Strains and culture conditions

The A. oligospora strains used in this study are listed in Table 1. A. oligospora strains were maintained on potato dextrose agar (PDA; Difco). Trap induction, capture rate, and survival rate assays were all performed on low-nutrient medium (LNM: 2% agar, 1.66 mM MgSO₄, 5.4 mM ZnSO₄, 2.6 mM MnSO₄, 18.5 mM FeCl₃, 13.4 mM KCl, 0.34 mM biotin, and 0.75 mM thiamin). C. elegans N2 strain was maintained on NGM with E. coli (OP50) as the food source.

Download:

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

Table 1. Arthrobotrys oligospora strains used in this study.

https://doi.org/10.1371/journal.pbio.3002400.t001

RNA-sequencing analyses

A. oligospora TWF154, grown on LNM, was transferred to fresh 9 cm LNM plates. After approximately 5 days (25°C, dark), the mycelium was exposed to approximately 2,500 to 3,000 C. elegans N2 L4-stage larvae per plate, and hyphal collections were taken at 2, 4, 10, 24, and 48 h after exposure, respectively. For the 2 and 4 h time points, nematodes were washed out from the Petri dishes prior to hyphal collection. The hyphae were harvested using a cell scraper and immediately frozen in liquid nitrogen. Total RNA was isolated by the Trizol-Phenol-Chloroform method, and cDNA libraries were prepared using the Illumina TruSeq Stranded mRNA kit. Prepared cDNA was sequenced on an Illumina NextSeq500 for 76 cycles (high-output), and 2 × 75 bp paired-end reads were obtained. Each biological replicate consisted of 45 to 76 M reads (approximately 85 to 144× coverage). Raw reads were filtered with fastp 0.14.1 and a minimum length of 68 was imposed, resulting in a minimum mean quality for 96% of Q30 read bases and 100% of >Q34 read bases. Salmon 0.9.1 was used to quantify transcript expression. Bootstrapped abundance estimates (200 bootstrap samples) were computed [36] and pseudoaligned using Kallisto [37]. Sleuth 0.29.0 [38] was employed to determine differential expression using the likelihood ratio test across all conditions, resulting in p-values and q-values (adjusted p-values) used in all subsequent analysis. In addition, a Wald test was performed to calculate the coefficient of the full model for the studied condition, denoted by beta. beta coefficients were computed to be comparable (but not equal) to the log₂ fold change between samples exposed to nematodes and those harvested before the addition of nematodes [38]. The resulting dataset is available from the NCBI GEO database (http://www.ncbi.nlm.nih.gov/geo/; series record: GSE233568). A q-value of <0.01 and a beta >1 or <−1 were used to define transcripts with altered expression. Sleuth [38] and Jvenn [39] were used to perform principal component analysis and build Venn diagrams, respectively. Transcript clusters were obtained for the differentially expressed genes using clust [40]. Selected lists of differentially expressed transcripts were run through Omicsbox 2.0.36 [41] for functional enrichment analysis, with p-value < 0.01 (Fisher’s exact test) as the threshold and the “Remove double IDs” option. The initial results obtained for both “Interpro IDs” and “Interpro GO IDs” were consolidated using the “reduce to most specific” option (p-value < 0.01). Heat maps were built on GraphPad Prism.

Secretome of A. oligospora

The secretome of A. oligospora was predicted using a combination of tools [42]. First, the entire transcriptome was run on SignalP 5.0 (cutoff = 0.8) [43], Phobius [44], and TMHMM 2.0 [45] to predict the presence of N-terminal signal peptides and transmembrane regions. Transcripts with a signal peptide (as defined by the concordant detection by SignalP and Phobius) and no transmembrane regions (according to both Phobius and TMHMM) were recorded as “list A.” To enhance our confidence in the final list of predicted secreted transcripts, which may represent proteins bearing either noncanonical secretion domains or signal peptide domains mistakenly categorized as N-terminal transmembrane domains by previous programs, we employed the following methodology: Transcripts lacking a signal peptide, as determined by the absence of SignalP and/or Phobius outputs, and featuring precisely one transmembrane region exceeding 18 amino acids within the initial 60 amino acids of the encoded protein were categorized as “list B.” Transcripts that exhibited a signal peptide, consistently detected by both SignalP and Phobius, and harbored one or more transmembrane domains (defined by Phobius and TMHMM, and not present in “list B”) were designated as “list C.” Subsequently, “list B” and “list C” were analyzed using WolfPSort [46], with transcripts achieving an “Extracellular” score of ≥17 recorded as “list B1” and “list C1.” Finally, list A, list B1, and list C1 were separately run on MitoFates [47]. Transcripts predicted to be targeted to mitochondria (cutoff = 0.5) were excluded, yielding the final compilation of predicted secreted transcripts. This comprehensive approach identified 913 secreted transcripts from list A (84.2%), 17 from list B1 (1.6%), and 154 from list C1 (14.2%). The final list (1,084 secreted transcripts) was run on EffectorP-fungi 3.0 [48] to detect putative effectors, yielding 417 putative effectors.

