Stage-specific Proteomes from Onchocerca ochengi, Sister Species of the Human River Blindness Parasite, Uncover Adaptations to a Nodular Lifestyle

Despite 40 years of control efforts, onchocerciasis (river blindness) remains one of the most important neglected tropical diseases, with 17 million people affected. The etiological agent, Onchocerca volvulus, is a filarial nematode with a complex lifecycle involving several distinct stages in the definitive host and blackfly vector. The challenges of obtaining sufficient material have prevented high-throughput studies and the development of novel strategies for disease control and diagnosis. Here, we utilize the closest relative of O. volvulus, the bovine parasite Onchocerca ochengi, to compare stage-specific proteomes and host-parasite interactions within the secretome. We identified a total of 4260 unique O. ochengi proteins from adult males and females, infective larvae, intrauterine microfilariae, and fluid from intradermal nodules. In addition, 135 proteins were detected from the obligate Wolbachia symbiont. Observed protein families that were enriched in all whole body extracts relative to the complete search database included immunoglobulin-domain proteins, whereas redox and detoxification enzymes and proteins involved in intracellular transport displayed stage-specific overrepresentation. Unexpectedly, the larval stages exhibited enrichment for several mitochondrial-related protein families, including members of peptidase family M16 and proteins which mediate mitochondrial fission and fusion. Quantification of proteins across the lifecycle using the Hi-3 approach supported these qualitative analyses. In nodule fluid, we identified 94 O. ochengi secreted proteins, including homologs of transforming growth factor-β and a second member of a novel 6-ShK toxin domain family, which was originally described from a model filarial nematode (Litomosoides sigmodontis). Strikingly, the 498 bovine proteins identified in nodule fluid were strongly dominated by antimicrobial proteins, especially cathelicidins. This first high-throughput analysis of an Onchocerca spp. proteome across the lifecycle highlights its profound complexity and emphasizes the extremely close relationship between O. ochengi and O. volvulus. The insights presented here provide new candidates for vaccine development, drug targeting and diagnostic biomarkers.


Immunoglobulin domain proteins
Several I-set domain proteins were robustly identified (≥2 peptides) only in iuMf, including a homolog (nOo_02540) of ZIG-1 and two homologs of the C-terminal moiety of titin (supplemental Fig. S3). In C. elegans, several products of the zig gene family are expressed in the PVT interneuron of the pre-anal ganglion, and are essential for maintenance of correct axon positioning in the ventral nerve cord specifically at the L1 stage (1). The functions of titin in C. elegans are more complex, since some isoforms link the dense body of the I-band to the A-band in body wall muscle sarcomeres (2), whereas others bind to lamins in the nuclear envelope and co-ordinate nuclear organisation during interphase (3). However, since titin cannot be detected in the muscle of C. elegans L1 (2), the O. ochengi homologs are perhaps more likely to be nuclear isoforms with a role in the maturation of Mf.
A single I-set domain protein with homology to C. elegans VER-3 (nOo_01479) was observed in AF and iuMf, which is a member of the "vascular endothelial growth factor receptor related" family. The expression of this protein is restricted to the ALA neuron in the dorsal ganglion of the head and the pharyngeal saucer and anal sphincter muscles of C. elegans, suggesting a role in neuronal guidance during morphogenesis (4). Surprisingly, two I-set domain proteins were robustly identified only in AM and iuMf, comprising a homolog of UNC-89 and a member of the neuronal cell adhesion molecule family. Whereas UNC-89 is involved in assembly of thick filaments in muscle (5), neuronal cell adhesion molecules are expressed on neurons and glia where they organise cell-to-cell contacts during the formation of neural networks (6). Finally, an I-set domain protein homologous to C. elegans DIM-1 was unique in being observed only in vL3 and NF. This protein maintains a strong connection between the myofilament lattice and the muscle cell membrane in C. elegans (7), but DIM-1 homologs have also been identified as powerful immunogens from the infective L3 of gastrointestinal parasites such as Ascaris suum (8) and Trichostrongylus colubriformis (9).

