The High Osmolarity Glycerol (HOG) Pathway Functions in Osmosensing, Trap Morphogenesis and Conidiation of the Nematode-Trapping Fungus Arthrobotrys oligospora

Hog1, a mitogen-activated protein kinase (MAPK), has been identified in diverse fungal species, and it regulates various cellular processes, such as osmoadaptation, nutrient-sensing, and pathogenesis. However, the roles that Hog1 plays in nematode-trapping fungi were previously unclear. Here, we characterized orthologs of Saccharomyces cerevisiae Hog1 and membrane mucin Msb2 in the nematode-trapping fungus Arthrobotrys oligospora. We generated gene deletion mutants of HOG1 and MSB2 in A. oligospora, and characterized their roles in osmosensing, growth, and trap morphogenesis. We found that both hog1 and msb2 mutants were highly sensitive to high osmolarity. Predation analyses further revealed that hog1 and msb2 deletion caused a reduction in trap formation and predation efficiency. Furthermore, HOG1 is required for conidiation in A. oligospora, demonstrating its critical role in this developmental pathway. In summary, this study demonstrated that the conserved Hog1 and Msb2 govern physiology, growth and development in the nematode-trapping fungus A. oligospora.


Introduction
The ability to sense and respond appropriately to the environmental changes is essential for any organism. For example, eukaryotic cells regulate their internal osmolarity in response to osmolarity changes in the environments. In the budding yeast Saccharomyces cerevisiae, cells respond to increases in external osmolarity by increasing intracellular glycerol synthesis [1]. In the fungal kingdom, osmosensing has been linked to fungal development, metabolism and virulence [2]. In some plant pathogens, such as Fusarium graminearum and Botrytis cinerea, the pathway regulating responses to hyperosmotic stress is also involved in hyphal development and pathogenesis [3,4].
Carnivorism is universal in the animal kingdom; however, certain plants and fungi that inhabit in low-nutrient environments have also evolved the ability to prey on small animals such as insects and nematodes. To access nitrogen resources, the Venus flytrap, Dionaea muscipula, attracts prey by emitting of volatile organic compounds that simulate food odors [5]. Similarly, a number of fungi, known as nematode-trapping fungi (NTF), can develop complex trapping structures to capture and

Identification and Deletion of the KU70, HOG1 and MSB2 Homologs in A. oligospora
The KU70 homologs of A. oligospora were identified in Blast2GO 5 Pro [23] by performing a local BLAST analysis. We used the amino acid sequences of Ku70 orthologs from S. cerevisiae (UniProt ID P32807) as the queries and the proteome of the A. oligospora TWF154 strain as the database. The same method was used to identify the HOG1 and MSB2 homologs of A. oligospora. The amino acid sequences of Hog1 and Msb2 orthologs from Aspergillus nidulans (UniProt ID Q9P419, Q5AXD9), Neurospora crassa (Q96TL5), F. graminearum (P0C431, I1RNP9), and S. cerevisiae (P32485, P32334) were used as the queries.
KU70, HOG1 and MSB2 were deleted by means of a homologous recombination-based strategy, which reported previously for A. oligospora [24,25]. Primers used for generating the KU70, HOG1 and MSB2 deletion mutants are described in the Supplementary Materials.