DUF3129 gene detection and orthologs distribution

The DUF3129 domain is defined as the conserved domain with accession numbers IPR021476 (InterPro) and PF11327 (Pfam). Multiple amino acid sequence alignments were executed on MAFFT [49] and subsequently visualized using Jalview [50]. Genes in the family of DUF3129 were characterized as those containing the DUF3129 domain of TEP1. Therefore, DUF3129 genes were found using the domain as query for a BlastP against filtered model proteins of each genome. On the mycocosm database, we used the blast section of the mycocosm portal [51] with options: aligment program blastp protein versus protein, database Gene Catalog proteins, Expect 1.0E-5, Word Size 3, Filter low complexity regions, Perform gapped alignment, and Scoring Matrix BLOSUM62. Local alignments were produced using blastp option of blast 2.9.0 [52], with options -outfmt 6 -max_target_seqs 500 -evalue 1e-5 -num_threads 16, on databases produced using makeblastdb -dbtype prot for each genome. Orthology groups for the 19 selected high-quality genomes (S5 Table) were computed using funannotate_compare within funannotate 1.8 [53], relying on ProteinOrtho v5 [54] with options -synteny -cpus = 24 -singles -selfblast. Orthogroups were selected among all the computed orthogroups as those containing at least one of the DUF3129 genes obtained from the blastp local alignment. To study gene gain and loss of the DUF3129 family, we inferred gene duplication and loss events from gene trees constructed with all DUF3129 genes in 19 high-quality fungal genomes and the species tree depicted in S3 Fig. Gene trees were constructed using the sequence alignment tool Muscle5 [55] and the software for phylogenetic tree inference, IQ-TREE 2 [56]. The species tree was built based on concatenation of single-copy orthologs from the 19 fungal genomes using Muscle5 and IQ-TREE 2. Further refinement of the gain/loss pattern was computed through SpeciesRax [57] with default settings. DUF3129 genes, obtained through blastp, that were not clustered in orthogroups of at least 4 genes were not suitable to compute gene trees; consequently, for such genes, we assessed shared lineage manually. Finally, genes without a shared lineage were counted as species-specific gained genes. The final results were summarized and visualized using TreeViewer [58]. Chromosomal maps of the DUF3129 genes were drawn using MapChart [59].

Targeted gene deletions and complementation

Targeted gene deletions were generated from either wild-type or the ku70 mutant of A. oligospora TWF154, as described previously [18]. Briefly, an overlap PCR-based construct was obtained by fusing the sequences 1.5 kb upstream and downstream of the target genes to a hygromycin resistance cassette (amplified from vector pAN7-1) or a clonNAT resistance cassette (amplified from vector pRS41N). The construct was then introduced into the protoplast of either A. oligospora TWF154 or the ku70 background strain via PEG-mediated transformation. Protoplasts of A. oligospora were generated according to previously described methods [18]. Approximately 5 × 10⁶ conidia were cultured in 100 mL of PDB for 24 h at 25°C and 200 rpm. Fungal hyphae were collected by centrifugation and washed by MN buffer (0.3 M MgSO₄, 0.3 M NaCl). The mycelium was then mixed with 5 mL of 100 mg/mL VinoTastePro enzyme mix in MN buffer overnight at 25°C and 200 rpm. Protoplasts were filtered through 3 layers of miracloth and washed with STC buffer (1.2 M sorbitol, 50 mM CaCl₂, 10 mM Tris–HCl (pH 7.5)). Next, 5 × 10⁵ protoplasts were mixed with 5 μg of construct DNA and incubated on ice for 30 min. Five times volume of PTC buffer (40% PEG 4000 [w/v], 10 mM Tris–HCl (pH 7.5), 50 mM CaCl₂) was then added and allowed to incubate for 20 min at room temperature. Then, molten regeneration agar (1% agar, 0.5 M sucrose, 0.3% yeast extract [w/v], 0.3% casein acid hydrolysate [w/v]) containing 100 μg/mL hygromycin B or 250 μg/mL clonNAT was added to the protoplasts and poured into Petri dishes. Transformants were then screened for gene deletion. Primers used for gene deletion and genotyping are listed in S1 Table.

For gene complementation, a wild-type copy of the target gene along with flanking sequence (2 kb upstream and 1 kb downstream) was amplified, fused with a G418 marker, and transformed into mutant protoplasts. The complemented strains were then genotyped to confirm gene rescue.

