2.1 Ethics statement

Ethics approval for infecting Syrian hamster with N. americanus was obtained from the Animal ethics Committee (CEUA) of the Federal University of Minas Gerais (Protocol# 51/2013) according to the Brazilian Guidelines on Animal Work. Ethics approval for infecting Syrian hamsters with A. ceylanicum was obtained from the Institutional Animal Care and Use Committee (IACUC) of the Baylor College of Medicine (protocol HA/HC658). Experimental human hookworm infections and egg collection for both N. americanus and A. ceylanicum were conducted with written informed consent from all participants according to the National Health and Medical Research Council (NHMRC) National Statement on Ethical Conduct in Human Research under the James Cook University Human Research Ethics Committee (JCU HREC) approval H8339. During the experimental infection study, participants’ health status were monitored via weekly self-assessed symptom scoring and regular lab visits to provide blood samples for pathology tests. Participants were advised on risks of hookworm release into the environment and were advised to maintain sanitation practices. Upon completion of the study, participants were offered the choice to keep their worm infections to maintain the lifecycle for provision of research material or to deworm with over-the-counter deworming medication.

2.2 ES products of adult hookworms

Adult N americanus ES products were isolated and purified as described [15]. Briefly, N. americanus adults were recovered at necropsy from Syrian hamsters, washed in PBS and maintained at 37°C, 5% CO₂ in RPMI 1640 supplemented with 100 μg/ml streptomycin, 100U/ml penicillin and 0.25 μg/ml amphotericin B. Culture supernatant was harvested the next day, centrifuged at 1,500 x g for 3 minutes and the supernatant representing the ES products aliquoted and stored at -80°C. Adult A. ceylanicum were recovered from hamsters at day 40 after infection with 150 A. ceylanicum L3. Adult worms were similarly maintained (above) in RPMI 1640 supplemented with 100 U/ml ampicillin, 100 μg/ml streptomycin, culture supernatant harvested 24 hours later, clarified by centrifugation at 1,500 x g for 10 minutes, passed through a 0.45 μm filter, concentrated, buffer exchanged to PBS pH 7.4 in a Centricon-3 microconcentrator (Amicon, USA), and stored at -80°C.

2.3 L3 ES products

Fecal samples were obtained from one healthy volunteer each previously experimentally inoculated with either 10 A. ceylanicum or N. americanus L3 larvae via the skin on two separate occasions, approximately 80–90 days between inoculations. The fecal material containing either A. ceylanicum or N. americanus eggs were cultured with activated charcoal on 15cm petri dishes at 25°C in a humid chamber for seven days. L3 that migrated to the edge of the dish were harvested and transferred to a 50 ml tube. The larvae were then washed five times in PBS supplemented with 500 μg/ml streptomycin, 500 U/ml penicillin and 1.25 μg/ml amphotericin B (5X Gibco Antibiotic-Antimycotic, Thermo Fisher Scientific, USA), followed by three washes in 200 μg/ml streptomycin, 200 U/ml penicillin and 0.5 μg/ml amphotericin B (2X Gibco Antibiotic-Antimycotic). The larvae were then cultured at 37°C in a tissue culture flask with RPMI 1640 supplemented with 200 μg/ml streptomycin, 200U/ml penicillin and 0.5 μg/ml amphotericin B (2X Gibco Antibiotic-Antimycotic), 2 X L-glutamine and 1% glucose at a concentration of 5000 larvae per ml. Culture supernatant was harvested 48 hours later, centrifuged at 500 x g, 2000 x g and 4000 x g for 30 minutes at 4°C before being stored at -80°C for later use.

2.4 S-trap proteomics

Culture supernatants containing ES products were pooled and concentrated and buffer exchanged in 3 kDa Amicon Ultra-15 centrifugal filter with at least four volumes of 50 mM tetraethylammonium bromide (TEAB) at 4°C and 4,000 x g. In total, 105 μg of N. americanus adult ES, 41 μg of N. americanus L3 ES, 130 μg of A. ceylanicum adult ES, and 55 μg of A. ceylanicum L3 ES products split across two replicates were prepared. Samples were lyophilized and solubilized in 50 mM TEAB 5% (w/v) SDS and processed in S-trap micro columns (Protifi, USA) according to the manufacturer’s instructions. In short, solubilized proteins were reduced with 20 mM tris(2-carboxyethyl)phosphine hydrochloride for 15 minutes at 55°C and alkylated with 40 mM iodoacetamide in the dark for 10 minutes at room temperature followed by acidification to a final concentration of around 2.5% (v/v) phosphoric acid and incubation with 100 mM TEAB in 90% (v/v) methanol (binding/wash buffer) to facilitate binding to the S-trap column. Proteins were washed on the S-trap three times with binding/wash buffer and incubated with Solu-trypsin (Sigma, USA) at a 1:10 trypsin to protein ratio for 18 hours. After incubation, peptides were eluted sequentially with 50 mM TEAB, then 0.2% (v/v) formic acid followed by 50% (v/v) acetonitrile and lyophilized for further processing.