Generation of Target Gene Deletion Mutants
In a previous study [26], we generated deletion mutants via homologous recombination and determined that the homologous recombination rate in A. oligospora is extremely low (~3%). Therefore, here, we generated a non-homologous end-joining-deficient strain (ku70) to increase the efficiency of targeted genes deletion in fungi [27][28][29]. An overlap PCR-based construct (donor DNA) was obtained by fusing the 1.5 kb long sequences flanking the open reading frames of the target genes (amplified from genomic DNA of the A. oligospora TWF154 strain) to the hygromycin-B resistance cassette (amplified from vector pAN7-1 [30]). The construct was introduced into protoplasts of the A. oligospora TWF154 strain via PEG-mediated transformation. Generation of protoplasts in A. oligospora was carried out based on previously described protocols [26]. For transformations,~10 6 protoplasts were gently mixed with 5 µg of donor DNA and incubated on ice for 30 min, after which five volumes of PTC (40% w/v PEG 4000, 10 mM Tris-HCl pH 7.5, 50 mM CaCl 2 ) were added and incubated at room temperature for 20 min. Then, 50 mL of molten regeneration agar (0.3% w/v yeast extract, 0.3% w/v casein, 20.5% w/v sucrose) at 45 • C containing 100 µg/mL hygromycin B was added to the protoplast mixture and poured into Petri dishes. After 7 days at 25 • C, transformants grown on selective medium were screened for gene replacement by rapid genomic extraction and PCR. Briefly, a small cluster of aerial hyphae was added to PCR tubes containing 30 µL of DNA TE buffer (10 mM Tris-HCl pH 8, 0.5M EDTA pH 8), which were placed in a microwave and heated for 2 min. After the samples cooled down to room temperature, they were re-heated in the microwave. Finally, the samples were placed in a −20 • C refrigerator to cool down for 10 min. Two µL of the samples were then used as template for PCR reactions.
For deletion of the A. oligospora HOG1 gene (EYR41_010451) and MSB2 (EYR41_007180), we fused the overlap PCR-based construct with flanking the open reading frame of the target gene to a nourseothricin sulfate resistance cassette (amplified from vector pRS41N [31]). The construct was introduced into protoplasts of the ku70 (control) strain via PEG-mediated transformation. For transformations,~10 5 protoplasts were used and the concentration of nourseothricin sulfate was 250 µg/mL.

Phenotypical Characterization
To quantify trap formation, fungal isolates were grown on low nutrient medium (LNM) for 5 days and then transferred to a fresh LNM plate for a further 48 h (25 • C, dark). Thirty N2 Caenorhabditis elegans L4-stage larvae were then introduced onto the fungal plates for 6 h, after which the animals were removed by washing with sterile water. Twenty-four hours after exposure, micrographs were taken of fungi using a Nikon SMZ 745T stereo microscope. Three images were randomly taken of each fungal plate using 40× magnification, and the sum of all traps in the three images was recorded.
To estimate survival rates of C. elegans upon exposure to hog1 and msb2 mutant lines, 80 young adult stage N2 C. elegans were introduced onto mutant fungal plates for 6 h. Numbers of moving nematodes were assessed every 2 h for a total of 12 h using a stereo microscope.

Statistics
For trap quantification analysis, a two-tailed unpaired Student's t test was performed to determine the statistical difference between the control and experimental samples using GraphPad Prism 8. For survival rate analysis, a two-tailed unpaired Student's t test was conducted to determine the statistical difference between the control and experimental groups after being exposed to C. elegans for 12 h. A value of p < 0.05 was considered statistically significant.

Identification of the KU70 Gene in A. oligospora
The homologous recombination rate in A. oligospora was extremely low (~3%) in a previous study [26]. To enhance the recombination rate, we generated a non-homologous end-joining-deficient strain (ku70) which has been reported to increase the efficiency of gene deletion in fungi [27][28][29]. We surveyed the genome of A. oligospora strain TWF154 [26] to search for the orthologs of the S. cerevisiae KU70 gene, and identified EYR41_000878 (KU70).
The ku70 mutants did not exhibit any obvious growth defects under normal growth conditions ( Figure 1A). To examine if the deletion of KU70 in A. oligospora influenced trap morphogenesis, we compared phenotypic differences between the TWF154 (WT) and the ku70 lines upon exposure to C. elegans. We observed that independent ku70 mutants formed traps and exhibited nematocidal activity comparable to that of the wild-type strain ( Figure 1B). Southern blot analysis revealed that some ku70 deletion strains we obtained contained multiple ectopic integrations of the repair DNA ( Figure 1C). Therefore, we decided to conduct gene disruption experiments in the ku70-2 mutant line to facilitate homologous recombination.