Intracellular transport
A striking feature of the AF and iuMf proteomes was the significantly enrichment of domains associated with intracellular trafficking of vesicles, including Arf and "Adaptin_N" in AF; and Arf, Snf7 and WD40 in iuMf (Fig. 4, supplemental Fig. S5). ADP-ribosylation factor (Arf)-like proteins identified in iuMf included homologs of ARL-1 (also robustly detected in AF) and ARL-3 from C. elegans. When expression of these small GTPases was inhibited by RNAi, C. elegans embryos arrested at an early stage of development in the case of ARL-1, and at the 2-fold stage of elongation when ARL-3 was targeted (10). Moreover, adult hermaphrodites subjected to RNAi targeting either ARL-1 or ARL-3 were sterilised within 24 hr (10). Subsequent studies have demonstrated that ARL-1 is essential for morphological transitions of the ER during early embryogenesis (11), and both ARL proteins are required for the production of caveolin-1 bodies that fuse with the oocyte plasma membrane following ovulation and fertilization, which may be a mechanism to prevent polyspermy (12). The "Adaptin_N" domains enriched in AF were located in three subunits from adaptorrelated protein complexes (AP)-1 and AP-2, and two subunits from coatomer protein (supplemental Fig. S5). The AP complexes mediate the binding of clathrin-coated vesicles to membranes: AP-1 is involved in transport between the trans-Golgi network and endosomes, whereas AP-2 is responsible for traffic between the plasma membrane and early endosomes (13). Conversely, coatomer forms the coat of COP-I vesicles that co-ordinate retrograde transport from the Golgi back to the ER, and is dependent on Arf GTPases for membrane recruitment (14). Both AP-1 and AP-2 are key players in the regulation of Notch signalling during the binary fate decisions that underpin neurogenesis (15), and are also required for secretion of a fibroblast growth factor homolog, EGL-17, which directs the positioning of sex myoblasts in the developing vulva of C. elegans (16).
The Snf7 domain proteins that were significantly overrepresented only in iuMf included three charged multivesicular body protein (CHMP) members from the endosomal sorting complex required for transport (supplemental Fig. S5). Although these have cytoplasmic roles in the biogenesis of late endosomes and especially membrane scission (17), we identified a homolog of CHMP-1, which functions secondarily within the nuclear Polycomb group to silence genes epigenetically (18). Moreover, a WD40-domain protein from this iuMf-enriched group, Sec13 (supplemental Fig. S5), also has a dual cytoplasmic and nuclear role. This protein is found both in COP-II vesicles (associated with ER) and in the nuclear pore complex, where it localises to the kinetochores during mitosis (19). Taken together, these observations suggest that intracellular vesicular transport and nuclear gene regulation are key processes during oogenesis, embryogenesis and the maturation of iuMf.

Galectins
In O. volvulus, both Ov-GBP-1 and Ov-GBP-2 are expressed throughout the parasite lifecycle (including in iuMf), although unlike Ov-GBP-2, Ov-GBP-1 was also detected in eggshells, uterine tissues and pseudocoelomic fluid within AF (20). Furthermore, ESP of immature Mf from L. sigmodontis have been shown to contain orthologs of Ov-GBP-1 and Ov-GBP-2 (21). In the current study, only the Ov-GBP-2 ortholog, which unlike Ov-GBP-1 contains a signal peptide (20), was detected in NF (supplemental Fig. S4). Notably, two additional Ov-GBP-1like galectins (nOo_08204 and nOo_10621) were identified in all WBE except vL3 (supplemental Fig. S4). High levels of IgE reactivity to galectins have been reported both in onchocerciasis patients and in humans exposed to the dog heartworm D. immitis, suggesting a role in allergic-type immunopathology (22,23). Moreover, galectins in D. immitis ESP and expressed on the adult parasite surface have been implicated in fibrin clot lysis via tissue plasminogen activator-dependent induction of the host fibrinolytic system (24,25).

Enzymes specifically enriched in vL3
Filarial chitinases were first identified from transcripts that were found to be highly expressed in L3 cDNA libraries from the rodent filaria Acanthocheilonema viteae (26). Moreover, in O. volvulus, Ov-CHI-1 was localised exclusively within secretory granules of the glandular oesophagus in L3 (27). As revealed by RNAi experiments in A. viteae, chitinases are required for moulting of L3 (although chitin is not present in nematode cuticles) and hatching of Mf, and they may have other roles in the physiology of AF (28) and during larval development in the vector (27). In the current study, we identified a single chitinase (nOo_05212) in all WBE (supplemental Fig. S6A) with a predicted molecular weight of 42 kDa; this corresponds to the size of a second O. volvulus chitinase reported by Wu et al. (27) in adult worms of both sexes, as well as L3. The other five chitinases we detected were strictly vL3-specific (supplemental Fig. S6A), although some lacked full-length chitinase domains and probably represent split gene models. The anthelminthic closantel and a group of natural compounds, -carboline alkaloids, were recently demonstrated to inhibit the activity of Ov-CHI-1 in vitro, which led to significant inhibition of the L3 -L4 moult (29,30).
The second large enzyme family that was enriched in vL3 was the cathepsin-like proteases. Of these, only three cathepsin-L-like precursors were restricted to vL3 extracts (supplemental Fig. S6B), which corresponds with the high expression of cathepsin-L-like cysteine proteases in the L3 of both O. volvulus and Brugia pahangi, and their essential role in moulting (31,32). We also identified a cathepsin-F-like protease in all WBE and NF that was orthologous to a B. malayi transcript reported by Guiliano et al. (31); and a cathepsin-Zlike protease from all WBE that was 99% identical to that identified in O. volvulus (33) (supplemental Fig. S6B). Finally, a cathepsin-B-like protease was detected in iuMf and vL3 only (supplemental Fig. S6B). In addition to their role in moulting, cathepsin-Z-like proteases are expressed in the hypodermis, pharynx and gonads of adult C. elegans, and disruption of the gene leads to partial embryonic lethality and morphological defects at later stages (34).
The final domain specifically enriched in vL3 was the Ras superfamily of small GTPases. In the current study, most Ras proteins were detected in all WBE, but a homolog of C. elegans Rab3 (nOo_03881) was robustly identified in vL3 and iuMf only (supplemental Fig. S6C). This Ras protein is required for regulation of synaptic vesicle recruitment in the nervous system of the pharynx and is not expressed exclusively in larvae (35). However, other Ras superfamily members are specifically expressed during embryogenesis and larval development in C. elegans, and have roles in morphogenesis and axon path-finding (36,37). Thus, the functions of Ras superfamily members such as nOo_03881 deserve to be explored during filarial development.