Phenotypic characterization

Induction of A. oligospora trap morphogenesis by C. elegans was performed as previously described [8]. Briefly, fungal strains were grown on 3-cm LNM plates for 48 h before exposure to 30 C. elegans L4 larvae. After 6 h, nematodes were removed, and 24 h later, random images within 0.5 cm of the plate edge were captured using a Zeiss Stemi 305 microscope with a Zeiss Axiocam ERc 5s camera at 40X magnification. Trap number was quantified from the images using Fungal Feature Tracker [60].

For rapamycin treatment, the final concentration of LNM with 4 μg/mL rapamycin was prepared by directly adding 10 mg/mL rapamycin stock to LNM. To quantify the C. elegans survival rate, WT A. oligospora was inoculated onto 3-cm LNM with and without rapamycin and cultured for 3 days, 25°C. After addition of 30 L4 C. elegans, the number of free-living worms was counted every 3 h for a total of 12 h. To quantify nuclei number between hyphal and trap cells, trap induction +/− rapamycin was performed as above using the H2B::mCherry strain. After 24 h, fungal cultures were stained with SCRI Renaissance 2200 (SR2200), which stains fungal cell walls without arresting growth, and imaged using an LSM980 Airyscan 2 confocal microscope (Carl Zeiss). Nuclei number was manually counted within each fungal cell.

To investigate the role of DNA replication in trap morphogenesis, fungal strains grown on 3-cm LNM were incubated with the DNA replication inhibitor, hydroxyurea (HU), at a final concentration of 200 mM for 30 min before exposure to C. elegans.

To estimate the C. elegans escape rate from A. oligospora traps, approximately 200 L4 to young adult worms were introduced per fungal plate to induce trap formation. After 24 h, another 50 young adult worms were added to the traps for 10 min before shaking slightly and rinsing with 1 mL of deionized water. The number of uncaptured nematodes was manually counted, and the nematode escape rate was calculated as (uncaptured nematodes/50) * 100%.

To characterize the growth of fungal mycelia inside nematodes after penetration, fungal strains were transformed with a cytosolic GFP plasmid. A chunk of agar with mycelia from the mother culture was inoculated onto a 5-cm LNM plate and allowed to grow for 3 days. Following 24 h of trap induction by addition of worms, additional adult worms were inoculated onto the formed traps and imaged using the Andor Revolution WD system with a Nikon Ti-E automatic microscope and an XON Ultra 888 EMCCD camera (Andor).

Third-generation HCR in situ hybridization

In situ hybridization was performed with slight modifications to the previously described third-generation in situ HCR method [61]. First, wild-type A. oligospora were grown on a 24-well LNM plate (each well approximately 1.9 cm²) for 48 h. Approximately 100 young adult C. elegans were then added onto the fungal hyphae. After 18 h of contact, the samples were fixed with 1 mL of 1% formaldehyde in PBS for 10 min at room temperature and 50 rpm. Immediately after fixation, lysing enzyme (Sigma-Aldrich) was added to a final concentration of 100 mg/mL and incubated at room temperature for 2 h. After digestion, the lysed hyphae were permeabilized with 70% ethanol overnight at 4°C and 50 rpm, followed by Triton treatment (0.1% in PBS) for 5 min at room temperature and 50 rpm. The plate was then prewarmed to 37°C for the following steps. During detection, 200 μL of probe hybridization buffer was added to the sample for 10 min. The sample was then hybridized with a probe solution (0.4 pmol in 100 μL of probe hybridization buffer) overnight in a humidified dark chamber at 37°C. Amplification was performed overnight to generate long HCR amplification polymers, which were then imaged using an LSM980 Airyscan 2 confocal microscope (Carl Zeiss) with a C-Apochromat 63×/1.2 numerical aperture (NA) water immersion objective. Images were acquired with Multiplex Mode (SR-4Y).

TEP1-GFP fusion protein construction

To construct the TEP1-GFP fusion protein using In-Fusion cloning, TEP1 and 2 kb upstream were amplified using the primers 3105_LGFP_F1 (5′-AGGGAACAAAAGCTGGTACCTACATGCAGAGTCTGTGGTTAC-3′) and 3105_LGFP_R1 (5′-ACCGCCACCGGCGCGCCTGTTACCCTTTTTCTCAGCTGGC-3′). PCR amplification was performed with Phusion High-Fidelity DNA Polymerases (NEB). The amplicon was then cloned into a GFP plasmid with a geneticin resistance cassette using the In-Fusion HD PCR Cloning Kit (Clontech).

Intrinsically disordered regions (IDRs) and three-dimensional structure prediction

To identify IDRs in TEPs, we utilized 2 commonly used disorder predictors, that is, PONDR VSL2 [62] and PONDR VL3 [63]. These predictors can be accessed at the PONDR website (http://www.pondr.com). We predicted protein structure using AlphaFold2 [64].

