Genome-wide identification and characterization of Fusarium circinatum-responsive lncRNAs in Pinus radiata

Abstract

Background: One of the most promising strategies of Pine Pitch Canker (PPC) management is the use of reproductive plant material resistant to the disease. Understanding the complexity of plant transcriptome that underlies the defence to the causal agent Fusarium circinatum, would greatly facilitate the development of an accurate breeding program. Long non-coding RNAs (lncRNAs) are emerging as important transcriptional regulators under biotic stresses in plants. However, to date, characterization of lncRNAs in conifer trees has not been reported. In this study, transcriptomic identification of lncRNAs was carried out using strand-specific paired-end RNA sequencing, from Pinus radiata samples inoculated with F. circinatum at an early stage of infection.

Results: Overall, 13,312 lncRNAs were predicted through a bioinformatics approach, including long intergenic non-coding RNAs (92.3%), antisense lncRNAs (3.3%) and intronic lncRNAs (2.9%). Compared with protein-coding RNAs, pine lncRNAs are shorter, have lower expression, lower GC content and harbour fewer and shorter exons. A total of 164 differentially expressed (DE) lncRNAs were identified in response to F. circinatum infection in the inoculated versus mock-inoculated P. radiata seedlings. The predicted cis-regulated target genes of these pathogen-responsive lncRNAs were related to defence mechanisms such as kinase activity, phytohormone regulation, and cell wall reinforcement. Co-expression network analysis of DE lncRNAs, DE protein-coding RNAs and lncRNA target genes also indicated a potential network regulating pectinesterase activity and cell wall remodelling.

Conclusions: This study presents the first comprehensive genome-wide analysis of P. radiata lncRNAs and provides the basis for future functional characterizations of lncRNAs in relation to pine defence responses against F. circinatum.

Keywords: Conifer defence; Fusarium circinatum; Pinus radiata; RNA-Seq.; Transcriptomics network; lncRNA.

Fig. 1

Survival probability plot determined using the Kaplan-Meier estimate for P. radiata seedlings inoculated with F. circinatum

Fig. 2

The pipeline of lncRNA identification and functional prediction

Fig. 3

Characterization of lncRNA transcripts showed differences with the characteristics of protein-coding transcripts in P. radiata. A Transcript size distribution for lincRNAs, lncNATs, intronic lncRNAs and protein-coding RNAs. B Number of exons per transcript for lincRNAs, lncNATs, intronic lncRNAs and protein-coding RNAs. C Exon size distributions for lincRNAs, lncNATs, intronic lncRNAs and protein-coding RNAs. D FPKM distribution of lncRNAs and protein-coding RNAs. (E) GC content of lncRNAs and protein-coding RNAs

Fig. 4

Hierarchical clustering plot of the differentially expressed lncRNAs of P. radiata in response to F. circinatum. The plot shows the scaled expression levels of these lncRNAs. Different columns represent different libraries, and different rows represent the differentially expressed lncRNAs. Red: relatively high expression; Blue: relatively low expression

Fig. 5

Enriched GO terms visualization of the DE lncRNA targeted genes constructed by REVIGO. Connections are based on the structure of the GO hierarchy. The colour of the bubble reflects the p-value obtained in the functional enrichment analysis, while its size indicates the frequency of the GO term in the underlying UniProt-GO Annotation database. Highly similar GO terms are linked by edges in the graph, where the line width indicates the degree of similarity

Fig. 6

Two co-expression modules were identified among DE lncRNAs, DEGs and targeted genes using CEMiTool package. (A) Dendrogram of samples clustered according to their condition. (B) Gene set enrichment analysis (GSEA)-based identification of two gene co-expression modules. Red coloring denotes a positive NES score, while blue coloring denotes a negative NES score. (C) Expression profiles for both expression modules (M1, M2). Each line represents a transcript and its change in expression across conditions. (D) Barplot for top GO terms enriched in M1 module. x-axis and colour transparency display - log₁₀ of the Benjamini-Hochberg (BH)-adjusted p-value. Dashed vertical line indicates BH-adjusted p-value threshold of 0.05