2.5 Solid-phase sample clean-up

Solid-phase stage tips were prepared by stacking Empore Octadecyl C18 (Supelco/Sigma-Aldrich, USA) disc punches in a 200 μl pipette tip. All centrifugation steps were performed at 1,000 x g for at least 1 minute. Stage tips were supported with an adaptor over microtubes and C18 disc punches were centrifuged with 100 μl of methanol followed by 100 μl of 0.2% (v/v) trifluoroacetic acid (TFA) in water. Tryptic peptides eluted from the S-traps were resuspended in 0.2% (v/v) TFA and loaded onto the stage tip over two rounds of centrifugation. Peptides bound to the C18 discs were washed three times with 200 μl of 0.2% (v/v) TFA before a final elution step with 100 μl of 60% acetonitrile 0.2% formic acid (v/v) in water. Eluted peptides were dried, lyophilized, resuspended in 0.1% (v/v) formic acid in water (Solvent A), and submitted for mass spectrometry analysis.

2.6 Data acquisition

All samples were analyzed in duplicate with a NanoLC 415 (Eksigent, USA) liquid chromatography system running on Solvent A and 0.1% (v/v) formic acid in 100% HPLC grade acetonitrile (Solvent B) coupled to a TripleTOF 6600 (AB Sciex, USA) mass spectrometer running the Analyst TF 1.8.1 software (Sciex, USA). Digested peptides from all replicates were first loaded on a Proteocol C18 G 3μm 200Å 10 x 0.3 mm trap column (Trajan, Australia) under 10 μl/minute of Solvent A for 5 minutes and separated on a Proteocol C18 G203 3 μm 200Å 250 × 0.3 mm column (Trajan, Australia) with linear gradient of 3–40% solvent B over 90 minutes and at microflow rates of 5 μl/minute. Mass spectrometer parameters were: curtain gas = 35, ion source gas 1 = 25, ion source gas 2 = 30, ionspray voltage floating = 5000, and turboheater temperature = 300°C. Ions were collected and analyzed with 250 ms TOF MS, a collision energy of 5 and a declustering potential of 80 followed by 31 experiments of 50 ms product ion data dependent acquisition scans on product ions with a charge state between 2–5, a collision energy spread of 5, and a declustering potential of 80. The mass windows were 400–1250 Da and 100–1500 Da for the TOF MS and product ion scans, respectively.

2.7 Database search and protein identification

Mass spectrometry data files for each replicate were pooled by species and life stage and searched against the respective protein databases reported for A. ceylanicum [16] and N. americanus [15] appended with the respective host proteome database from UniProt Mesocricetus auratus (hamster) proteome UP000189706 retrieved 24^th May 2023 for adult life stages and Homo sapiens proteome UP000005640 retrieved 23^rd May 2023 for L3 life stages and 381 sequences from the universal protein contaminant library [17] using the Fragpipe V20.0 software [18, 19]. The indexed retention time peptide (Biognosys, Switzerland) sequence was also added to the database as it was added before the acquisition step. Combined databases were appended with an equal number of Fragpipe generated decoy proteins. Searches were performed with the following parameters: precursor mass tolerance of 20 ppm, fragment mass tolerance of 20 ppm, isotype error of 0/1/2, with mass calibration and parameter optimization enabled. Enzyme parameters were set to strict trypsin with two missed cleavages allowed and a peptide length of 7–50 amino acids. Variable modifications of Met oxidation (max of three) and protein N-terminal acetylation with up to a total of three modifications per peptide were allowed. All fixed modifications were enabled. MSBooster was enabled to predict RT and Spectra. PSM validation and Percolator was enabled to with a minimum probability of 0.5. ProteinProphet was set to a maxppmdiff of 2,000,000. The false discovery rate filter was set to 0.01. Search results were further filtered to exclude any proteins from contaminants, the host proteome, and proteins with less than two unique peptides detected.

2.8 Protein functionality annotation database construction and functional enrichment testing

All proteins in the current N. americanus protein set [15] were assigned functional annotations using results from InterProScan v5.42 [20] to identify gene ontology [21] classifications and InterPro functional domains [22], GhostKOALA v2.2 [23] for KEGG [24] annotations, and PANNZER [25] and Sma3s [26] for additional gene naming. Proteins predicted to be secreted were annotated using SignalP v5.0 [27] to identify signal peptides and transmembrane domains. Proteins were characterized as putatively secreted if they contained a signal peptide and contained fewer than two transmembrane domains. The same information was retrieved from a database provided in the previously published A. ceylanicum annotation [16]. For both species, significant functional enrichment for GO terms was performed using GOSTATS v2.50 [28] and for InterPro domains and KEGG pathways using WebGestalt v2019 [29] (FDR-adjusted P ≤ 0.05, minimum two proteins detected, against a background of all annotated proteins for all enrichment testing). For both hookworm species, a comprehensive list of cysteine-rich secretory / SCP/TAPS / CAP domain proteins were identified by searching for associated InterPro domains (IPR035940, IPR001283, IPR014044, IPR035940, IPR035109, IPR002413, and IPR018244).