Identification of the HOG1 and MSB2 Genes in A. oligospora Suggests a Conserved Role in Osmosensing
We were interested in examining the role of the HOG MAPK pathway in responses of A. oligospora to C. elegans presence. First, we surveyed the genome of A. oligospora strain TWF154 [26] to search for the orthologs of the S. cerevisiae HOG1 and MSB2 genes, and identified two targets, i.e., EYR41_010451 (HOG1) and EYR41_007180 (MSB2), respectively. The overall sequence identities between S. cerevisiae and A. oligospora Msb2 and Hog1 proteins were ~44.5% and ~54.8%, respectively. The predicted A. oligospora Msb2 protein has a domain architecture similar to that of S. cerevisiae Msb2 (

Identification of the HOG1 and MSB2 Genes in A. oligospora Suggests a Conserved Role in Osmosensing
We were interested in examining the role of the HOG MAPK pathway in responses of A. oligospora to C. elegans presence. First, we surveyed the genome of A. oligospora strain TWF154 [26] to search for the orthologs of the S. cerevisiae HOG1 and MSB2 genes, and identified two targets, i.e., EYR41_010451 (HOG1) and EYR41_007180 (MSB2), respectively. The overall sequence identities between S. cerevisiae and A. oligospora Msb2 and Hog1 proteins were~44.5% and~54.8%, respectively. The predicted A. oligospora Msb2 protein has a domain architecture similar to that of S. cerevisiae Msb2 (Figure 2A,C), including an N-terminal signal sequence (22 amino acids), a large extracellular domain (amino acids 23 to 793), and one transmembrane domain (TM; 23 amino acids). A. oligospora Hog1 also has a similar domain structure to that of S. cerevisiae Hog1 ( Figure 2B,D), including a highly conserved protein kinase domain (amino acids 1 to 272), a common docking (CD) domain (amino acids 273 to 284) and a Pbs2-domain binding 2 (PBD-2) domain (amino acids 290 to 320). Both the CD domain and the PBD-2 are required for the activation of Hog1 by Pbs2 in S. cerevisiae [32]. These results indicate that these two genes might play important roles, including pathogenesis, in nematode-trapping fungi. We therefore generated targeted gene deletion mutants for HOG1 and MSB2 mutants via a homologous recombination-based strategy reported previously for A. oligospora [24,25]. The deletion of HOG1 or MSB2 genes was confirmed by Southern blot analysis (Supplementary Material Figure S1).
[2], so we investigated the effect of treating our A. oligospora hog1 and masb2 mutants with excess sodium chloride (NaCl). On potato dextrose agar (PDA) with 1% (w/v) NaCl, the hog1 mutant had not grown after incubation for 4 days, whereas the msb2 mutant showed a growth defect relative to the ku70 control ( Figure 3A). These results indicate that the osmoregulatory role of the HOG MAPK cascade is conserved in A. oligospora. Given that the msb2 mutant exhibited only a slight growth defect in the presence of excess NaCl, it is plausible that Msb2 is not the only receptor operating upstream of the HOG pathway.
The hog1 mutant did not exhibit any obvious growth defects under normal growth conditions ( Figure 3A), but a more detailed morphological analysis using scanning electron microscopy (SEM) revealed a defect in conidial formation ( Figure 3B). In contrast, though the msb2 mutant presented reduced growth in PDA alone, its conidiation was as robust as the wild-type ( Figure 3B). Thus, Hog1 appears to play a role in osmoregulation and the asexual development of A. oligospora. In contrast, Msb2 appears to control hyphal growth and to participate in osmoregulation.  The HOG MAPK pathway is important in regulating responses to hyperosmotic stress in fungi [2], so we investigated the effect of treating our A. oligospora hog1 and masb2 mutants with excess sodium chloride (NaCl). On potato dextrose agar (PDA) with 1% (w/v) NaCl, the hog1 mutant had not grown after incubation for 4 days, whereas the msb2 mutant showed a growth defect relative to the ku70 control ( Figure 3A). These results indicate that the osmoregulatory role of the HOG MAPK cascade is conserved in A. oligospora. Given that the msb2 mutant exhibited only a slight growth defect in the presence of excess NaCl, it is plausible that Msb2 is not the only receptor operating upstream of the HOG pathway.
The hog1 mutant did not exhibit any obvious growth defects under normal growth conditions ( Figure 3A), but a more detailed morphological analysis using scanning electron microscopy (SEM) revealed a defect in conidial formation ( Figure 3B). In contrast, though the msb2 mutant presented reduced growth in PDA alone, its conidiation was as robust as the wild-type ( Figure 3B). Thus, Hog1 appears to play a role in osmoregulation and the asexual development of A. oligospora. In contrast, Msb2 appears to control hyphal growth and to participate in osmoregulation.

