*2.5. Microscopy Observations*

To determine the e ffect of *MrHex1* deletion on WB formation, a transmission electron microscope (TEM) analysis was conducted as described before [27]. The spores of the WT and mutants were inoculated in SDB for three days and the mycelia were harvested by filtration. After washing twice with distilled water, fungal samples were fixed in 2.5% glutaraldehyde in 0.1 M phosphate bu ffer solution (PBS; pH, 7.2) at 4 ◦C for 12 h, rinsed three times in the phosphate bu ffer, and fixed overnight in 1% osmium tetroxide bu ffered in 0.1 M cacodylate (pH, 7.0) at 4 ◦C. After rinsing three times with the phosphate bu ffer, samples were dehydrated in an ethanol gradients, infiltrated with a gradient series of epoxy propane, and then embedded in Epon resin for sectioning [27]. The ultrathin samples were treated in 2% uranium acetate and then lead citrate prior to the observations under a TEM (H-7650; Hitachi).

The mycelia collected from SDB were also used for fluorescent staining. After washing with PBS, the mycelia of each strain were jointly stained with DAPI (46-diamidino-2-phenylindole, Sigma-Aldrich, St. Louis, USA) and Calcofluor white (CW, Sigma-Aldrich) to detect nuclei and cell septa, respectively. A stock solution of DAPI (100 μg/mL) was prepared in water and diluted to 1–2 μg/mL in PBS for staining for 30 min. After washing with PBS three times, the samples were then treated with CW solution (4 μg/mL) bu ffered in 10% potassium hydroxide for 1 min prior to the observations with an Olympus microscope (BX51-33P, Tokyo, Japan).

#### *2.6. Appressorium Induction and Insect Bioassays*

Appressorium formation of the WT and mutants were induced on both a hydrophobic surface and locust hind wings [28]. Briefly, the spores of each strain were inoculated into individual polystyrene petri dishes (6 cm in diameter) containing 2 mL MM-Gly at a final concentration of 2 × 10<sup>5</sup> conidia/mL. After incubation for 24 h, the appressorium di fferentiation rates were recorded for > 300 conidia under a microscope. The locust hind wings were surface sterilized in 37% H2O2 for 5 min, washed twice with sterile water and immersed in conidial suspensions (2 × 10<sup>7</sup> spores/mL) for 20 s. The inoculated wings were lined on 0.8% water agar at 25 ◦C for 16 h. The Student's *t*-tests were conducted to compare the di fferences between strains.

Insect bioassays for the WT and mutants were conducted using the newly emerged last instar larvae of the mealworm *Tenebrio molitor* and silkworm *Bombyx mori*. Conidia were harvested from the two-week old PDA plates and suspended in 0.05% Tween-20 at the concentration of 1 × 10<sup>7</sup> conidia/mL. Insects were chilled on ice before immersion in spore suspensions for 30 s. In addition, injection assays were performed using the silkworm larvae. Each insect was injected at the second proleg with 10 μL of the suspensions each containing 1 × 10<sup>6</sup> conidia/mL. The mortality was recorded every 12 h and the median lethal time (LT50) was calculated by Kaplan–Meier analysis [31]. The control insects were treated with 0.05% Tween-20. Each treatment had three replicates with 15 insects each and the experiments were repeated twice.

#### **3. Results and Discussions**

#### *3.1. Length Variation of the Hex1 Proteins with Conserved C-termini*

The single copy and complete ORF of *Mrhex1* (MAA\_00782) encodes a protein possessing 392 amino acid (aa) residues and containing a carboxyl-terminal S1\_Hex1 domain (75 aa) like other proteins such as Hex-1 of *N. crassa* and HexA of *A. fumigatus* [18,20], however, with substantial total length variations between each other (Figure 1A). Further survey of the S1\_Hex1 domain proteins catalogued in GenBank obtained 460 proteins (single copy within each genome) from those fungal species belonging to the clade Sordaromyceta of the subphylum Pezizomycotina (Ascomycota) (Table S1). Unexpectedly, the substantial length variation was further evident for the Hex1 proteins from di fferent fungal species, ranging from 79 aa (EPQ66756, *Blumeria graminis f. sp. tritici*) to 2958 aa (ERF74742, *Endocarpon pusillum*) (Table S1). The misannotation of some of these proteins could not be