Proteins with shared orthologs among nematode species were quantified using BLAST [30] (NCBI blastp v2.7.1+, default settings) to identify sequence similarity hits (E ≤ 10⁻⁵) between each pairwise combination of A. ceylanicum [16], N. americanus [15], and N. brasiliensis [31]. "Reciprocal best BLAST hits" were identified when two proteins from different species were each other’s best match across the proteins. That is, Protein A from species one is best matched to Protein B from species two, and protein B from species two is best matched to protein A from species one. The reciprocal best hit criterion is used to reduce false positives and increase confidence in the assignment of orthologous relationships, and to ensure that results are 1:1 comparable across species. These "reciprocal orthologs" were used to directly compare orthologs across species. In this study, orthologs shared between A. ceylanicum, N. americanus, and N. brasiliensis are referred to as "three-species orthologs". Orthologs of Schistosoma mansoni proteins (PRJEA36577.WBPS16 [32]) were also identified for each A. ceylanicum protein using the same BLAST approach.

Previously-published raw RNA-seq reads for A. ceylanicum (19 samples across the life cycle stages L3, L4, adult female and adult male [33]) and N. americanus (two L3 exsheathed and two mixed adult replicates [34]) were downloaded. Trimmomatic v0.36 [35] was used to trim adapters and the STAR aligner [36] (v2.7.5b; 2-pass mode, basic) was used to map RNA-seq reads to the current respective assemblies downloaded from WormBase Parasite [32]. Read counts for every gene in every sample were annotated using the current A. ceylanicum [16] and N. americanus [15] gene annotations using featureCounts [37]. For every protein, complete functional annotations, peptide and spectral counts per proteomics sample, BLAST results against other species and normalized gene expression values (FPKM) for all RNA-seq samples are provided for N. americanus (S1 Table) and for A. ceylanicum (S2 Table).

3.1 Proteomic results overview and functional annotation

Proteomics samples were collected and quantified for samples from ES products from N. americanus L3 (1,594 spectra identifying 580 peptides in 103 proteins), N. americanus adult (1,371 spectra identifying 690 peptides in 106 proteins), A. ceylanicum L3 (936 spectra identifying 290 peptides in 57 proteins), and A. ceylanicum adults (13,451 spectra identifying 3,441 peptides in 512 proteins). Relative abundance values for each protein are represented by the normalized spectral abundance factor (NSAF), representing total spectral counts normalized to protein length and total spectral intensity in each sample. Among N. americanus life stages, 91 proteins were L3-specific and 94 were adult-specific ESP, while 12 ESPs were shared between the two stages (Fig 1A). For A. ceylanicum, 41 proteins were L3-specifc compared to 496 adult-specific ESP, while 16 ESPs were shared between these two life cycle stages (Fig 1B). Identified proteins were superimposed over a distribution plot of previously published RNA-seq gene expression levels to compare both life stages for N. americanus [34] (Fig 1C) and A. ceylanicum [33] (Fig 1D). In both species, proteins detected only in the adult secretome (blue) tended to have higher gene expression in the adult stage while most proteins detected only in the L3 stage (yellow) had corresponding higher gene expression in the L3 stage. Most proteins detected in both life stages appeared to have overall a relatively high gene expression compared to the other life stage-specific proteins (Fig 1C and 1D, Red).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The number of detected ESPs and their relative gene expression levels in L3 and adult stages for both N. americanus and A. ceylanicum.

Venn diagrams showing the counts of proteins in each stage are shown for (A) N. americanus and (B) A. ceylanicum. The relative gene expression levels of detected proteins and all other genes (based on published datasets) are shown for (C) N. americanus and (D) A. ceylanicum.

https://doi.org/10.1371/journal.pntd.0012780.g001

To facilitate inferred functional characterization and enrichment analysis for detected protein sets of interest, databases were constructed to functionally annotate the current N. americanus [15] proteome, including InterPro functional domains [22], KEGG enzymes/pathways [24], gene ontology (GO) classifications [21], PANNZER [25], and Sma3s [26] gene naming, and signal peptide annotations (with SignalP [27]). The same data were retrieved from the previously published study of the A. ceylanicum ES proteome [16]. Additionally, proteins with shared orthologs among nematode species were analyzed using BLAST [30] to identify reciprocal best BLAST hits among A. ceylanicum [16], N. americanus [15], and N. brasiliensis [31] (i.e., two proteins from different species were each other’s best match across all proteins). All findings including the functional annotations, BLAST statistics, and gene expression data for all proteins are provided for N. americanus in S1 Table and for A. ceylanicum in S2 Table.

