Gene expression differences between western redcedar seedlings resistant and susceptible to cedar leaf blight

Western redcedar (T. plicata) is an important Cupressaceae both at economic and cultural levels in the Pacific Northwest of North America. In adult trees, the species produces one of the most weathering-resistant heartwoods among conifers, making it one of the preferred species for outdoor applications. However, young T. plicata plants are susceptible to infection with cedar leaf blight (D. thujina), an important foliar pathogen that can be devastating in nurseries and small-spaced plantations. Despite that, variability in the resistance against D. thujina in T. plicata has been documented, and such a variability can be used to breed T. plicata for resistance against the pathogen. This investigation aimed to discern the phenotypic and gene expression differences between resistant and susceptible T. plicata seedlings to shed light on the potential constitutive resistance mechanisms against cedar leaf blight in western redcedar. The study consisted of two parts. First, the histological differences between four resistant and four susceptible families that were never infected with the pathogen were investigated. And second, the differences between one resistant and one susceptible family that were infected and not infected with the pathogen were analyzed at the chemical (C, N, mineral nutrients, lignin, fiber, starch, and terpenes) and gene expression (RNA-Seq) levels. The histological part showed that T. plicata seedlings resistant to D. thujina had constitutively thicker cuticles and lower stomata densities than susceptible plants. The chemical analyses revealed that, regardless of their infection status, resistant plants had higher foliar concentrations of sabinene and α-thujene, and higher levels of expression of transcripts that code for leucine-rich repeat receptor-like protein kinases and for bark storage proteins. In conclusion, the data collected in this study shows that constitutive differences at the phenotypic (histological and chemical) and gene expression level exist between T. plicata seedlings susceptible and resistant to D. thujina. Such differences have potential use for marker-assisted selection and breeding for resistance against cedar leaf blight in western redcedar in the future.

Methods
This dataset includes transcriptomic and gene expression data from an investigation on the impact of Didymascella thujina (cedar leaf blight) on Thuja plicata (western redcedar). The datasets included in this submission were collected and produced as follows below. The information shown here was extracted from the manuscript:

RNA extraction, mRNA enrichment, library production and sequencing

RNA extraction from foliage of three CLB+ and three CLB- seedlings was done using a modified version of the protocol of Rajakani et al. (2013) [...]. mRNA enrichment was done using protocol C of the Thermo Scientific™ MagJET mRNA Enrichment Kit (Life Technologies Inc., Burlington ON, Canada). Libraries were made using the NEB Next® Ultra™ RNA Library Prep Kit for Illumina® v. 1.2. (New England BioLabs® Inc., Ipswich MA, USA). DNA was purified as required using the Thermo Scientific GeneJET NGS Cleanup Kit (Life Technologies Inc.), and size selection (∼450 bp fragment size) was completed with the Thermo Scientific MagJET NGS Cleanup and Size Selection Kit (Life Technologies Inc.). Libraries were barcoded using the NEB Next® Multiplex Oligos for Illumina® - Index Primers Set 1 (New England BioLabs® Inc.). Quality control and quantification of the individual libraries was done with a DNA 1K Analysis Kit (Bio-Rad Laboratories, Mississauga ON, Canada) in an Experion™ Automated Electrophoresis Station (Bio-Rad Laboratories). The final pool consisted of 40 ng of DNA per library. Pair-ended 100 base sequencing was completed in a single lane of an Illumina® HiSeq 2000 sequencer at Genome Quebec Innovation Centre (Montreal QC, Canada).

Assembling and annotation of the reference transcriptome

[...] All of the processes described below were completed on the WestGrid Hermes cluster (https://www.westgrid.ca/) hosted at the University of Victoria using customized shell, Python and R scripts. HPC GridRunner was used to enhance annotation searches such as BLAST and HMMER.

Paired-end FASTQ Illumina® 1.9 (Phred-33 ASCII) compressed files were produced for each sample after sequencing. Each file was checked for quality before and after trimming using FastQC v. 0.11.2 (Andrews, 2014). Trimming was done in Trimmomatic v. 0.33 (Bolger et al. 2014; [...]). The reference transcriptome was built using Trinity v. 2.0.6 (Grabherr et al., 2011) with the default settings for paired-end data, and its statistics were calculated in PRINSEQ v. 0.20.1 (Schmieder and Edwards, 2011). Annotation was completed using Trinotate v. 2.0.2 (http://trinotate.github.io; [...]).

Differential expression analyses

The downstream analyses (Haas et al., 2013) were conducted using the assembled contigs and contig variants from Trinity v. 2.0.6 (Grabherr et al., 2011) instead of the smaller number of corresponding deduced genes. Trinity refers to the contigs and variants as “transcripts” [...]. Reads were mapped to the assembly with RSEM v. 1.2.20 (Li and Dewey, 2011) and fragments per kilobase of transcript per million mapped (FPKMs, Trapnell et al. 2010) were calculated. Normalization was achieved in edgeR (Robinson et al., 2010) by computing the trimmed mean of M-values (TMM; Dillies et al. 2013, Robinson and Oshlack 2010). The differential expression (DE) analysis was completed by comparing all samples in pairs using the default settings in edgeR, and then extracting and merging the sequences that had a minimum fold-change of four and a maximum false discovery rate of 0.001 from all the samples into a single matrix.