Novel global effector mining from the transcriptome of early life stages of the soybean cyst nematode Heterodera glycines

Abstract

Soybean cyst nematode (SCN) Heterodera glycines is an obligate parasite that relies on the secretion of effector proteins to manipulate host cellular processes that favor the formation of a feeding site within host roots to ensure its survival. The sequence complexity and co-evolutionary forces acting upon these effectors remain unknown. Here we generated a de novo transcriptome assembly representing the early life stages of SCN in both a compatible and an incompatible host interaction to facilitate global effector mining efforts in the absence of an available annotated SCN genome. We then employed a dual effector prediction strategy coupling a newly developed nematode effector prediction tool, N-Preffector, with a traditional secreted protein prediction pipeline to uncover a suite of novel effector candidates. Our analysis distinguished between effectors that co-evolve with the host genotype and those conserved by the pathogen to maintain a core function in parasitism and demonstrated that alternative splicing is one mechanism used to diversify the effector pool. In addition, we confirmed the presence of viral and microbial inhabitants with molecular sequence information. This transcriptome represents the most comprehensive whole-nematode sequence currently available for SCN and can be used as a tool for annotation of expected genome assemblies.

Figure 1

Species distribution of predicted homologues to H. glycines. Homologues were predicted using a BLASTX search against several protein databases at an e-value cutoff of 1e–5. The top 20 species with the most homologues are shown here. The resulting species evolutionary relationship was obtained from NCBI Taxonomy Browser and visualized using IcyTree.

Figure 2

H. glycines orthologs in proteomes from sequenced nematodes with diverse feeding behaviors. The interior numbers represent predicted H. glycines proteins that only have orthologs identified in one of the seven other nematode species examined. Exterior numbers represent sequenced nematode proteins with no unique orthologs in the early parasitic H. glycines transcriptome.

Figure 3

Identification and characterization of ‘Candidatus Cardinium hertigii’-associated transcripts within the H. glycines early life stage transcriptome. Transcripts from the H. glycines transcriptome were extracted and mapped against the proteome for Candidatus Cardinium hertigii to identify potential endosymbiont-associated transcripts, resulting in the identification of 468 of the 839 described proteins for this endosymbiont within the SCN early life stage transcriptome (a). Available gene ontology annotation was added to the endosymbiont-associated transcripts by BLAST2GO and grouped by the parent terms molecular function (b), cellular component (c), and biological process (d).

Figure 4

Variation of known effectors in the H. glycines early life stage transcriptome. Protein variants of previously published H. glycines effectors^– were identified using a BLASTP search at a 1e-5 cutoff and counted. Known effector sequences with >70% amino acid identity were grouped into stylet-secreted effector families (SSEF) to facilitate the analysis. Available functional annotation for effector families is indicated as follows: ANN = annexin-like; SLP1 = SNARE-like protein 1; ENG = endoglucanase; CHI = chitinase; VAP = venom allergen-like protein; CBP = cellulose-binding protein; CLE = CLAVATA3/EMBRYO SURROUNDING REGION (CLE)-like; CSP = circumsporozoite protein; CM = chorismate mutase.

Figure 5

Gene structure, protein functional domain architecture, and isoform protein products for GLAND13. Domain architecture and the retained protein domains in each of two isoforms, IS-1 and IS-2 (a). Expression of each isoform (transcripts per million) in pre-parasitic second-stage juveniles (ppJ2) and parasitic J2 (pJ2) life stages during a compatible (c) or incompatible (I) host interaction (b). The first isoform was significant for life stage change (p-value is 9.07E-11), while the second isoform was significant for both life stage and host interaction changes (p-value 1.29E-5).

Figure 6

Gene structure, protein functional domain architecture, and isoform protein products for HgCLE2. Domain architecture and the retained protein domains in each of eight isoforms, IS-1 and IS-8 (a). Shown in red are the retained introns. Each retained intron was modified as a result of AS. Dark grey boxes correspond to a modified VD1 domain due to AS. Expression of each isoform (transcripts per million) in pre-parasitic second-stage juveniles (ppJ2) and parasitic J2 (pJ2) life stages during a compatible (c) or incompatible (I) host interaction (b). Red boxes highlight transcripts that were statistically different for both life stage and host interaction groups.

Figure 7

Secreted effector protein prediction in the early life stage transcriptome of H. glycines. Predicted peptides from the transcriptome were put through two separate pipelines to identify candidate effectors. One pipeline utilized prediction of a signal peptide and lack of a predicted transmembrane domain (TMD) while the other utilized N-Preffector, a machine learning algorithm. Numbers shown here are predicted peptides remaining after each step in the pipeline.