3.2 Functional characterization of L3-specific hookworm ESPs

Of the 91 proteins unique to the ES from N. americanus L3 (Fig 1A), 53 sequences contained signal peptides for secretion (58.2%, P < 10⁻¹⁵ for enrichment, binomial distribution test). Among the 16 proteins with a relative protein abundance (NSAF) higher than 0.015, sperm-coating protein (SCP), domain-containing proteins, and Kunitz type protease inhibitors accounted for three proteins each. The top five most abundant proteins in order of relative abundance were annotated as a nematode fatty acid and retinoid binding protein (FAR), a hyaluronoglucosaminidase, a CAP domain protein and two SCP domain-containing proteins (Table 1A; the full gene list is provided in S3 Table). Among these 91 N. americanus L3-specific ES products, 14 InterPro domains/families were significantly enriched (Tables 2A and S4A), including InterPro domains associated with metallopeptidases, CUB domains (most exclusively found in secreted proteins [38]), hyaluronidases, and CAP domains (which were also enriched among the adult-specific ESPs). Five KEGG pathways were significantly enriched in the N. americanus L3 ESP and four were unique compared to the adult ESP. The KEGG pathway for "peptidases and inhibitors" was significantly and highly enriched (FDR-adjusted P <0.001) in both the N. americanus L3 and adult ESP while the pathways for glycosaminoglycan degradation, betalain biosynthesis, bacterial toxins and isoquinoline alkaloid biosynthesis were uniquely enriched in the L3 ES. A total of 30 molecular function and 24 biological process GO terms were found to be significantly enriched in the ESPs unique to N. americanus L3, including many peptidase terms also enriched in the adult-specific ESPs, and L3-specific terms including catalytic activity and metal ion binding (Tables 2A and S4A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The top five most abundant ESPs from protein sets defined in Fig 1A and 1B.

Complete protein lists with additional functional annotations are provided in S1 and S1 Tables for N. americanus and A. ceylanicum respectively. NSAF = Normalized spectral abundance factor, adjusted for protein length and total number of spectra per sample.

https://doi.org/10.1371/journal.pntd.0012780.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Significantly enriched InterPro domains, KEGG pathways and Gene Ontology molecular function and biological process terms among L3-specific ESPs of both N. americanus and A. ceylanicum.

https://doi.org/10.1371/journal.pntd.0012780.t002

In A. ceylanicum, of the 41 ESPs unique to the A. ceylanicum L3 secretome (Fig 1B), 31 contained signal peptide sequences (75.6%, P < 10⁻¹⁵ for enrichment). Among the 12 proteins with an NSAF higher than 0.015, three contained Netrin/Tissue inhibitor of metalloproteinases (TIMP) domains while two were SCP-like proteins. Like N. americanus, the most abundant protein identified in the A. ceylanicum L3 secretome was also a nematode FAR (Tables 1 and S3B), followed by a papilin, two SCP-like proteins, and a serine protease inhibitor-like superfamily member. Five InterPro domains and families were significantly enriched in the A. ceylanicum L3 ESP (Tables 2B and S4B), but only two were unique to the L3 ESP compared to the adult ESP (two LysM associated terms). The Venom allergen 5-like family (part of the SCP/TPX-1/Ag5/PR-1/Sc7, SCP/TAPS family), Protease inhibitor I35 (TIMP) as well as Netrin InterPro terms were enriched in both A. ceylanicum L3 and adult ESP. A total of 14 molecular function were significantly enriched in the L3-specific ESPs, five of which were L3-specific (three oxidoreductase activity terms, copper ion binding and monooxygenase activity), and nine terms were also enriched among adult ESPs. These terms shared with adult ESPs primarily consisted of peptidase activity terms, similar to N. americanus.

3.3 Functional characterization of adult-specific hookworm ESPs

Of the 94 adult-specific ESPs in N. americanus (Fig 1A), 54 contained signal peptide sequences (57.4%, P < 10⁻¹⁵ for enrichment, binomial distribution test), representing a similar proportion as the L3-specific ESPs. In this adult secretome, 20 proteins had a NSAF higher than 0.015, 11 of which were annotated as SCP-like proteins or SCP domain-containing proteins (S3C Table). The top five most abundant proteins in order of abundance were a CAP superfamily protein, a Shtk domain protein, a SCP-like protein, a SCP domain-containing protein, and an Ancylostoma-associated secreted protein-related protein (Table 1C).

In A. ceylanicum, 277 out of 496 adult-specific ESPs contained signal peptide sequences (55.8%, P < 10⁻¹⁵), representing a smaller proportion compared to the L3-specific ESPs, although the total number of proteins was much higher in adults. Overall NSAF values were lower due to the higher number of proteins, but nine proteins had a NSAF higher than 0.01, with four of these annotated as CAP domain proteins, three as Ancylostoma-specific proteins, two annotated as SCP domain-containing proteins, and another two as Ancylostoma-associated secreted proteins (ASPs; S3D Table). Of note, some overlap was present between annotations, and among the nine most abundant proteins, two SCP domain-containing proteins, and one Ancylostoma-associated secreted protein were also annotated as CAP domain proteins. Lastly, the two most abundant identified proteins were Ancylostoma-specific proteins, with no other known functional annotations assigned (Table 1D).