Protease inhibitor assay

To assess the inhibitory effects of protease inhibitor cocktail (PIC), A. oligospora mycelia were incubated in Potato Dextrose Broth (PDB) for 3 days. Following incubation, the mycelia were disrupted using sonication at the setting duty-cycle 50%, output control 2 (Microtip, Branson, Sonifier 250). The resulting mixture was then subjected to centrifugation at 16,000×g to remove the pellet. For protease inhibition, the supernatant was incubated with 1× PIC (P2714, Sigma-Aldrich) for 2 h (diluted 1:10 from the stock solution). Subsequently, 10 μL of the obtained supernatant were applied as spots onto a milk plate composed of 2% milk powder and 2% agar. The plate was incubated at 25°C for 24 h, after which imaging was performed.

To assess the potential roles of proteases in nematode colonization, A. oligospora culture was treated with 1× PIC for 2 h. Adult worms were then placed onto the traps and imaged using an Andor Revolution WD system with a Nikon Ti-E automatic microscope and an XON Ultra 888 EMCCD camera (Andor). Deionized water was used as the control.

Time-course transcriptional profiling of A. oligospora in response to C. elegans reveals a highly dynamic transcriptome landscape during nematode predation

To capture the different stages of nematode predation (that is, sensing, trap formation, adhesion, penetration, and digestion [3]), we profiled the A. oligospora transcriptome under low nutrient conditions at various intervals following exposure to nematode prey (Fig 1A). For sequencing and all subsequent experiments, we used our newly established A. oligospora strain (TWF154) [8]. Principal component analysis of all transcript reads showed clustering of biological replicates for all conditions sampled (pre-exposure control, 2, 4, 10, 24, and 48 hours post-exposure (hpe) samples), certifying low variability between replicates (Fig 1B). Notably, overall transcript expression at 2 hpe and 4 hpe (earlier stages of interaction) showed tighter clustering to each other and to the pre-exposure control than to later time points (10, 24, and 48 hpe time points) (Fig 1B).

Download:

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

Fig 1. A. oligospora responds to the presence of prey with a dynamic transcriptomic response.

(A) Overview of time-course RNA-seq experiment. hpe, hours post-exposure. (B) Principal component analysis of A. oligospora transcript expression upon exposure to C. elegans of 3 biological replicates for the indicated time points, as analyzed by RNA-seq. The data underlying this Figure can be found in S1 Data. (C) Venn diagrams depicting the number of A. oligospora transcripts that are up-regulated (left) or down-regulated (right) upon exposure to C. elegans. (D) The total number of transcripts with altered expression for each of the time points (“size of each list”) is shown. (E) Cluster analysis of transcript expression profiles revealed 5 common expression profiles. The 2,693 differentially expressed genes were subjected to soft clustering to identify 5 common expression profiles. The top enriched GO term for each cluster is listed, respectively. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002400.g001

A large number of transcripts were found to be down-regulated and up-regulated in response to C. elegans. Specifically, using thresholds of beta > 1 (up-regulation) or beta < −1 (down-regulation), 839, 959, 1,129, 678, and 622 transcripts were up-regulated, while 1,552, 1,471, 1,760, 1,141, and 996 transcripts were down-regulated at 2 hpe, 4 hpe, 10 hpe, 24 hpe, and 48 hpe, respectively (Fig 1C and 1D). Among all the time points sampled, we observed the greatest differential expression (both up- and down-regulation) at 10 hpe, which corresponds to a period of intense trap formation and adhesion between fungal and nematode cells. Across all time points, a total of 159 and 318 transcripts were up-regulated and down-regulated, respectively, and are likely to correspond to core genes required for predation (Fig 1C).

Furthermore, examination of expression levels across time for differentially expressed genes (q-value < 0.01 and beta > 1 or < −1) revealed 5 clusters of genes defined by their unique expression pattern during nematode predation (Fig 1E and S2 Table). Transcripts in the most highly occupied cluster (“cluster C0,” 994 genes) showed a quick increase in expression upon exposure to C. elegans, sustained expression until 10 hpe, and a subsequent decrease from 24 hpe to 48 hpe (Fig 1E). This cluster was enriched in transcripts related to ribosomal biogenesis, nucleosome assembly, and translation (Fig 1E). Showing a reciprocal expression pattern to C0, cluster C3 was the second most highly occupied group, encompassing 927 transcripts. This cluster was enriched in transcripts with functions in oxidation–reduction processes that harbor metabolism, stress response, and enzymatic activity (Fig 1E). Cluster C1 transcripts (310), which showed gradual increases in expression that peaked at 10 hpe, were enriched for genes containing signal peptides. Interestingly, we also observed a strong overrepresentation of genes belonging to the DUF3129 (domain of unknown function) protein family (Fig 1E), also known as Egh16-like virulence factors [65–67]. Cluster C2 (234 transcripts) expression was significantly reduced upon contact with nematodes, followed by gradual restoration of expression to pre-exposure levels by the terminally sampled time point. This cluster primarily comprises genes involved in transmembrane transport, as well as N,N-dimethylaniline monooxygenase activity (Fig 1E). Cluster C4 transcripts (228) were characterized by a sharp decrease in expression at 24 hpe. Interestingly, C4 transcripts were found to be involved in nitrogen metabolism (specifically the conversion of nitrogen-containing compounds into usable nitrogen forms) and/or to contain domains bearing nitrate assimilation and oxidoreductase activity. Altogether, these data indicate that NTF react to the presence of nematode prey with a very rich and dynamic transcriptomic response.

