Identification of candidate infection genes from the model entomopathogenic nematode Heterorhabditis bacteriophora

Parasitic nematodes (PN) continue to place a considerable burden on human health and agricultural production. PN diseases cause a variety of unfavorable conditions ranging from lethargy to fever, diarrhea, blindness and death. An estimated 1–1.2 billion people are infected with PNs leading to an increase of up to 52.1 million disability adjusted life years with hookworm infections contributing more than 40% of the lost years alone [1]. Current control strategies of deworming, while effective in the short-term, are inadequate due to frequent reinfection and the development of drug resistant nematode populations. Furthermore, attempts at creating a vaccine have been hindered due to the lack of good animal models and effective antigens [2].

While PN infections are understood to be immunomodulatory in nature, the molecular mechanisms of infection are poorly understood [3]. While efforts have been made to study parasitism in PNs directly, the requirement of a vertebrate host for development makes in vitro cultivation currently impossible, thereby making access to parasitic life stages and the interactions between the host immune system and the parasite difficult [4]. Even though a great deal has been learned about nematode biology and development using the free-living nematode Caenorhabditis elegans, it is not a parasite and thus does not allow for relevant investigations of PN infection mechanisms. For this reason, interest in developing Heterorhabditis bacteriophora into a model organism in order to study nematode parasitism has recently grown [5–8].

Heterorhabditis bacteriophora is a member of the Eurhabditis clade, which also contains other PNs such as the vertebrate hookworms Ancylostoma ceylanicum and Necator americanus as well as the model organism C. elegans. Heterorhabditis bacteriophora is an entomopathogenic nematode (EPN) which uses host insects and the mutualistic bacteria Photorhabdus luminescens to successfully reproduce [9]. The free-living stage of H. bacteriophora, the infective juvenile (IJ), is a developmentally arrested stage analogous to the infective stage of hookworms and the dauer of C. elegans [10, 11]. The IJs seek host insects to colonize and reproduce. Once established, the IJs resume their development and progress through the complete life cycle of H. bacteriophora. After 2 to 3 generations of reproduction, the nutrition of the host’s cadaver is exhausted and juveniles begin to arrest in mass as IJs. These IJs leave the cadaver and begin to search for a new host. Unlike hookworms and most other PNs, H. bacteriophora and its bacterial symbiont P. luminescens can be manipulated and cultured in vitro. Additionally, advanced molecular tools (e.g. gene silencing by RNAi) are being developed for H. bacteriophora as well as the recent publication of its genome, making H. bacteriophora potentially an excellent alternative model for nematode parasitism [5, 8, 12–14]. Furthermore, the ability to propagate H. bacteriophora in the immunology model Drosophila melanogaster, allows the study of host responses to PN infection mechanisms [15–17].

While the basic tools to develop H. bacteriophora as a model organism have been or are in the process of being developed [8, 13], genes directly involved in parasitism are still poorly understood. Studies examining the transcriptome of H. bacteriophora have been performed, but our results described here is the first study, to our knowledge, to utilize advanced next-generation sequencing technologies and the published H. bacteriophora genome to examine the transcriptional program during host invasion [18, 19]. A better understanding of this crucial transition period could help define parasitism genes and possibly allow the establishment of new interventions that prevent infection of hosts. To begin teasing apart the molecular biology of the early infection, we performed next generation RNA sequencing on H. bacteriophora IJs incubated in Manduca sexta hemolymph plasma for 9 h. Comparative analysis of the transcriptome with untreated controls found a number of differentially expressed genes (DEGs) which cover a number of different functional categories. Furthermore, a subset of the DEGs is conserved across Clade V parasitic nematodes. This subset of genes may serve as potential targets for future studies investigating nematode parasitism.