The Hog1 Pathway is Required for Proper Formation of Traps during Nematode Induction
To examine if the HOG pathway functions in A. oligospora trap morphogenesis, we compared phenotypic differences between the control and the mutant fungal lines upon exposure to C. elegans. We observed that both hog1 and msb2 mutants formed traps, but trap numbers were reduced when compared to the control ( Figure 4A). Quantification of trap morphogenesis revealed that the trap numbers in the hog1 and the msb2 mutant reduced to ~70% and ~50% of the ku70 control, respectively ( Figure 4C). Next, we conducted SEM analysis to further examine the detail trap morphology in the wild-type and mutant strains. We observed that the hog1 mutant exhibited defective development of three-dimensional trap structures; however, the traps formed by the msb2 mutant were comparable to those of the ku70 control ( Figure 4B).
Both hog1 and msb2 mutants developed fewer traps upon C. elegans exposure; therefore, we evaluated their trapping efficiency. In the ku70 control, only ~28% of the nematodes applied to A.

The Hog1 Pathway Is Required for Proper Formation of Traps during Nematode Induction
To examine if the HOG pathway functions in A. oligospora trap morphogenesis, we compared phenotypic differences between the control and the mutant fungal lines upon exposure to C. elegans. We observed that both hog1 and msb2 mutants formed traps, but trap numbers were reduced when compared to the control ( Figure 4A). Quantification of trap morphogenesis revealed that the trap numbers in the hog1 and the msb2 mutant reduced to~70% and~50% of the ku70 control, respectively ( Figure 4C). Next, we conducted SEM analysis to further examine the detail trap morphology in the wild-type and mutant strains. We observed that the hog1 mutant exhibited defective development of three-dimensional trap structures; however, the traps formed by the msb2 mutant were comparable to those of the ku70 control ( Figure 4B). J. Fungi 2020, 6, x FOR PEER REVIEW 7 of 11 oligospora fungal plates had survived after 12 h ( Figure 4D). In contrast, survival rates of nematodes exposed to the hog1 and msb2 mutants were ~50 to 60% ( Figure 4D). Together, the reduction in trap numbers and the development of inadequate trapping structures demonstrate that the hog1 and msb2 mutants display impaired predation ability when compared to the wild-type.  Both hog1 and msb2 mutants developed fewer traps upon C. elegans exposure; therefore, we evaluated their trapping efficiency. In the ku70 control, only~28% of the nematodes applied to A. oligospora fungal plates had survived after 12 h ( Figure 4D). In contrast, survival rates of nematodes exposed to the hog1 and msb2 mutants were~50 to 60% ( Figure 4D). Together, the reduction in trap numbers and the development of inadequate trapping structures demonstrate that the hog1 and msb2 mutants display impaired predation ability when compared to the wild-type.

Discussion
NTF hold great potential to be utilized as biological control agents in agricultural settings. However, the interactions between nematodes and NTF have not been investigated in detail at the molecular level. In recent years, a number of studies have begun to uncover the molecular mechanisms underlying the biology of these specialized predators. For example, it has been reported that two MAP kinases, Slt2 in the cell wall integrity (CWI) pathway and Ime2 in the inducer of meiosis 2 (Ime2) pathway, are involved in A. oligospora trap formation and nematocidal activity [33,34]. These data suggest that MAPK pathways play important roles in interaction between NTF and their prey.
The Hog1 MAPK pathway is thought to be primarily involved in osmoregulation in S. cerevisiae, whereas its orthologs in various filamentous fungi often have additional biological functions [2,10]. In Figure 5, we illustrated schematic models for the role of Hog1 and Msb2 playing in S. cerevisiae, M. oryza, and A. oligospora. In S. cerevisiae, functionally redundant osmosensors activate downstream Hog1 MAPK to respond to changes in osmolarity in the extracellular environment [13,15]. In M. oryzae, Osm1 (ortholog of Hog1) activated by upstream osmosensor Sln1 plays a critical role in osmoregulation but is dispensable for plant infection [35]. In contrast, M. oryzae Msb2 activates Pmk1, an ortholog of the S. cerevisiae Fus3 MAPK-mediating mating process, to regulate appressorium formation and penetration [36]. In other plant pathogens, such as Ustilago maydis, deletion of MSB2 also impedes the differentiation of appressoria on inductive surfaces [37]. In Fusarium oxysporum, which does not produce appressoria, mutants lacking Msb2 present reduced penetration of host roots [38]. These findings indicate that Msb2 has evolved diverse functions in different fungal species.

