Intercellular communication is required for trap formation in the nematode-trapping fungus Duddingtonia flagrans

Abstract

Nematode-trapping fungi (NTF) are a large and diverse group of fungi, which may switch from a saprotrophic to a predatory lifestyle if nematodes are present. Different fungi have developed different trapping devices, ranging from adhesive cells to constricting rings. After trapping, fungal hyphae penetrate the worm, secrete lytic enzymes and form a hyphal network inside the body. We sequenced the genome of Duddingtonia flagrans, a biotechnologically important NTF used to control nematode populations in fields. The 36.64 Mb genome encodes 9,927 putative proteins, among which are more than 638 predicted secreted proteins. Most secreted proteins are lytic enzymes, but more than 200 were classified as small secreted proteins (< 300 amino acids). 117 putative effector proteins were predicted, suggesting interkingdom communication during the colonization. As a first step to analyze the function of such proteins or other phenomena at the molecular level, we developed a transformation system, established the fluorescent proteins GFP and mCherry, adapted an assay to monitor protein secretion, and established gene-deletion protocols using homologous recombination or CRISPR/Cas9. One putative virulence effector protein, PefB, was transcriptionally induced during the interaction. We show that the mature protein is able to be imported into nuclei in Caenorhabditis elegans cells. In addition, we studied trap formation and show that cell-to-cell communication is required for ring closure. The availability of the genome sequence and the establishment of many molecular tools will open new avenues to studying this biotechnologically relevant nematode-trapping fungus.

Fig 1. The nematode-trapping fungus D. flagrans produces adhesive traps, spores, and chlamydospores.

(A) Formation of a three-dimensional trapping network. (B) Trapped nematode C. elegans. Chlamydospores, normal spores and traps are labelled with arrows. (C) Complete degradation of C. elegans and fungal growth inside the nematode. (D) Asexual spore. (E) Chlamydospore. (F) Conversion of single trap-compartments into chlamydospores (star) compared to the conversion of vegetative cells (arrow). (G, H) Glycogen staining with Lugol‘s iodine. (I, J) Visualization of ring-like accumulation of chitin (star) at the contact zone of trap cells with the nematode cuticle. The cell wall of the fungus was stained with calcofluor-white (CFW). Nuclei of C. elegans were labelled with GFP [78]. (K) A X. index adult trapped in multiple D. flagrans networks (stars) after 48 h of co-incubation. The nematode body is completely filled with hyphae, revealed by CFW staining. (L, M) Magnification of the framed area in (I) and turned 90 degrees counter clockwise. The star indicates the trapping network around the head region of the nematode.

Fig 2. Orthologous clusters in the proteomes of four nematode-trapping fungi.

A Venn diagram plotted by OrthoVenn shows shared orthologous protein clusters among D. flagrans, A. oligospora, Da. haptotyla, and Dr. stenobrocha proteomes. Most gene families were shared between the closely related A. oligospora and D. flagrans. All four nematode-trapping fungi shared 4646 orthologous genes clusters.

Fig 3. Predicted secreted proteins, SSPs and host localization of putative effectors.

(A) 638 proteins without transmembrane domain (TMHMM) and with signal peptide (SignalP 4.0) were predicted to be secreted by WoLF PSORT. 249 of these proteins are SSPs (< 300 amino acids). 117 putative effector proteins or virulence factors were predicted with EffectorP 2.0. The sub-cellular localization of the mature proteins in the host was predicted using WoLF PSORT (animal mode). (B) Expression of pefB in hyphae of D. flagrans and in hyphae co-cultivated for 24 h with C. elegans. The expression was normalized to actin. (C) Scheme of PefB. The 274 amino acid (aa) protein comprises a 20 aa signal peptide at the N-terminus and an NLS between amino acids 54 and 76. Three repeat sequences (R1, 2, 3) were predicted by RADAR.

Fig 4. Establishment of a transformation system and fluorescent proteins in D. flagrans.

(A-C) Comparison of the growth of D. flagrans wild type and transformants containing the hygromycin-B resistance gene hph (A), the nourseothricin resistance gene nat (B), or geneticin G418 resistance gene neo (C) on PDA supplemented with the corresponding drugs at inhibitory concentrations. Colonies were grown for 7 days at 28°C. (D-F) Comparison of the growth of D. flagrans wild type and corresponding transformants on PDA after 3 days incubation at 28°C. (G) Expression of GFP under the control of the A. nidulans oliC promoter. (H) Expression of mCherry under the control of the A. nidulans oliC promoter. (I, J) D. flagrans wild-type control with the same parameters as in (G, H). Exposure time for (G-J) was 500 ms.

Fig 5. Visualization of D. flagrans nuclei using a H2B-mCherry fusion protein.