Parasitic nematode infections continue to pose a considerable burden to human health. However, the obligate need for a vertebrate host has made efforts to study the molecular mechanisms of parasitic infection intractable and highlights the need for an analogous model nematode. Due to its life cycle and recent studies, the possibility of developing and using H. bacteriophora as a model for parasitic infections has grown. Basic advances in the development of a molecular toolbox and the publication of its genome provides a way for manipulating and identifying the function of parasitic genes [5, 8, 13, 19]. In order to create a smaller list of genes involved in infection for future investigation, we performed RNA-seq on H. bacteriophora strain TT01 IJs incubated in hemolymph plasma for 9 h. After differential expression analysis, a total of 1641 genes were identified as being differentially expressed after a 9 h exposure to hemolymph plasma. Hemolymph plasma from Manduca sexta was selected as an activation media because M. sexta is a natural host, and its large size facilitates extraction of large quantities of hemolymph. Previous studies identified hemolymph soaking as a valid method for inducing synchronous activation of entomopathogenic IJs, while also allowing for the mass activation of IJs necessary to produce enough high quality RNA for RNA-seq [19, 22–24]. Additionally, our pilot studies used to determine the maximum concentration of IJs per milliliter hemolymph plasma showed IJs develop into J4s within 48 h of soaking and continue to adults by 72 h (data not shown). After 1 week of soaking in hemolymph plasma, multiple generations of nematodes were present.

Assignment of GO terms to DEGs categorized 769 of the genes to probable functions. Some of the most represented functional categories included reproduction, reproductive process, developmental process, binding and catalytic activity. By checking for overrepresented GO terms, we found a number of genes involved in proteolysis, acid phosphatase activity, NADH dehydrogenase (ubiquinone) activity and hydrolases, as well as sequences associated with ribosomal constituents and translation elongation factors. Also, a number of overrepresented terms were also found in categories related to development such as body morphogenesis, carbohydrate metabolism and metabolic processes, consistent with the resumption of development by the IJs.

KEGG functional categories and pathways were also assigned to DEGs. KEGG assignments showed similar categories as the overrepresented GO terms. A large number of sequences were categorically assigned ribosome, ribosome biogenesis, messenger RNA biogenesis, cytoskeleton proteins and chaperones and folding catalysts suggesting upregulation in protein production is necessary to exit the developmentally arrested dauer stage. Also seen in the KEGG functional categories again are protein phosphatases and associated proteins, ubiquitin system and peptidases. While the KEGG and GO analysis does not provide enough resolution to identify parasitic genes directly, the recurrence and overrepresentation of proteolytic enzymes suggests some of these genes may be involved in the production of proteins/catalysts necessary for parasitism. This idea is further supported by an increase in the transporters category since gene products important for PN-host interactions are likely secreted into host tissues. The KEGG pathway assignments mirror the KEGG functional categories by also showing a number of genes involved in biosynthesis of secondary metabolites, protein processing in endoplasmic reticulum, ribosome biogenesis and lysosome.

In an attempt to identify common parasitic genes, a series of protein blasts were performed against each nematode clade and also against representative nematodes of Clade V. Blasting the protein sequences of the H. bacteriophora DEGs against all nematodes included in Clades I, III, IV and V (not including H. bacteriophora) returned 172, 405, 263 and 802 gene hits respectively. As expected, the majority of hits (397 genes) were found only in Clade V due to H. bacterophora’s close phylogenetic relationship with its members. Interestingly, while there is overlap with Clades III and IV, there are 405 hits with Clade III and only 263 with Clade IV. Given their phylogenetic relationship to Clade V, hits would be expected to decrease as the distance between clades increases. The blast searches by clade also suggest a subset of “pan-nematode” genes (130) which are shared by all four of the clades surveyed. GO term assignment identifies these genes as being involved in embryo development, reproduction, nematode larval development, life cycle development and growth. Given the apparent conservation of genes related to reproduction and development, it is possible these genes represent the basic genetic needs for the reproduction and development of nematodes.