Enriched InterPro domains among the adult-specific ESPs were similar for the two species (Tables 3A, 3B, S4C and S4D, and included very strong enrichment for CAP domains (P < 10⁻¹⁵), peptidases and inhibitors (P ≤ 4.1×10⁻⁵) and peptidase activity (P ≤ 2.7×10⁻⁹).

3.4 Functional characterization of ESPs present in both L3 and adult stages

The 12 proteins detected in both the L3 and adult ESPs in N. americanus (Fig 1A) are provided in S3E Table, and the 16 proteins detected in both the L3 and adult ESPs in A. ceylanicum (Fig 1B) are provided in S3F Table. The ESPs in both species include transthyretin-like proteins, FARs, galectins and serine proteinase inhibitors, while N. americanus ESPs also contain a DVA polyprotein and an immunogenic protein 3, while A. ceylanicum ESPs contain an enolase and a ferritin. The top five most abundant ESPs in both stages are shown in Table 1E for N. americanus and Table 1F for A. ceylanicum.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Significantly enriched InterPro domains, KEGG pathways and Gene Ontology molecular function and biological process terms among adult-specific ESPs of both N. americanus and A. ceylanicum.

https://doi.org/10.1371/journal.pntd.0012780.t003

3.5 ESPs conserved across hookworm species

Shared orthologs were identified using best reciprocal BLAST hits between each pairwise combination of A. ceylanicum [16], N. americanus [15], and N. brasiliensis [31] proteins. "Reciprocal best BLAST hits" were identified when two proteins from different species were each other’s best match across the proteins.

Fifteen orthologous proteins were detected only in the L3 secretomes of N. americanus and A. ceylanicum but not in the adult stages of these two species (Fig 2A, Tables 4A and S3G). Ten (66.6%) of these proteins contained a signal peptide for secretion, two sequences were annotated as tyrosinases, and another two proteins remained functionally unannotated. The top five most abundant proteins in order of abundance were annotated as papilin, TIMP, zinc metalloproteinase NAS-28, glutathione peroxidase, and a ground-like domain-containing protein. Sixteen orthologous protein pairs were detected only in the adult secretomes of N. americanus and A. ceylanicum (Fig 2A, Tables 4B and S3I). Thirteen (81.2%) sequences contained a signal peptide for secretion and two were inferred as neprilysin metalloproteases. Four proteins were inferred as CAP domain proteins and two of these CAP domain proteins were also inferred as SCP-like proteins. The remaining two were inferred as a CRISP1/2/3 protein. The top five most abundant orthologs among N. americanus and A. ceylanicum adult secretomes were a LYS-8 protein, FAR-1, a SCP-like protein, a transthyretin-like family protein and a glutamate dehydrogenase. Lastly, only five sequences were identified as orthologs detected in the secretomes of both life stages and hookworm species (Fig 2A, Tables 4C and S3H), and four (80%) sequences contained signal peptides. These proteins were annotated as a transthyretin-related-1 protein, a clade V-conserved protein, a transthyretin-like family protein, galectin, and a FAR.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. ESP counts in cross-species comparisons, matched by 1:1 reciprocal best BLAST orthologs.

(A) Overlap of L3 and adult ESPs from N. americanus and A. ceylanicum from the current study. (B) Overlap of L3 ESPs from N. americanus, A. ceylanicum, and N. brasiliensis (from Sotillo et al. 2014). (C) Overlap of adult N. americanus ESPs (current study and Logan et al. 2020) and A. ceylanicum (current study and Uzoechi et al. 2023).

https://doi.org/10.1371/journal.pntd.0012780.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Conserved stage-specific ESPs of N. americanus and A. ceylanicum.

https://doi.org/10.1371/journal.pntd.0012780.t004

Identified L3 secretome orthologs from N. americanus and A. ceylanicum were compared to the N. brasiliensis L3 secretome [31] to further identify three-species reciprocal orthologs. Of 29 reciprocal hits identified from the N. americanus and A. ceylanicum L3 secretomes (Fig 2A), eight were detected in reciprocal orthologs from the N. brasiliensis ES (Fig 2B). Some notable proteins include brorin, extracellular superoxide dismustase as well as NAS-28 and D-aspartate oxidase 3 (DDO-3), both of which were L3-specific reciprocal hits not detected in the N. americanus and A. ceylanicum adult secretomes (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. L3 stage-specific ESP orthologs also detected in N. brasiliensis.