precluded. Statistically, the major distribution of Hex1 protein length is within the regions 470–534 aa (26.7%, 123/460), 405–469 aa (24.3%, 112/460) and 145–209 aa (13.9%, 64/460) (Figure 1B). The last group includes those characterized in *N. crassa* (Hex-1, NCU08332, 176 aa) and *A. nidulans* (AnHex1, AN4965, 221 aa). Length variations were also evident in different species from the same genus. For example, the Hex1 homologues from *Metarhizium* genus vary from 392 aa (MAA\_00782 and MAN\_09889, *M. anisopliae*) to 423 aa (MAC\_08379, *M. acridum*) and 454 aa (NOR\_02601, *M. rileyi*). Likewise, the proteins from the *Aspergillus* and other fungal genera are also highly variable in total length (Table S1). Similar to this finding, length differences have also been observed between other proteins belonging to the same family. Some protein domains are functionally permissive to length variation (termed length-deviant domains) while some others are less tolerant to length alteration (termed length-rigid domains) [32]. Considering the conserved function of Hex1 in WB formation in different fungi [13], it is therefore length-deviant for Hex1 proteins in term of their full lengths. It was found that the Hex-1 cleavage occurred in *N. crassa* [20]. The mature and functional length of Hex1 proteins remains to be determined in different fungi.

**Figure 1.** Schematic structuring and phylogenetic analysis of the selected Hex1 proteins. (**A**) Schematic structuring of MrHex1 and the selected homologues. Selected proteins are: Hex-1 from *Neurospora crassa*, HexA from *Aspergillus fumigatus* and AnHex-1 from *A. nidulans.* (**B**) Length variations of the Hex1 proteins from different fungi. *X* axis represents the amino acid (aa) length region of proteins. *Y* axis represents the number of proteins belonging to different length regions. (**C**) Conservation analysis of the Hex1-domain sequences. The sequences (75 aa each) were extracted from 460 Hex1 proteins from different fungal species. The N and C letters labeled at the bottom represent the Nand C-termini of the Hex1 domains. (**D**) Phylogenetic analysis of the selected Hex1 proteins. Protein sequences were retrieved from the selected fungal species and aligned to generate a neighbor joining tree with a Dayhoff substitution model and 1000 bootstrap replicates.

Irrespective of clear length variations among Hex1 proteins, a highly conserved C-terminus S1\_Hex1 domain with 75 aa residues is evident in each Hex1 protein, a typical feature of the length-rigid domain (Figure 1A; Table S1). In particular, the characteristic and specific peroxisome-targeting signal 1 (PTS1) tripeptide S/A-R/S-L [17] is present at the C-terminal of MrHex1 and other proteins (Figure 1C; Table S1), which is di fferent from the consensus PTS1 motif S/A/C-K/R/H-L/M reported before [13,17]. In particular, the PTS1 motif A-S-L is found from the putative Hex1 proteins of the plant pathogen *Monosporascus* genus and an S-S-L pattern from the Hex1 proteins of the *Valsa* genus (Table S1), where the second residue of serine (S) has not been suspected before. A phylogenetic NJ tree generated with 21 selected Hex1 proteins revealed that the clustering pattern of these proteins largely correlated with fungal speciation relationships (Figure 1D). For example, consistent with previous analyses [33,34], the Hex1 proteins from *Metarhizium* species evolved following the trajectory from the specialists (*M. rileyi* and *M. album*) to the generalist species (e.g., *M. robertsii* and *M. brunneum*) with a broad host range. In this respect, *Hex1* might have evolved by following fungal divergence and speciation after its birth in the ancestor of the Pezizomycotina fungi.