Up-regulation of ribosome biogenesis and nuclear division during trap morphogenesis

To gain further insights into trap morphogenesis, we conducted Gene Ontology (GO) enrichment analysis in A. oligospora samples collected between 2 and 10 hpe to C. elegans. Analysis revealed “structural constituent of ribosome” (GO:0003735), “rRNA processing” (GO:0006364), “nucleolus” (GO:0005730), and “nucleosome” (GO:0000786) as the top 4 most highly enriched GO terms, indicating a strong up-regulation of ribosome biogenesis and protein synthesis during the initial hours of exposure to prey (Fig 2A). These results suggest that, despite nutrient deprivation, fungal cells can ramp up ribosome biogenesis machinery in preparation for trap development upon sensing prey. Ribosome biosynthesis is known to be regulated by the TOR signaling pathway [68,69]. Thus, to examine the effect of ribosome biosynthesis on trap morphogenesis, we utilized rapamycin to inhibit TOR activity. Supplementation of rapamycin to LNM yielded a significant reduction in the number of traps induced by C. elegans and an increase in the C. elegans survival rate (Fig 2B and 2C), demonstrating that up-regulation of the ribosome biogenesis pathway is crucial for trap development. Furthermore, using a strain expressing histone H2B::mCherry to determine nuclei number in fungal cells, we observed dramatically more nuclei in trap cells than in vegetative hyphae (an average of 10 versus 2 nuclei) (Fig 2E). Consistent with this observation, the GO terms “nucleolus” (GO:0005730) and “nucleosome” (GO:0000786) are highly enriched in 2 to 10 hpe differentially expressed genes (S1 Fig). In the presence of rapamycin, nuclei number per trap cell was reduced to approximately 5 nuclei (Fig 2E).

Download:

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

Fig 2. Ribosome biosynthesis is up-regulated in early time point.

(A) Expression profile of structural constituents of ribosome upon exposure to C. elegans. The data underlying this Figure can be found in S1 Data. (B) Trap induction and quantification of WT A. oligospora grown on LNM and LNM with 4 μg/ml rapamycin (scale bar, 500 μm). The data underlying this Figure can be found in S1 Data. (C) Survival rate of C. elegans on WT A. oligospora grown on LNM and LNM with 4 μg/ml rapamycin. The data underlying this Figure can be found in S1 Data. (D) Trap induction and quantification of WT A. oligospora on LNM and LNM with 200 mM HU (scale bar, 500 μm). The data underlying this Figure can be found in S1 Data. (E) Close-up images and quantification of the nuclei of the H2B::mCherry strain in the traps of A. oligospora grown on LNM and LNM with 4 μg/ml rapamycin or 200 mM HU (scale bar, 20 μm). The arrowhead indicates the trap cell; the arrow indicates vegetative hyphae. The data underlying this Figure can be found in S1 Data. hpe, hours post-exposure; HU, hydroxyurea; LNM, low-nutrient medium; WT, wild-type.

https://doi.org/10.1371/journal.pbio.3002400.g002

Given the nearly 5-fold increase in DNA content in trap cells over vegetative hyphae, we next asked whether inhibition of DNA replication might affect trap morphogenesis. We applied the known DNA synthesis inhibitor HU [70,71] to fungal cells induced with nematodes and found that HU treatment moderately inhibited trap development and reduced the number of nuclei to approximately 4 nuclei per trap cell (Fig 2D and 2E). Taken together, these findings suggest that, under nutrient deprivation, A. oligospora up-regulates ribosome biogenesis, translation, and nuclear division to prepare for trap development in response to nematode signals.

A. oligospora exhibits strong reliance on secretion-related genes during interactions with C. elegans

Strikingly, in our functional analysis, we observed a strong enrichment of differentially expressed genes with functions related to protein secretion. Indeed, applying a stringent threshold of beta > 2 nevertheless resulted in highly significant (q-value = 1.7 × 10⁻¹⁰¹) enrichment of the signal peptide category “SignalP-noTM” at 10 hpe (S3 Table). This suggests that the A. oligospora response to C. elegans involves vigorous induction of protein secretion.