https://doi.org/10.1371/journal.pntd.0012780.t005

3.6 Comparison to Schistosoma cercariae ESPs, also secreted during skin invasion

The cercaria stage of the flatworm Schistosoma also enters the host by invasion through the skin during the cercarial stage [39], so comparisons to ESP proteomes from this stage of Schistosoma may give insights into broader helminth-conserved skin invasion mechanisms. There are two such available proteomic studies of S. mansoni cercariae ESPs [40,41], but these predate the availability of the S. mansoni genome annotation, and many of the Uniprot IDs identified are long deprecated, so systematic 1:1 comparisons are difficult, but manual searching of the tables from the studies was performed using BLAST orthologs to S. mansoni proteins (PRJEA36577). No S. mansoni orthologs were identified among the five proteins detected in ESPs from both stages and both species of hookworms, but five of the nine ESPs detected in L3 in both hookworm species and adult-stage A. ceylanicum ESPs had significant S. mansoni orthologs. These included (i) enolase and (ii) aldolase/fructose-bisphosphate aldolase, the two most abundant proteins identified from lipid-induced cercarial secretions and among the most highly released from acetabular glands in Knudsen et al. 2005 [40], and both were detected in cercarial ESPs in Curwen et al. 2006 [41]. The other three proteins include (iii) peptidyl-prolyl cis-trans isomerase (detected in lipid-induced cercarial and acetabular gland secretions [40], and in ESPs in Curwen et al. 2006 [41]), (iv) superoxide dismutase (detected in lipid-induced cercarial secretions [40]) and (v) Smp_089670/ SM21.7 (detected in lipid-induced cercarial and acetabular gland secretions [40]). Finally, three orthologs of S. mansoni proteins were among the 15 proteins detected only in L3 in both hookworm species: (i) dipeptidyl peptidase / dpf-1, one of 16 proteins detected in cercariae ESPs, (ii) D-aspartate oxidase, and (iii) papilin/spondin, neither of which were detected in the previous studies. Both studies identified elastases as major abundant components of S. mansoni cercariae ESPs [40, 41], but no orthologs of these or any proteins annotated as elastases were identified in either hookworm species. Together, these results indicate several sets of proteins that appear to be very important for skin penetration, including helminth-conserved proteins (especially enolase and adolase), nematode-specific proteins (transthyretin proteins, FAR-1, galectins, and others) and enolases as very abundant trematode-specific proteins. BLAST results between A. ceylanicum and S. mansoni proteins used for this analysis are provided in S2 Table.

Metalloproteinases

Two biological process GO terms associated with the molting cycle were uniquely and highly enriched in the N. americanus L3 ESPs (Table 2A; "molting cycle, collagen and cuticulin-based cuticle", P = 2.0×10⁻⁶ and "molting cycle", P = 2.0×10⁻⁶), with six of 103 proteins assigned these terms. These six proteins were also linked to GO terms associated with zinc ion binding, proteolysis, and metallopeptidase activity. Inferred as either various types of metalloendopeptidases or metalloproteinases such as MTP-1, HCH-1, or NAS-28, all six proteins were also shown by InterProScan analysis to belong to the peptidase family M12; four belonged to the subgroup of astacin-like metallopeptidases, while all but HCH-1 would contain the CUB domain structural motif. Although it did not have an NSAF higher than 0.015, a NAS-28 astacin-like metallopeptidase associated with the molting cycle GO term was identified with the 16^th highest relative abundance among A. ceylanicum L3-specific proteins. Further, this NAS-28 was found to be ortholog between the L3 secretomes from both hookworms in this study and the previously published N. brasiliensis L3 secretome [31].

Astacin-like metalloproteases containing a CUB domain were shown to be crucial in cleaving cuticle collagen for successful molting in the free-living nematode Caenorhabditis elegans [46]. Disrupting the expression of such proteins resulted in body morphology or molting defects, as these proteins appear to be required for cuticle formation and shedding [46–48]. Additionally, rescue with the respective protein orthologs from the parasitic nematodes Haemonchus contortus and Brugia malayi was able to correct the defective phenotype, suggesting that the purpose of these proteins may be conserved across different free-living and parasitic nematode species [49, 50]. The collagenolytic activity of metalloproteases was also demonstrated to be important for tissue invasion as an astacin-like metalloprotease, Ac-MTP-1 from the dog hookworm Ancylostoma caninum was shown to degrade connective tissue substrates gelatin, fibronectin and laminin, and exposure to metalloprotease inhibitors significantly reduced the number of A. caninum L3 that penetrated skin in vitro [51]. Furthermore, vaccination of dogs with recombinant Ac-MTP-1 followed by infection challenge with A. caninum resulted in worm burdens and egg counts that were inversely correlated with anti-Ac-MTP-1 IgG2 titers [52].

Oxidoreductases and Glycoside hydrolases

Several other functional processes were enriched across the L3 secretome profiles. Notably, the molecular function GO term "oxidoreductase activity" was uniquely enriched in the L3 secretomes from both N. americanus and A. ceylanicum (S4A and S2B Tables; P = 0.010 and P = 4.5×10⁻³, respectively). Abundant proteins associated with this enrichment include D-aspartate oxidase (DDO-3), which catalyzes the transfer of electrons from reductant to oxidant molecules, and can facilitate different biological processes in nematodes or cestodes such as protecting against oxidative damage, facilitating the metabolism of D-amino acids, and neurotransmission [53–56]. DDO-3 was identified among the orthologous proteins of both the N. americanus and A. ceylanicum L3 secretomes, and N. brasiliensis L3 secretome, suggesting that this protein could be crucial to hookworm invasion and survival [31].