Given the relatively large number of hits with Clade V nematodes, another protein blast was performed in order to further investigate the presence of a set of PN genes. For this reason, this Clade V blast was limited to A. ceylanicum, H. contortus, N. americanus and C. elegans. While C. elegans is not a parasitic nematode, given the deep understanding and robust annotation of its genome, it was included in order to further separate Clade V nematode genes from the highly conserved genes of Clade V PNs. While a number of DEGs were found to hit each nematode surveyed, 75 genes were found to be shared by A. ceylanicum, H. contortus and N. americanus. Of these 75 genes, 30 have no GO, InterPro ID or protein motifs assigned in the ParaSite Biomart database. The 45 remaining genes contain a variety of different functional categories, but are predominantly represented by categories involved in growth and development. However, some proteins of possible parasitic interest are present. Most notably, 4 genes (Hba_14679, Hba_17394, Hba_20282 and Hba_21214) contain signal peptides with no transmembrane region. Hba_14679 contains a protein motif for leucine-rich repeats, which serve as structural supports for protein-protein interactions [25]. While a wide variety of proteins make use of these supports, given the presence of a signaling peptide, the product of Hba_14679 may play a role in host-parasite interactions external to the worm. Hba_20282 contains a motif for the delta subunit precursor of a translocon. While the exact function of this protein is unknown, translocons are known to transport peptides across membranes. Within eukaryotes translocons are commonly used to transfer polypeptides into the endoplasmic reticulum [26]. However, in prokaryotes, translocons can be assembled to export virulence factors outside of the cell [27, 28] While this protein may serve no other purpose than shuttling molecules into the endoplasmic reticulum, its conservation between 4 different parasites suggests that it may be needed for the transport of parasitism factors.

Hba_17394 and Hba_21214 are both enzymes which may be externally secreted since they both contain signal peptides. Hba_17394 is identified to have an aspartic peptidase motif. While further functional investigation is necessary for an exact identity, aspartic peptidases are known to have a range of functions from the digestion of peptides to the production of active proteins from precursor proteins [29, 30]. Within parasites, such a peptidase could be secreted for use as a virulence factor. Hba_21214 contains a protein motif for a glycoside hydrolase, family 19. Glycoside hydrolases are chitinases which break down glycosidic bonds between carbohydrates [31]. Chitinases are most commonly used as defense mechanisms to break down the cell walls of fungal and insect pathogens [32–34]. Within H. bacteriophora, the presence of a chitinase seems obligate since the digestion of chitin would be necessary upon infection of a host insect. However, given its presence across several parasitic nematodes, it is possible a secreted chitinase may serve a more general role in parasitism [35].

Given that any likely parasitism genes would potentially be upregulated upon exposure to a host, further analyses was carried out to identify any upregulated DEGs conserved within parasitic nematodes by comparison to A. ceylanicum (Bioproject PRJNA231479). Out of 861 upregulated DEGs, 239 had a false discovery rate less than 0.05. Of these 239 genes, 17 genes were identified as being potential conserved parasitism genes due to being orthologous to A. ceylanicum and containing a signal peptide without a transmembrane region and lacking orthology with free-living nematodes. The majority of these genes (16) were identified to have enzymatic protein motifs by InterPro ID. These enzymes include peptidases (7), phosphatases (2) and a kinase. The non-catalytic peptide is identified as having a transmembrane transport motif. While individually, the peptidases listed in Table 2 could be present for a number of different biological processes, their inclusion as secreted peptides orthologous to A. ceylanicum and non-orthologous with C. elegans suggests they are conserved parasitism genes. Furthermore, it is possible these genes are involved in the mechanisms of parasitism either as virulence factors or modulators of the host immune system.

Taken together, our RNA-Seq analysis reveals transcriptional changes in the regulation of a large number of genes, most of which have not been shown previously to play a role in parasitic responses. A significant proportion of these genes are conserved amongst closely related parasitic nematodes, suggesting the identification of a group of parasitism factors within nematodes. These genes provide ideal candidates for functional characterization using recently developed tools in H. bacteriophora to dissect the contribution of these genes in infection. While our selection criteria potentially misses some genes of interest, future studies using these candidates, in addition to further data-mining of our sequencing data, will not only provide functional insight into the process of nematode parasitism but may also shed light on the evolution of parasitism within nematodes as our data also reveals the existence of conserved genes amongst PN that are upregulated during early infection.