De novo transcriptome assemblies of five major European oilseed rape insect pests

Abstract

Objective

Insect pests can cause severe losses in oilseed rape yields across Europe. Genomic and transcriptomic information is very limited for these insects. The aim of our study was to provide transcriptomic resources on several oilseed rape herbivores that will support research into their biology and help develop new methods of sustainable pest management.

Data

Transcriptomes for larval stages of five major European pest species were de novo assembled by Trinity assembler. Total number of transcripts ranged from 112,247 for Ceutorhynchus pallidactylus to 225,110 for Ceutorhyncus napi. Intermediate numbers of 140,588, 140,998 and 144,504, were found for Psylliodes chrysocephala, Dasineura brassicae, and Brassicogethes aeneus, respectively. Bench-marking universal single-copy orthologues analyses for each dataset indicated high degree of completeness for all five species. The transcriptomes extend the list of genomic data on insect larvae that constitute major pests of oilseed rape. The data provide information on larval physiology and form a basis to develop highly specific RNA interference-based plant protection.

Objective

Brassica napus (oilseed rape, OSR) in the mustard family Brassicaceae is the second most important oilseed crop. However, since 1990 the average yields of OSR have been declining in Europe and Australia [1]. These declines have not mainly been linked to an increase in pests and pathogens but also to warm temperatures and low precipitation [2]. In Europe, the ban on the use of three neonicotinoids as active ingredients for seed treatment has also exacerbated pest problems [3]. Overall, insects account for an average yield loss of 13% on the global and 15% on a European scale [1].

Cabbage stem weevil Ceutorhynchus pallidactylus (Coleoptera, Curculionidae), rape stem weevil Ceutorhynchus napi (Coleoptera, Curculionidae), pollen beetle Brassicogethes aeneus (Coleoptera, Nitidulidae), cabbage stem flea beetle Psylliodes chrysocephala (Coleoptera, Chrysomelidae) and Brassica pod midge Dasineura brassicae (Diptera, Cecidomyiidae) are the most important insect pests affecting European oilseed rape crops [1]. These insects generally affect OSR in the larval stage although significant damage can also be caused by adult B. aeneus, which migrates into OSR field to feed on pollen and buds or adult P. chrysocephala that feed on seedlings [4]. The larvae of the weevils C. pallidactylus and C. napi feed inside the OSR stem, resulting in fewer pod production and severe losses [5], whereas larvae of P. chrysocephala mine the petiole [6]. The larvae of the midge D. brassicae feed on the inner pod wall and cause a complete loss of seeds [7].

Basic genomic information is lacking for these pest species except for B. aeneus (GCA_921294245.1- which is genomic assembly data) and GAPE01000000 a transcriptomic data set obtained from adult stage B. aeneus [8]. In this study, we have sequenced, assembled and annotated transcriptome assemblies for the larvae of the above-mentioned pests. This will lay the foundation for genomic and transcriptomic resources for these pest species, which can be utilized to develop methods of pest control in Brassica napus, such as host-induced gene silencing (HIGS) technology and spray-induced gene silencing (SIGS). These techniques have enormous potential as an eco-friendly alternative to pesticides for crop protection [9]. Additionally, these transcriptome resources will also be helpful to better understand the physiology and biochemistry of these insect species.

Data description

The larvae of C. napi were collected from infested summer oilseed rape after rearing the larvae for one generation. To infect the plants, beetles that were hatched in the lab from larvae collected in an experimental winter oilseed rape field in Lower Saxony, Germany in 2019 were used. Larvae of all other species were used directly for transcriptome analysis after collection. C. pallidactylus Msh. larvae were collected from the main shoots of plants from an experimental winter oilseed rape in Lower Saxony, Germany in 2020. Samples of P. chrysocephala (L.) were isolated from the petioles of the leaves of plants from an experimental winter oilseed rape field in Mecklenburg-Western Pomerania, Germany in 2020. The larvae of B. aeneus (F.) and D. brassicae (Macquart) were collected from the flowers and pods respectively of winter oilseed rape plants in an experimental field from the University of Göttingen in Göttingen in 2020. Based on macroscopic morphological characters, we sampled mixed first- to third-instar larvae of C. pallidactylus, B. aeneus, and D. brassicae, while C. napi sampling was biased to the late second- to early third-instar and P. chrysocephala sampling was biased to the first- to early second-instar.

RNA isolation and Illumina sequencing

Total RNA was extracted with Quick-RNA Tissue/Insect Kit (Zymo Research, Freiburg, Germany) according to manufacturer’s instructions. Stranded libraries were prepared from 1 µg total RNA after poly A-based mRNA isolation with NEBNext Ultra II Directional RNA library preparation kit (NEB, Frankfurt am Main, Germany) following the manufacturer’s guidelines.

