specific immune pathways, and susceptibility is mediated by modulating hormone responses. The implication of multiple specific clade D ethylene response factors and upregulation of JA-associated genes suggests that host defense in this pathosystem involves ethylene response factors to modulate jasmonic acid signaling.

Keywords: Alternaria; Brown leaf spot; Co-expression network; Comparative transcriptomics; Early blight; Gene regulatory network; Plant immunity; Tomato.

Fig. 1

All treatments induce distinct, mostly unique transcriptome responses. A Principal component analysis plot to visualize the distribution and clustering of samples based on their gene expression. Dots represent an individual sample replicate. Dot color represents treatment class. The axes represent the variance explained by the first two principal components. B Upset plot showing the intersections and unique sets of differentially expressed genes (DEGs) amongst the treatments. The x-axis shows the categories of the comparison sets, and the y-axis represents the number of DEGs within a comparison set. The horizontal bars represent the individual treatments, the black dots indicate which treatments are compared, and the vertical lines represent the intersection of treatments for a given comparison. The numbers above the vertical bars indicate the number of DEGs in a comparison set

Fig. 2

The early defense response to EBDC involves gene transcription, Ca2 + signaling, and hormone signaling. Expression patterns of tomato genes with significant differential expression (FDR < 0.05) in canonical immune signaling pathways after treatment with chitin or spores of either a virulent (1117–1) or avirulent (CS046) isolate of EBDC. The black dotted line is set to the up/down-regulation inflection point (LFC = 0); the red dashed line is set at the LFC threshold to be considered a DEG (LFC =|1|); blue circles and labels annotate notable genes with high differential expression

Fig. 3

WGCNA yields a discrete co-expression module associated with EBDC defense. A Visualization of the weighted gene co-expression network analysis (WGCNA). Nodes represent genes colored according to co-expression module membership. Edges indicate a measurable co-expression relationship with weaker connections having increasingly transparent lines. B Heatmap showing module-trait relationships. Rows represent co-expression modules, and columns represent the experimental treatment. Heatmap cells are colored according to the Pearson’s correlation coefficient r between a co-expression module eigengene and an experimental treatment class. Row dendrogram illustrates hierarchical clustering results for similar co-expression modules

Fig. 4

Blue module regulatory hubs show specific induction, while turquoise and yellow module regulatory hubs demonstrate generalized induction. The directional gene regulatory network (GRN) subnetworks showcase the GRN hub genes: blue for EBDC defense-associated, yellow for EBDC susceptibility-associated, and turquoise for chitin-associated co-expression modules with adjacent clustered heatmaps that are colored based on their log2 fold change (LFC). In the networks, nodes symbolize genes with their colors indicating module membership. Large nodes, annotated with their protein names, represent co-expression hub genes, while red nodes denote GRN hub genes