Proteins annotated with the glycoside hydrolase (GH) superfamily InterPro domain were uniquely enriched in the N. americanus L3 secretome (Table 2A; P = 3.0×10⁻⁴). GH family enzymes act on glycosidic bonds, and we identified a protein annotated as family 31 glycosyl hydrolase (GH31). GH31 family members are highly expressed in Ascaris suum ES products, where they may degrade and metabolize host carbohydrates [57]. Indeed, several proteins identified in the N. americanus L3 secretome that belonged to this family were also functionally mapped to the GO term of carbohydrate metabolic processes. For example, α-L-fucosidase is critical for biological processes such as embryo development, mobility and longevity in the pine wood nematode Bursaphelenchus xylophilus [58], while malate dehydrogenase participates in the citric acid cycle [59].

Hyaluronidases are a subset of the glycoside hydrolase superfamily as they cleave the β1–4 glycosidic linkages in hyaluronic acid [59,60]. Hyaluronidases facilitate skin invasion and enable hookworm larvae passage between keratinocytes through hydrolysis of extracellular matrix hyaluronic acid [61, 62]. We found that the hyaluronidase functional term was notably enriched in the N. americanus L3 secretome (Table 2A; P = 7.2×10⁻³). Intriguingly larval homogenates from the canine hookworm A. braziliense exhibit far greater hyaluronidase activity than Ancylostoma tubaeforme and A. caninum [61]. This disparity might be attributed to the route of infection as A. tubaeforme and A. caninum can infect via both the skin and oral routes whereas A. braziliense is an obligate skin-penetrator [61]. We have not detected proteins with hyaluronidase activity from the A. ceylanicum L3 secretome, and given that A. ceylanicum is infective via oral or skin routes, it might secrete less hyaluronidase than N. americanus [4].

Cysteine-rich secretory proteins

Numerous sequences inferred as cysteine-rich secretory proteins (CRISP), Ancylostoma-secreted proteins (ASP), SCP-like or SCP-domain containing proteins belonging to the CAP superfamily were identified across both hookworm species and life stages in this study. The CAP superfamily is also known as the SCP protein family and it encompasses a diverse group of proteins defined by the presence of cysteine-rich CAP motifs which make up structurally conserved disulfide stabilized CAP (sometimes referred to as SCP or PR-1) domains of 17–21 kDa in size [63]. Proteins of this family are often referred to as SCP/TPX-1/Ag5/PR-1/Sc7 (SCP/TAPS) or ASP proteins [15,16,64]. SCP/TAPS proteins are highly expressed in the ES products of the larval and adult developmental stages of numerous helminths [15, 16, 31, 45, 63, 65–68]. Functional terms associated with the venom allergen 5-like protein (VAL) family were also enriched in the N. americanus adult secretome and the secretome from both life stages of A. ceylanicum. A subset of SCP/TAPS, VALs appear to preferentially bind hydrophobic molecules such as sterols or fatty acids [69, 70].

Both VALs and the broader family of SCP/TAPS proteins in helminths associate with activities such as immune modulation [71, 72], host invasion, and transition to parasitism [73–75]. Some ASPs or VALs confer vaccine efficacy [76, 77], although vaccination with Na-ASP-1 induced an IgE mediated urticaria in endemic region study volunteers with pre-existing hookworm protein-specific IgE [72, 78]. Despite the significance and enrichment of this protein family, little is known about their molecular mechanisms of action, and research is needed to elucidate its functional contribution to hookworm biology [64].

In this study, we detected 124 of these cysteine-rich secretory proteins in A. ceylanicum adult, four in A. ceylanicum L3, 43 in N. americanus adult and 11 in N. americanus L3. All of the cysteine-rich secretory proteins identified in each species are identified in S1 and S2 Tables.

Additional ESPs of interest

The potential development of vaccines based the identification of orthologous proteins from both larval and adult stages presents an opportunity for pan-hookworm protection. The orthologous proteins from the N. americanus and A. ceylanicum L3-specific secretome (Table 4A) provide a larger pool of novel immunization candidates not detected in the adult secretome, that could target the larvae during its parenteral migration. Indeed, a recent clinical trial with repeated rounds of short-term exposure with N. americanus L3 followed by anthelmintic treatment induced a protective response that was associated with skin responses, indicating that antibody-mediated killing of L3 in the skin generates meaningful immunity [79]. Furthermore, the three-species L3 orthologs offer a group of proteins that have been down-selected from the profiled L3 secretome that can readily be evaluated in the N. brasiliensis mouse challenge model due to the presence of orthologs in the N. brasiliensis secretome. Aside from the previously discussed DDO-3 and NAS-28, a tissue inhibitor of metalloproteases (TIMP/Netrin) and a tyrosinase copper-binding oxidoreductase containing a Shtk domain were identified among the three-species orthologs. Furthermore, TIMP/Netrins were highly represented in the proteins unique to the A. ceylanicum secretome and another tyrosinase/Shtk domain protein was identified among ESP orthologs from all three species.