To better understand the role of protein secretion in the fungal prey response, we predicted the A. oligospora secretome using a combination of outputs from SignalP [43], Phobius [44], TMHMM [45], and WolfPSort [46]. A total of 1,084 peptides were predicted to be secreted in A. oligospora, representing 8.25% of the transcriptome. From this list, 549 transcripts (51% of 1,084) showed altered expression in our RNA-seq dataset (q-value < 0.01; beta > 1 or < −1) (Fig 3A). Expression pattern and functional enrichment analysis of the 549 transcripts revealed that a large proportion were up-regulated at 4 to 10 hpe (Fig 3B) and displayed an overrepresentation for a few select protein domains (S4 Table). Within the predicted secretome, the “DUF3129,” “Glycoside hydrolase superfamily,” “Peptidase S8/S53,” and “Carbohydrate-binding WSC” protein domains were highly enriched (S4 Table).

Download:

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

Fig 3. A. oligospora‘s response to C. elegans is dependent on protein secretion.

(A-D) Percentage and expression pattern analysis of the predicted secretome (A, B) and effectorome (C, D) that showed differential expression upon exposure to C. elegans. The data underlying this Figure can be found in S1 Data. (E) High-resolution image of traps stained by SR2200. Images were taken 24 h after induction (scale bar, 50 μm). The arrowhead indicates the trap cell; the arrow indicates vegetative hyphae. (F, G) Quantification of trap numbers (F) induced by exposure to 30 C. elegans nematodes and nematode escaping rate. (G) Among trapping of 50 C. elegans (mean ± SEM). The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002400.g003

Work in pathogenic fungi has demonstrated an important role for effector proteins in infection [72–74]. In this context, we therefore reasoned that effector proteins might play a role in nematode predation by manipulating nematode cellular processes or influencing behavior during trapping. To investigate this hypothesis, we used the predicted secretome to identify putative effectors in A. oligospora, resulting in a list of 417 predicted transcripts encoding effectors (representing 38.5% of the secretome) (Table 2). Two hundred and nine predicted effectors showed differential expression in the presence of C. elegans (q-value < 0.01; beta > 1 or < −1; Fig 3C), with the majority of the predicted effectors induced at 4 or 10 hpe (Fig 3D). While 259 predicted effectors lacked Pfam domains, functional enrichment analysis of the differentially expressed predicted effectors encoding known domains again revealed the prevalence of the “DUF3129” domain, as well as “peptidase M43,” “cysteine-rich secretory protein-related,” and “ribonuclease T2-like” domains.

Download:

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

Table 2. Number of secretome and effectorome in A. oligospora.

https://doi.org/10.1371/journal.pbio.3002400.t002

To further investigate the role of secretion in NTF, we investigated how disruption of protein secretion might affect A. oligospora predation by leveraging a genetic mutant deficient in the highly conserved syntaxin gene EYR41_000055. Syntaxin belongs to the t-SNARE family, is known to regulate the fusion of secretory vesicles with the plasma membrane during the late stages of the secretory pathway, and has been shown to enhance protein secretion when overexpressed in the budding yeast Saccharomyces cerevisiae [75,76]. We generated a targeted gene deletion mutant in the plasma membrane t-SNARE protein sso2 in the ku70 background, which facilitates homologous recombination in A. oligospora. While the sso2 mutant exhibited no defects in trap development (Fig 3E and 3F), we found that sso2 mutant traps exhibited reduced functionality compared to wild-type traps. Specifically, 70% of nematodes were able to escape sso2 traps, in contrast to the nearly 0% escape rate from wild-type A. oligospora traps (Fig 3G and S1 and S2 Movies). These data demonstrate that protein transport and secretion, likely of nematode-adhesive molecules, is important for trap function.

Members of the DUF3129 gene family are expanded in the genome of nematode-trapping fungi

Having identified significant enrichment of the DUF3129 domain in our predicted secretome and effectorome datasets, we further probed the role of DUF3129 domain-containing genes in A. oligospora. A specific examination of this family in our RNA-sequencing demonstrated that several members of the family were prominently induced in A. oligospora when exposed to nematodes (Fig 4A). Notably, among the 27 DUF3129 transcripts, 15 exhibited peak induction at 10 hpe to C. elegans (Fig 4A), suggesting a potential role of DUF3129 proteins in trap development. A. oligospora genome encodes 27 DUF3129 domain-containing genes, and except for one, all of these DUF3129 domain genes are predicted to contain a signal peptide by SignalP, indicating that this domain likely defines an expanded family of secreted proteins. Furthermore, approximately 40% (11 genes) of the family members were identified as putative effectors in our analysis. We further examined the distribution of the DUF3129 genes in the A. oligospora genome and found them scattered across 8 chromosomes without apparent genomic clustering (S2 Fig).

