Transcriptional analysis of chitosan-induced resistance
Priming of gene expression normally follows a characteristic pattern: differential expression is low, transient or often non-detectable after treatment with the elicitor only (i.e. Chitosan + Mock) and enhanced differential expression occurs upon subsequent infection (i.e. Chitosan + B. cinerea ) compared to infected plants that were not pretreated with the chemical (i.e. Water + B. cinerea ) (Conrath et al., 2006, Martinez-Medina et al., 2016). Importantly, the expression kinetics are also key points for the establishment of priming. To further determine the priming basis of chitosan-induced resistance, we performed whole transcriptomic analysis at 6, 9 and 12 hours post infection (hpi) with B. cinerea. These time points were selected as they cover the early, non-symptomatic start of the B. cinereainfection process. Unsupervised data analysis was first performed to observe global changes in the experiment. For this, we did a 2D principal component analysis (PCA) at different hours post infection. This analysis shows that chitosan treatments did not trigger major changes in transcription, however, it was the infection with B. cinerea which greatly impacts the experiment (Fig 3a). Moreover, whereas separation can be observed between Mock- and B. cinerea -infected replicates at 9 and 12 hpi, no obvious differences could be seen at the early time point of 6 hpi.
Genes with similar expression profiles were grouped, resulting in the identification of 1,722 differentially-expressed genes (DEGs) across all three treatments and time points. Hierarchical clustering separated the genes into four crude groups when compared to the Water + Mock treatment at the first time point (6 hpi, Fig 3b): Cluster i consists of genes that were repressed by B. cinerea infection; clusterii represents genes induced by chitosan treatment only; clusteriii includes genes repressed by B. cinerea infection and by treatment of chitosan at the later time points; cluster ivconsists of genes induced by B. cinerea infection and by treatment of chitosan only (Fig 3b). Overall patterns aligned with the previous finding that infection with B. cinerea had a large-scale, more extensive and differential response on tomato transcription compared to treatment with chitosan (Fig 3a). Moreover, the analysis demonstrates that application of chitosan results in a higher number of genes repressed than induced, with the exception of some highly induced genes in cluster iv . Distinct differences were evident between treatment with chitosan compared to infection withB. cinerea , e.g. a large group of genes in cluster ivdifferentially induced by B. cinerea at 9 and 12 h, as well as a large group of genes repressed by the pathogen in cluster i . This indicates that chitosan works as a priming agent that does not directly trigger major effects in gene transcription.
To study the different signalling pathways and specific genes responsible for priming of chitosan against B. cinerea , a two-way ANOVA identified 8,471 differentially expressed genes (DEGs) among all three treatments and time points. This global list of DEGs was subsequently used for focussed pairwise analysis to identify transcripts changing between treatments at each time-point. Venn diagrams demonstrates that the effect of chitosan on its own did not trigger major changes in gene transcription: only 15, 36 and 20 genes were differentially expressed in Chitosan + Mock vs Water + Mock treatments at 6, 9 and 12 hpi, respectively (Fig 3c). However, the effect of chitosan was much more pronounced after plants had been infected withB. cinerea . This combination resulted in the differential expression of 543, 2,011 and 2,967 genes at 6, 9 and 12 hpi, respectively, of which 260, 991 and 723 DEGs were induced only in the chitosan B. cinerea treatment (Fig 3c). In comparison, Water +B. cinerea treatments displayed differential expression of 327, 1,134, and 2,697 genes at 6, 9 and 12 hpi, respectively, of which 70, 116 and 501 DEGs were specific to the Water + B. cinereatreatment (Fig 3c). These results demonstrate that there is a subset of genes potentially responsible for chitosan-induced priming for a faster and more robust response against B. cinerea.
To further identify early-acting signalling pathways and genes involved in chitosan-induced priming, further analyses were performed on genes corresponding to the 260 probes differentially expressed only in the Chitosan + B. cinerea treatment at 6 hpi. Gene overrepresentation analysis was performed to identify biological processes and molecular functions of enriched genes. For biological processes, pathways such as response to stimulus, chemical and auxins were overrepresented (Table 1). Moreover, for molecular function, cysteine-type peptidase activity, transcription factor activity, sequence-specific DNA binding and nucleic acid binding transcription factor activity were enriched (Table 1).