Quantitative differences between stages
In addition to the chitinases and cathepsins already highlighted as enriched in L3, a cuticlin homolog previously identified in post-infective L3 and L4 cDNA libraries from B. pahangi (38) was one of the more abundant vL3-specific proteins (supplemental Table S5). However, as expected, the dominant protein found only in vL3 was abundant larval transcript protein-1, the key immunomodulatory protein secreted by L3 during establishment of infection (39,40). Other immunomodulators quantified in vL3 included two isoforms of activation-associated secreted proteins (ASPs), for which ASP-1 was more abundant than ASP-2 (supplemental Fig. 7A). Members of the ASP family have angiogenic properties (41) and stimulate Th-1-biased cellular responses, which has led to development of Ov-ASP-1 as an adjuvant for viral vaccines (42).
Preparations from iuMf exhibited the greatest number of stage-specific proteins, reflecting their greater proteomic diversity overall. By far the largest group in this stage were RNA-binding proteins such as splicing factors (supplemental Fig. 7A; supplemental Table S5), suggesting that the incomplete developmental status of iuMf is characterised by heightened activity in the spliceosome. The most abundant unique protein in iuMf was an enzyme involved in pyrimidine synthesis (uridine phosphorylase), although mitochondrial proteins such as malate dehydrogenase, hexokinase type II, L-2-hydroxyglutarate dehydrogenase, and apoptosis-inducing factor-1 were also well represented. Interestingly, a fatty acid retinoid-binding protein precursor orthologous to Ov-FAR-2 was very abundant in the iuMf-specific dataset, even though Ov-FAR-2 was identified from O. volvulus L3 cDNA libraries (43). This protein is clearly distinct from Oo-FAR-1 (52% coverage, 25% identity), an antigen expressed in all lifecycle stages which was evaluated in a vaccine trial against O. ochengi (44).
Sex-specific gene and protein expression have been explored in B. malayi, and in agreement with an expression array study (45), we found protein serine-threonine phosphatases (PSPs) to be enriched in AM in both the Pfam ("Metallophos", Fig. 4) and the Hi-3 analyses (supplemental Fig. 7A). Alongside the major sperm proteins (the most abundant unique proteins in AM; supplemental Fig. 7A; supplemental Table S5), PSPs are among the most distinctive male-specific genes in nematodes, which have been associated with spermatogenesis in C. elegans (46), Haemonchus contortus (47), Ascaris suum (48), Trichostrongylus vitrinus (49) and Toxocara canis (50). In C. elegans, mutation of one PSP family member, calcineurin, leads to aberrant sperm morphology and a reduction in quantity (46), whereas RNAi of another PSP group (Glc-seven phosphatases) results in defective sperm chromosome segregation (51). Other male-specific filarial proteins reported previously that are corroborated in the current study include PDZ-domain proteins [associated with AM in B. malayi (52)] and adenylate kinase isoenzyme-1, which is considered a male-enriched protein in B. malayi (52) and is also the dominant surface protein on AM in L. sigmodontis (21). Furthermore, we noted that the Pfam and Hi-3 datasets concurred in the overrepresentation of malic enzyme in AM ( Fig. 4 and supplemental Fig. 7A). Adult Onchocerca spp. are homolactate fermenters, and the presence of cytosolic NADP-dependent malic enzymes in several species has been reported previously from mixed-sex adult worm extracts (53,54).These enzymes are a key inhibitory target of the adulticidal drug, suramin (53), although to the best of our knowledge, potential differences in malic enzyme activity between AF and AM of Onchocerca spp. have not been investigated previously. Finally, AM showed a remarkable diversity of uncharacterised proteins lacking any conserved domains, several of which were highly abundant (supplemental Fig. 7A; supplemental Table S5).
In contrast with AM, AF exhibited a marked paucity of unique proteins (six in total; supplemental Fig. 7A; supplemental Table S5). Although this is probably due largely to the overlap between the AF and iuMf proteomes, the implication is that AM and AF display very similar protein expression patterns outside the reproductive tract. The most abundant stage-specific protein in AF was a cuticular collagen, although cystathionine beta-synthase was also well represented. Notably, RNAi experiments have demonstrated that this enzyme is highly expressed in murine granulosa cells and is required for oocyte maturation (55). Two other unusual proteins were uniquely abundant in AF: a homolog of a mitochondrial porin in C. elegans, TOMM-40, and a bactericidal permeability-increasing superfamily domain protein (supplemental Fig. 7A). TOMM-40 forms a hydrophilic pore in the mitochondrial translocase and is required for importation of nuclear-encoded preproteins into the organelle (56). This porin is essential for embryogenesis and larval development, since inactivation of TOMM-40 causes lack of responsiveness to the insulin-like peptide DAF-28 and larval arrest in C. elegans (57). The bactericidal permeability-increasing superfamily domain protein is more enigmatic, but another nematode protein containing this domain is NRF-5 in C. elegans, which is secreted from the intestine and may be involved in drug detection and transport, in addition to trafficking of yolk proteins to oocytes (58). Moreover, OBI-1 in Pristionchus pacificus harbours the same domain and is required for neutralisation of a sex pheromone secreted by its beetle host, which can otherwise induce arrest of embryogenesis and inhibition of exit from the dauer (dormant) larval stage (59). Intriguingly, P. pacificus OBI-1 is expressed in various tissues of the vulva as well as in the amphid neuron sheath. Taken together, these observations suggest that previously unexplored AF-enriched proteins may be essential for partitioning of resources during embryogenesis and perhaps host-parasite crosstalk or sexual communication.