Discussion
NTF hold great potential to be utilized as biological control agents in agricultural settings. However, the interactions between nematodes and NTF have not been investigated in detail at the molecular level. In recent years, a number of studies have begun to uncover the molecular mechanisms underlying the biology of these specialized predators. For example, it has been reported that two MAP kinases, Slt2 in the cell wall integrity (CWI) pathway and Ime2 in the inducer of meiosis 2 (Ime2) pathway, are involved in A. oligospora trap formation and nematocidal activity [33,34]. These data suggest that MAPK pathways play important roles in interaction between NTF and their prey.
The Hog1 MAPK pathway is thought to be primarily involved in osmoregulation in S. cerevisiae, whereas its orthologs in various filamentous fungi often have additional biological functions [2,10]. In Figure 5, we illustrated schematic models for the role of Hog1 and Msb2 playing in S. cerevisiae, M. oryza, and A. oligospora. In S. cerevisiae, functionally redundant osmosensors activate downstream Hog1 MAPK to respond to changes in osmolarity in the extracellular environment [13,15]. In M. oryzae, Osm1 (ortholog of Hog1) activated by upstream osmosensor Sln1 plays a critical role in osmoregulation but is dispensable for plant infection [35]. In contrast, M. oryzae Msb2 activates Pmk1, an ortholog of the S. cerevisiae Fus3 MAPK-mediating mating process, to regulate appressorium formation and penetration [36]. In other plant pathogens, such as Ustilago maydis, deletion of MSB2 also impedes the differentiation of appressoria on inductive surfaces [37]. In Fusarium oxysporum, which does not produce appressoria, mutants lacking Msb2 present reduced penetration of host roots [38]. These findings indicate that Msb2 has evolved diverse functions in different fungal species.
In A. oligospora, we found that the hog1 and the msb2 mutants were hypersensitive to osmotic stress, especially for the hog1 mutant that was completely inhibited when excess NaCl was present in the culture medium. In some pathogenetic fungi, deletion of HOG1 has been shown to reduce conidia numbers [20,39], and we also found that our A. oligospora hog1 mutant was defective in conidial formation. Thus, Hog1 not only participates in the osmoregulation, but also regulates asexual reproduction in A. oligospora.  In A. oligospora, we found that the hog1 and the msb2 mutants were hypersensitive to osmotic stress, especially for the hog1 mutant that was completely inhibited when excess NaCl was present in the culture medium. In some pathogenetic fungi, deletion of HOG1 has been shown to reduce conidia numbers [20,39], and we also found that our A. oligospora hog1 mutant was defective in conidial formation. Thus, Hog1 not only participates in the osmoregulation, but also regulates asexual reproduction in A. oligospora.
Both the hog1 and the msb2 mutants were capable of forming traps upon exposure to C. elegans, but in both cases, trap numbers were reduced. Notably, the hog1 mutant displayed a defect in developing proper three-dimensional trap structures, although those of the msb2 mutant were comparable to ku70 control traps. These observations potentially explain the decreased nematocidal activity of our mutants, as assessed by C. elegans survival.
Our msb2 mutant line shared several but not all of the phenotypes displayed by the hog1 mutant. Accordingly, we speculate that additional functionally redundant osmosensors also regulate Hog1 activation ( Figure 5), a scenario that has also been described for the HOG pathway of budding yeast [13,14]. Besides, in U. maydis and M. oryzae, it has been reported that compared to individual deletion mutants of SHO1 or MSB2, the sho1 msb2 double mutant completely abolished appressorium formation [36,37]. Therefore, we hypothesize that SHO1 and SLN1, which are both present in the genome of A. oligospora, likely also function as upstream osmosensors. Overall, our study demonstrates that the Hog1 MAPK pathway displays a conserved role in osmoregulation, and is also critical in regulation of conidiation and trap morphogenesis in the NTF A. oligospora.