Download:

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

Fig 4. The TEP family is an important component of the A. oligospora response to C. elegans.

(A) Expression profiles of TEP transcripts upon exposure to C. elegans. The data underlying this Figure can be found in S1 Data. (B) Histogram depicting the number of genes containing the DUF3129 domain encoded in the 2,512 fungal genomes (S5 Table). The data underlying this Figure can be found in S1 Data. (C) Orthogroup analysis cloud map. Each dot represents a DUF3129 protein from the indicated species, with lines connecting proteins belonging to the same orthogroup. The data underlying this Figure can be found in S1 Data. (D) High-resolution confocal imaging of 4 highly expressed TEP transcripts in WT traps. The WT strain was induced by nematodes for 24 h. Hairpins were conjugated with Alexa Fluor 647, 594, 546, and 514 (red) (scale bar, 20 μm). The arrowhead indicates the trap cell; the arrow indicates vegetative hyphae. (E) An image of trap induction of the TEP1-GFP strain (scale bar, 200 μm). The arrowhead indicates the trap cell; the arrow indicates vegetative hyphae. (F) Confocal images of the localization of TEP1-GFP (scale bar, 100 μm). (G) Images of traps of the WT, tep1, tep4, tep5, and tep1 TEP1 following a 6-h exposure to 30 C. elegans nematodes; the images were captured 24 h later (scale bar, 200 μm). The arrowhead indicates the trap cell. (H) Quantification of the nematode escaping rate of 50 C. elegans (mean ± SEM). The data underlying this Figure can be found in S1 Data. BF, bright field; hpe, hours post-exposure; NTF, nematode-trapping fungi; TEP, Trap Enriched Protein; WT, wild-type.

https://doi.org/10.1371/journal.pbio.3002400.g004

We next looked for DUF3129 genes distribution within the fungal kingdom by scanning 2,504 fungal genomes comprised in the mycocosm fungal database [51] and 8 genomes of NTF (S5 Table). Interestingly, the majority of species (approximately 99%) contain fewer than 5 DUF3129 genes in their genomes, whereas this number dramatically increased to more than 20 in several species of NTF (Fig 4B). In order to verify and expand these initial findings, we performed orthogroup analysis of the DUF3129 genes in 19 selected high-quality genomes of diverse fungi (S5 Table). A total of 108 DUF3129 genes were distributed across 12 of these genomes. Of these DUF3129 genes, 86 were grouped into 23 orthogroups (Fig 4C). As shown in Fig 4C, the majority of the A. oligospora DUF3129 genes were specific to the Arthrobotrys genus. Additionally, our analysis revealed that the family of genes containing DUF3129 has expanded specifically within the NTF. This association implies that their expansion, followed by sequence divergence, likely contributed to the unique biology of the NTF species (S3 Fig). Our investigation into gain/loss patterns indicated that most of the duplication events involving DUF3129 genes occurred within Ascomycete NTF clades. Notably, even within the A. oligospora branch, we detected an additional 15 DUF3129 gene duplication events. This demonstrates that gene expansion has even occurred after speciation from A. vermicola. On the other hand, only a few duplication events were observed in non-NTF. The number of DUF3129 genes within plants, insects, humans, and other nematode pathogens remained limited, ranging from 0 to 6. These findings further emphasize an evolutionary expansion of DUF3129 genes, that is potentially linked to the predatory lifestyle of NTF.

To further investigate the potential roles of the DUF3129 domain-containing proteins, we first determined their expression pattern by in situ hybridization chain reaction (HCR). Probes for the 4 most highly expressed DUF3129 family genes, EYR41_003105, EYR41_001312, EYR41_005998, and EYR41_0011738, were designed and hybridized to A. oligospora culture at 10 hpe to C. elegans, where they showed high enrichment in trap cells (Fig 4D). We then investigated the expression and localization of the most highly expressed member, EYR41_003105, through fusion to GFP. EYR41_003105 was prominently enriched in traps (Fig 4E) and localized on the surface of the trap cell (Fig 4F), while showing weak or no signal in regular hyphae (Fig 4E). Moreover, live imaging revealed the flow of EYR41_003105 from hyphae into traps (S3 Movie). These results provide strong evidence that DUF3129 family genes accumulate in traps at both the transcriptional and translational level. Therefore, we named this class of highly expressed DUF3129 members Trap Enriched Protein (TEP) genes.