DIG-1 in nodule fluid
The etymology of "DIG-1" is derived from the phenotype of C. elegans mutants, which have a displaced gonad and associated defects in the positioning of HSN neurons that innervate the uterine and vulval musculature (60), as well as a twisted pharynx (61). Unlike its homolog in O. ochengi (nOo_01384), C. elegans DIG-1 has a classical signal peptide and a predicted molecular weight of ~1.37 MDa, rendering it the largest secreted protein described to date (62) (supplemental Fig. S9). It has roles during embryogenesis in the normal fasciculation of head sensory neurons (62) and ensheaths ganglia and fascicles in the ventral nerve cord and head region of adults (63). The presence of the DIG-1 homolog (and other basement membrane proteins involved in neuron development) within the nodule fluid is surprising. These proteins may simply be expressed at high levels during embryogenesis and released during parturition, or they could have evolved to incorporate entirely different roles in the context of parasitism, such as immunomodulation or establishment of a permissive environment for the adult worms. Whatever their function, orthologs of nOo_01384 have been reported from the ESP of both B. malayi (64,65) and L. sigmodontis (21), and orthologs are also present in the genomes of D. immitis and O. volvulus. Notably, most of the filarial orthologs are approximately half the size of C. elegans DIG-1, with much reduced six-bladed -propeller domains in the central region and an absence of a signal peptide, but full retention of immunoglobulin-fold and von Willebrand factor type-A domains in the N-and C-terminal portions, respectively (supplemental Fig. S9). However, we detected a unique peptide from the C-terminal portion of the O. volvulus ortholog (OVOC8391) which was not present in nOo_01384, suggesting that the multitude of intron-exon boundaries has led to errors in gene models. Nevertheless, the L. sigmodontis ortholog appeared to be distinct from the other filarial DIG-1-like proteins in the extent of its diminution, with a predicted molecular weight of only ~226 kDa, and a complete absence of -propeller and von Willebrand factor type-A domains (supplemental Fig. S9). Interestingly, this ortholog did retain a signal peptide (supplemental Fig. S9).  . ochengi and (B) wOo, and dendrograms were generated by hierarchical clustering based on pair-wise distance. Stage-specific clusters of expression and selected proteins of interest are highlighted for iuMf (purple), vL3 (red), NF (black), AM (blue) and AF (pink). UNC, uncoordinated; G protein, guanine nucleotide-binding protein; ASP, activation-associated secreted proteins; ACAD, acyl-CoA dehydrogenase; TGF, transforming growth factor; MSP, major sperm protein; PSP, protein serine-threonine phosphatase; CBS, cystathionine beta-synthase; BPI, bactericidal permeability-increasing. Complete lists of proteins identified in each cluster are provided in supplemental Table S5.