2 × 100 bp paired-end read sequencing by DNA nanoball sequencing technology (DNBSEQ) was performed at BGI Genomics, Hong Kong.

Data filtering, transcriptome assembly and quality

The raw reads were assessed for quality by FASTQC and cut adapt was used to remove adaptor sequences [10]. De novo assemblies of the transcriptomes of five samples were performed from the RNA-seq data. The transcriptomes assemblies were generated with Trinity software version (2.11.0) [11] with default settings. The transcriptome assemblies were filtered for Brassica napus transcripts prior to submission to GenBank. Transvestigator [12] was implemented to prepare the transcriptome assemblies for the submission which ensures that the predicted ORFs are in positive strand (Data file 1). The associated raw data for each species are presented in Data files 2 to 6. The total number of transcripts was 112,247 for C. pallidactylus, 140,588 for P. chrysocephala, 140,998 for D. brassicae, 144,504 for B. aeneus, and 225,110 for C. napi. Based on all transcript contigs, the average contig length was highest for C. pallidactylus with 1543 bp and lowest for C. napi with 936.03 bp. Intermediate average contig lengths were found for D. brassicae, P. chrysocephala, and B. aeneus with 1476.4, 1461.07, and 1135.3 bp, respectively. Detailed assembly statistics for all samples are summarized in Data file 7. The completeness of the transcriptome assemblies was determined by BUSCO (Benchmarking Universal Single-Copy Orthologs) [13]. BUSCO searches against 3285 marker genes from diptera lineage found 84.1%, of complete universal single-copy genes for D. brassicae; BUSCO searches against 2124 marker genes from Endopterygota lineage found completeness of 95.4% for C. pallidactylus and B. aeneus, 94.3% for C. napi, and 92.9% for P. chrysocephala, which indicates a moderate to high level of completeness for all species (Data file 8).

Annotation

Functional annotation of the transcriptome assemblies generated by Trinity was performed by Trinotate v3.2.2 pipeline (https://trinotate.github.io/) [14]. The trinotate software automates the functional annotation of the transcripts and predicted protein sequences. The Trinotate pipeline employs several software including Hmmer v.3.1b1 [15], a protein domain identification (PFAM) software [16], Tmhmm v.2.0c prediction of transmembrane helices in proteins [17], Rnammer v.1.2 to predict ribosomal RNA [18], SignalP v.4.1 predicts signal peptide cleavage sites [19], prediction of gene ontology GOseq [20], eggnog v.3.0 search for orthologous groups [21]. Open reading frames (ORFs) were predicted with Transdecoder v5.5.0 (http://transdecoder.github.io) and only those ORFs were retained that had a minimum length of at least 100 amino acids, which were then blasted using BlastP against NCBI protein database with an E-value cut-off of 10 − 3 [22]. Finally, all the hits were loaded into a Sqlite database generated by Trinotate to produce a full annotation report for each species (Data file 9).

For the efficiency of HIGS and SIGS technology, the uptake and systemic transport of double stranded RNA (dsRNA) in insect pests is important [23]. One pathway possibly involved in these processes is mediated by Sid-1-like (SIL) genes, encoding dsRNA-selective, dsRNA-gated channel proteins [23]. Because the number of SIL genes are highly variable between insect species [24], we searched for expressed orthologs and found expression of SIL gene orthologues in larvae of three species: Three annotated orthologues were found in C. napi and two each in C. pallidactylus and B. aeneus (Data file 9).

Beside the uptake and systemic transport of dsRNA in insects, the development of species-specific dsRNAs for RNA interference against essential target genes is required for HIGS and SIGS technology [9]. Previously, Sect. 23, a subunit of the coat protein complex II vesicle transport complex that promote transport vesicle formation in the endoplasmatic reticulum was proven to be an effective target gene leading to high mortality in Colorado potato beetle, Leptinotarsa decemlineata [25] and western corn rootworm, Diabrotica virgifera virgifera [26]. In larvae of all five species analyzed, transcripts of orthologues of Sect. 23 were identified (Data file 9). Given their previous effectivity, these transcript sequences might provide a starting point to test the susceptibility of the five OSR pests for dsRNA-mediated gene silencing and develop species-specific HIGS and SIGS approaches.

Table 1 Overview of data files/data sets

Limitations

The insect larvae pooled were not staged i.e. they do not have specifically defined developmental stages. Additionally, no biological replicates were performed.