Having established the expression and localization of TEP genes to traps, we next examined their involvement in trap development and function. Targeted gene deletion mutants of EYR41_003105 (tep1), EYR41_0011738 (tep4), and EYR41_009694 (tep5) were generated and assessed for their ability to develop traps in response to C. elegans. We found that all mutants developed comparable number of traps to wild type in response to C. elegans, suggesting that these TEP genes do not play a role in nematode-sensing and trap morphogenesis (Figs 4G and S4). Instead, given TEP protein localization to the surface of trap cells, we hypothesized that TEPs may be critical for the function of the traps themselves. Adding C. elegans onto the traps of wild-type A. oligospora leads to their immediate capture. However, in the tep1 mutant, only 10% of nematodes were captured after 10 min of exposure to the traps, with trap efficacy restored upon reconstitution of TEP1 (Fig 4H and S1 and S4 Movies), indicating a potential role for TEP1 in trap adhesion needed to catch C. elegans.

Despite the fact that all 27 A. oligospora DUF3129 proteins share a conserved domain in the central portion of their sequence, a sequence alignment of the 4 most strongly expressed DUF3129 genes demonstrates high sequence diversity among these proteins (S5 Fig).

Structural predictions of TEP protein sequences indicated that the C-terminus region of the TEPs exhibit a high degree of disordered (S6 Fig). Because IDRs of proteins can drive liquid–liquid phase separation (LLPS), or the compartmentalization of proteins into “membraneless organelles” [77], it is possible that TEPs form liquid condensates in traps (S6 Fig). Aggregation of TEPs and other proteins might be required for trap adhesion. Taken together, these data support the idea that the DUF3129 family may have expanded in NTF and contribute to the function of traps.

Late time point induced protease and its potential role in digesting nematodes

At mid to late time points, we observed hyphal growth inside nematodes following penetration of the cuticle, corresponding to digestion. Hypothesizing proteases may be important for nematode digestion, we searched the transcriptome for proteases predicted by the MEROPs peptidase database. We identified a total of 323 protease-encoding genes; of these, 27, 56, 87, 123, and 19 genes were characterized as encoding aspartic, cysteine, metallo, serine, and threonine peptidases, respectively. Notably, genes encoding each class of protease (15 aspartic, 14 cysteine, 39 metallo, 62 serine, and 3 threonine) were differentially expressed in response to nematodes (Fig 5A). Consistent with a role in digestion, many of the protease genes were highly expressed at late time points when nematodes have already been captured (Fig 5B).

Download:

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

Fig 5. Late time point induced protease and its potential role in digesting nematodes.

(A) Expression profiles of differentially expressed protease transcripts upon exposure to C. elegans. The data underlying this Figure can be found in S1 Data. (B) TPM of the differentially expressed proteases. Arrows indicate the 4 highly up-regulated proteases. The data underlying this Figure can be found in S1 Data. (C) Image and quantification of hyphal colonization after penetration for 6–18 h and treatment with 1× PIC (scale bar, 100 μm). The data underlying this Figure can be found in S1 Data. (D) Image and quantification of hyphal colonization in C. elegans after being captured by traps for 6 h in WT, mcp1, and amp1 (scale bar, 100 μm). The data underlying this Figure can be found in S1 Data. hpe, hours post-exposure; PIC, protease inhibitor cocktail; TPM, Transcript Per Million; WT, wild type.

https://doi.org/10.1371/journal.pbio.3002400.g005

Given the large number of predicted proteases and potential for redundancy in digestive function, we reasoned that mutation of a single protease might not be sufficient to see obvious changes in digestion of nematodes. Therefore, to investigate the potential involvement of proteases in hyphal colonization and prey digestion, we utilized a protease inhibitor cocktail (PIC) to inhibit overall protease activity after trap morphogenesis and measured hyphal growth inside nematode bodies. Visualization and quantification of hyphal growth using cytosolic GFP expressed in the NTF revealed delayed hyphal colonization with PIC treatment (Fig 5C), suggesting an important role of proteases in both hyphal growth and prey digestion. Additionally, as a control, we conducted a milk plate assay, assessing the proteolytic activity of A. oligospora with or without PIC treatment and demonstrated that inhibitor-treated samples showed a smaller clear zone, but some residual enzymatic activity persisted despite the treatment (S7 Fig).

We next investigated individual proteases in the fungus. From our RNA-seq dataset, we identified 2 metallopeptidase genes that were highly differentially expressed at 24 hpe (Fig 5A and 5B). To investigate the involvement of metallopeptidases in nematode digestion, we generated targeted gene deletions of mep1 (MEtalloPeptidase 1) and amp1 (AMinoPeptidase 1) in the ku70 background. Both mutants were able to generate traps in response to nematodes, but as visualized by cytosolic GFP displayed minor defects in hyphal colonization (Fig 5D). Together, our findings suggest that proteases, specifically metallopeptidases, play a role in facilitating invasive hyphal growth and colonization of nematodes.