H2B-mCherry was expressed under the H2B promoter in D. flagrans. (A-C) Hyphae. (B) Nuclei were visualized using Hoechst 33342, (C) Merge of (A) and (B). (D-G) Nuclei in a conidium and a chlamydospore. (H-J) Nuclei in a trap. (I) Calcofluor white (CFW) staining of the cell wall. (J) Merge of (H) and (I). (K) The fluorescent tag allowed to track the infection inside the nematode. Merged image of the GFP, mCherry and CFW channel. Nuclei of C. elegans were labelled with GFP. The cell wall of the fungus was stained using CFW. The contrast of the CFW channel was adjusted to see the weak staining of the fungal cell inside the nematode (purple) and is displayed as “Fire” using the ImageJ lookup table. The pixel values of the cell wall outside the nematode are oversaturated (white). (D-K) Images are maximum intensity projections.

Fig 6. Analysis of the putative effector PefB.

(A) Secretion of the PefB-laccase fusion protein led to blue-green colored colonies on LNA plates containing 1mM ABTS. The wild-type strain remained uncolored (circle). (B) The N-terminal GFP fusion protein of PefB without signal peptide (GFP-PefBΔSP) localized in the cytoplasm and nuclei of D. flagrans. Nuclei were visualized using Hoechst 33342. (C) A C-terminal GFP fusion protein of PefBΔSP expressed in C. elegans shows nuclear localization. The gene was expressed using the elt-3 promoter. The cells indicated with the arrows are probably glia and neuronal cells. The picture was overexposed to see the nuclei. The pharynx of C. elegans shows red fluorescence due to the expression of the marker plasmid. In addition to the weakly stained nuclei on the side of the pharynx, large nuclei of intestinal cells below the pharynx showed the GFP signal (boxed area). (D) Another individual of C. elegans was stained with DAPI. In intestinal nuclei GFP and DAPI signals co-localized.

Fig 7. Targeted deletion of the sofT gene in D. flagrans using homologous recombination or Cas9 RNP.

(A) Deletion of the sofT gene using homologous recombination. (B) Scheme of the sofT-gene locus from wild type (upper panel) and the ΔsofT mutant (lower panel). In the ΔsofT strain the ORF (3.8 kb) is replaced by the 1.8 kb hygromycin resistance gene hph. Two 1 kb probes and the restriction enzyme HindIII were used for Southern analyses. (C) Confirmation of the ΔsofT mutant using Southern blot analysis. Genomic DNA was digested using HindIII. The 3‘-probe hybridized with a 2.9 kb fragment in wild-type, while in contrast it hybridized with a 4.4 kb fragment in the ΔsofT mutant. The membrane was stripped and hybridized with a hph-probe yielding in a single 4.4 kb fragment only in the ΔsofT mutant. (D) Deletion of the sofT gene using the Cas9 RNP transformation system. The 20bp target sequence in the sofT gene is underlined and located in the first exon of the gene. The PAM sequence (AGG) is shown in red. The ΔsofT mutant T3 showed insertion of the hygromycin-B resistance cassette three nucleotides upstream of the PAM sequence. (E, F) Single integration of the hygromycin-B resistance cassette in the ΔsofT mutant T3 was verified using Southern analysis. Genomic DNA of wild-type and the ΔsofT mutant T3 was digested using the XbaI restriction enzyme resulting in a 3.6 kb fragment after hybridization with the hph-probe.

Fig 8. The ΔsofT mutation led to the reduction of aerial mycelium, incomplete trap closure, absence of hyphal anastomoses and decrease in virulence.

(A) Growth of D. flagrans wild-type on PDA after 7 days. Note the production of aerial mycelium shown by the colony view from the side. (B) Prior to ring closure a small hyphal peg (arrow) is growing towards the trap-forming hypha. (C) Growth of the D. flagrans ΔsofT mutant on PDA after 7 days. Aerial mycelium was reduced as seen in view of the colony from the side. (D) Deletion of the sofT gene prevents ring closure during trap formation. (B+D) Fungal cell walls were stained using CFW. Images are maximum intensity projections. (E) Lack of hyphal anastomoses in the ΔsofT-mutant strain T3. Fungal cell walls were stained using CFW (magenta). The mutant strain (stained only in magenta) was grown on the same slide as the control strain (wild-type background) expressing cytoplasmic GFP (stained in green + magenta) for 24 h. Hyphal anastomoses were absent despite close physical contact of the hyphae in the ΔsofT mutant (stars) compared to numerous hyphal anastomoses in the control strain. (F) Complex formation of trapping networks and trapping of C. elegans by the ΔsofT mutant. (G, H) Comparison of the nematicidal activity (in %) of D. flagrans wild type and the ΔsofT mutant on LNA after 24 and 36 h, respectively. The nematicidal activity is divided into captured (alive, black bar) and digested (dead, grey bar) state of the nematode. Significance was tested using an unpaired t test (P < 0.05: ns = not significant; * = significant). (I) Comparison of the germination time and fungal growth of D. flagrans wild-type and the ΔsofT mutant. Significance was tested using an unpaired t test (P < 0.05).