As reflected in its name, TIMPs inhibit metalloproteases. Their N-terminal domains share structure with netrin domains, which results in overlap between the TIMP and Netrin functional annotations in this study [80, 81]. However, sequence similarity may not translate to functional activity. AceES-2, previously identified from A. ceylanicum was shown to be structurally similar to other netrin domain-containing TIMPs yet does not inhibit matrix metalloproteases [82]. Due to their structural properties, the immunomodulatory activity of TIMPs may be due to cytokine binding, and therefore unrelated to the inhibition of metalloprotease catalysis [82–85]. Whereas the precise functions of tyrosinases and/or Shtk domain proteins in nematode larvae have yet to be elucidated, tyrosinases of adult stages, (a type of oxidoreductase) assist in the cross-linking of proteins during eggshell production [86, 87]. Congruent with our gene ontology analysis of the A. ceylanicum L3 secretome and uniquely enriched therein, some tyrosinases participate in the melanin synthesis. Teladorsagia circumcincta appears to upregulate melanin production in response to direct sunshine and melanization may protect free-living larvae from environmental UV and genotoxicity [88]. More research is needed to determine the functional implications of tyrosinases and Shtk domain proteins, as one of the most abundant proteins in the adult N. americanus secretome was a Shtk protein without any tyrosinase annotation, and sequences associated with both functional terms featured prominently within a few different protein compartments in this study.

Transthyretin-like or transthyretin-related protein orthologs were detected from both development stages of N. americanus and A. ceylanicum, and the adult-specific secretome also included proteins belonging to this family (Table 4B and 4C). Notably, transthyretins were not detected uniquely in the L3 secretome. In vertebrates, transthyretins function as a transporter of thyroid hormones but transthyretin-like proteins in nematodes, despite containing a similar transthyretin-like domain, do not appear to demonstrate thyroid binding properties [89, 90]. Transthyretin-like proteins and transthyretin-related proteins of nematodes have been implicated in biological processes such as larval development, protection against oxidative stress and the clearance of apoptotic cells [91–93]. H. contortus transthyretin also showed promise as a vaccine candidate, inducing a marked antigen-specific antibody response and reduction in parasite load in goats [94].

A FAR ortholog from N. americanus and A. ceylanicum was detected in both the L3 and adult stage ESPs (Table 4C), and a separate FAR (FAR-1) was identified among the more abundant adult-specific secretome proteins in both species (Table 4B). In addition, we identified highly abundant L3-specific FAR proteins that were unique to each species (Table 1A and 1B). FARs bind lipids and retinol and are crucial for scavenging fatty acids and other lipid-based molecules from the host as nematode parasites have a limited capacity for lipid synthesis [95–97]. FARs have been described to be immunogenic [98,99] or immunomodulatory, and it was postulated that the latter property may be was facilitated by the ability of FAR to bind lipid-based signaling molecules, which interferes with their associated immune pathways [71,97].

Several of the orthologous secretome proteins identified in this study have minimal to no functional annotation assigned by any of the annotation tools. Given their interspecies overlap, these proteins could be implicated in crucial hookworm processes and further investigation into these proteins could yield novel insights into hookworm biology. We note, however, that due to the draft nature of these nematode genomes, some proteins and orthologous matches may have been missed from the proteomic identification altogether. Nonetheless, based on BUSCO [100] genome completeness analysis in earlier reports, we expect to have annotated 97.4% of all N. americanus proteins [15] and 94.5% of all A. ceylanicum proteins [16]. Hence, we do not expect a substantial number of missing annotations. It is also important to note that data generated by this study may be limited by the lack of biological replicates as it was difficult to culture a large number of L3 for ESP production due to the low number of L3 the participants were originally infected with and sample loss through processing. Further, only one participant out of six developed a patent A. ceylanicum infection that was sufficient to generate an adequate amount of L3 required for ESP production. A larger study in the future with more participants could overcome this limitation and further enrich this dataset through the production and analysis of a larger quantity of ESP. Moreover, biological replicates for adult hookworm ES products were also omitted due to the difficulty in obtaining sufficient number of adult worms in hamsters. The percentage of larvae that mature to adulthood in hamsters is low, and we had ethical concerns around the number of hamsters required to generate larger quantities of ES products.

Finally, comparisons to ESPs secreted from the cercarial stage of the flatworm S. mansoni provided a broader phylogenetic perspective of the data, since schistosomes also enter the host by invasion through the skin by the cercariae stage [39]. This analysis allowed for the phylogenetic categorization of several sets of proteins that appear to be very important for skin penetration, including helminth-conserved skin penetration ESPs (especially enolase and aldolase), nematode-specific proteins (transthyretin proteins, FAR-1, galectins, and others as described above) and enolases as very abundant trematode-specific proteins.

Disclaimer

The views, opinions and/or findings expressed are those of the authors and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government.