#### *3.2. MrHex1 E*ff*ecting on Fungal Growth, Sporulation and Stress Responses*

By checking the previous RNA-seq transcriptome data, relative to the conidial sample, *MrHex1* was found to be highly transcribed by the fungus during the formation of appressoria on locust wings [35]. To determine the function of *MrHex1* in *M. robertsii*, the gene was deleted and the obtained null mutant was also complemented by the verification of RT-PCR analysis (Figure 2A). Phenotypic growth assays showed that the deletion of *MrHex1* substantially reduced the fungal growth rate when compared with the WT and complemented (Comp) strains (Figure 2B,C). In addition, we found that the sporulation ability of Δ*MrHex1* was severely ( *P* = 3.94 × <sup>10</sup>−4) impaired when compared with that of the WT (Figure 2D). Unexpectedly, the gene-rescued mutant Comp also had a reduced level of conidiation when compared with the WT ( *P* = 2.64 × <sup>10</sup>−6). Otherwise, relative to the WT, both the null and rescued mutants did not show obvious defects in their stress responses against the challenges with the detergent SDS, osmotic stressors KCl and sorbitol, antifungal agen<sup>t</sup> farnesol or cell wall biosynthesis inhibitors CW and Congo red (Figure 3).

The requirement of Hex1 for asexual growth and sporulation has also been found in a few fungal species like *N. crassa* [17], *A. oligospora* [21], *F. graminearum* [19] and *M. oryzae* [12]. However, in contrast, the Δ*HexA* of *A. fumigutas* showed normal growth and sporulation like the WT strain [18]. Thus, similar to the observation of functional divergence between the conserved transcription factors in di fferent fungi [36,37], Hex1 also shows functional alterations in di fferent fungi. The fact that no obvious di fferences were observed in the stress responses between WT and Δ*MrHex1* provided further supports of species-dependent functional variations of Hex1 in di fferent fungi. For example, it has been found that, in contrast to Δ*MrHex1* and Δ*HexA* of *A. fumigatus* [18], *Hex1* null mutant of *A. oligospora* was sensitive to osmotic stress [21]. However, relative to the WT of *A. fumigatus*, Δ*HexA* became sensitive to SDS, farnesol, CW and Congo red [18], which was not the case for Δ*MrHex1* as we showed.

*J. Fungi* **2020**, *6*, 172

**Figure 2.** Gene deletion and phenotypic characterizations: (**A**) RT-PCR verification of gene deletion and complementation. Comp, the complemented mutant; β-*Tub*, the β-tubulin gene used as a control. (**B**) Phenotypic characterization of the wild-type (WT) and mutants after growth on potato dextrose agar (PDA) for 14 days. (**C**) Time-scale growth assays by measuring colony diameters. (**D**) The quantification of conidial production by WT and mutants after growth on PDA for 18 days. (**E**) Culture filtrates of different strains after reaction with ninhydrin. (**F**) Photometric estimation of the leaked amino acids after reaction with ninhydrin. The unit absorbance of A570 was normalized to the mycelium dry weight. Error bar on top of each column represents standard deviation.

#### *3.3. Requirement of Mrhex1 for Woronin Body Formation and Maintaining Cell Integrity*

Hex1 is the major WB protein in Pezizomycontina fungi [13,16,20]. To determine the function of MrHex1 in WB formation in *M. robertsii*, mycelial samples of the WT and mutant strains were subject to TEM analysis. As a result, the dense and characteristic WBs were evident on both sides of the WT cell septa but absent in Δ*MrHex1*. For Comp, after the examination of multiple section samples, the WT-like distribution of WBs was not observed but the WBs were found to be plugged or anchored

in proximity to the septum pore (Figure 4A). Thus, MrHex1 is similarly required for WB formation in *M. robertsii*. This kind of WB number and positioning differences between WT and the complemented mutant has also been found in *F. graminearum* [19] and *A. oligospora* [21]. It is noteworthy that WB positioning and localization are associated with the WB enveloping protein (i.e., the Woronin sorting complex protein, WSC) and a tethering protein Leashin (Lah) [38]. The *N. crassa* WSC-like protein (NCU07842 vs. MAA\_02499, 71% identity at amino acid level) is present in *M. robertsii*. However, in contrast to the finding in *Aspergillus* fungi [39], the large and nonconserved Lah-like protein remains elusive in *M. robertsii*. In addition, it has been known that the proper function of some genes requires their positions preferentially located in genomes [40]. The importance of the *Hex1* gene positioning remains to be determined for function. It could not be precluded at this stage that the imperfect issue of gene rescue might be due to the non-original position insertion.

**Figure 3.** Stress response assays. The spores of the WT and mutants were inoculated on PDA (**A**), PDA amended with farnesol at 50 μM (**B**), PDA plus Calcofluor white at 200 μg/mL (**C**), PDA plus KCl at 1.5 M (**D**), PDA plus sodium dodecyl sulphate (SDS) at 0.01% (**E**)**,** PDA plus Congo red at 250 μg/mL (**F**) and PDA plus sorbitol at1M(**G**). The phenotypes were photographed after inoculation with 2 μL of spore suspensions (started at 2 × 10<sup>7</sup> conidia/mL) diluted 10-fold for three days.

The anchoring of WBs to the septum pore in fungal cells is essential for preventing cytoplasmic leakage after cell damage [11], and maintaining cell integrity and heterogeneity [13]. We first performed ninhydrin reaction assays to determine if any difference between WT and mutants in terms of the amino acid leakage in culture filtrates. The results indicated that a deep purple color, the result of amino acid reaction with ninhydrin, was evident for the Δ*MrHex1* sample but not for the WT and Comp strains (Figure 2E). Consistently, the photometric assays indicated that the A570 value of the Δ*MrHex1* sample was significantly higher than those of the WT (*P* = 4.92 × <sup>10</sup>−4) and Comp (*P* = 4.03 × <sup>10</sup>−4) (Figure 2F). It was also found that the A570 value of Comp was higher than that of the WT (*P* = 0.0063) for an unclear reason. We also performed the joint fluorescent staining of different strains for detecting the distribution pattern of the nuclei within each hyphal cell. The results showed that only one nucleus was observed within one hyphal cell of the WT and Comp whereas more than one nucleus were frequently evident in Δ*Mrhex1* cells, especially within the cells close to the injured end (Figure 4B). MrHex1 is therefore functionally important in maintaining cell integrity and heterogeneity in *M. robertsii*. Likewise, it has been shown that the hyphal heterogeneity of *A. oryzae Hex1* null mutant was affected [15]. It has also been shown that the peroxisome-related WB formation affects fungal secondary metabolisms [13], which remains to be determined in *M. robertsii*.

**Figure 4.** Microscopic observations: (**A**) transmission electron microscope observation showing the presence or absence of Woronin bodies (arrowed) in the WT and mutant cells. Bar, 0.5 μm; (**B**) the co-staining of the mycelium cells for detecting nuclei and septa (arrowed). The broken end of the Δ*MrHex1* mycelium is arrowed for its bright field image. Bar, 5 μm.

#### *3.4. Defects of Mrhex1 Null Mutant in Appressorium Formation and Topical Infection of Insects*

We then performed infection structure induction and insect bioassays with the WT and mutant strains. Appressorium formation was induced on both the hydrophobic surfaces and locust hind wings. As a result, we found that appressorium production was considerably impaired for Δ*MrHex1* when compared with the WT and Comp under both conditions (Figure 5A). Statistically, the rate of appressorium production by Δ*Mrhex1* (23.3% ± 2.53) declined significantly ( *P* < 0.001) when compared with those formed by WT (83.6% ± 5.69) and Comp (82.9% ± 4.38) on a hydrophobic surface. The failure of septal pore sealing might lead to the defects in building up turgor pressure within appressorium cells. Considering that the mutants of *M. robertsii* with impaired abilities in generating cellular turgor pressure could still form appressoria [10,27,41], the defect of Δ*MrHex1* in appressorium formation might not be due to the turgor generation failure of the mutant. The exact mechanism between WB and infection structure formations requires further investigation.

**Figure 5.** Appressorium induction and insect survival assays. ( **A**) Microscopic examination of appressorium formation by the WT and mutants on hydrophobic surface (upper panels) and locust hind wings (lower panels). CO, conidium; AP, appressorium. Bar, 5 μm. (**B**) Survival of the mealworm larvae after topical infection. ( **C**) Survival of the silkworm larvae after topical infection. ( **D**) Survival of the silkworm larvae after injection.

Consistent with the mutant defect in appressorium formation, the topical infection of the mealworm and silkworm larvae revealed that the virulence reduction of Δ*MrHex1* was evident (Figure 5B,C). Thus, the LT50 value of Δ*MrHex1* (4.98 ± 0.18 days) was significantly higher than those of the WT (3.94 ± 0.12 days; χ2 = 25.12, *P* < 0.0001) and Comp (3.80 ± 0.15 days; χ2 = 22.24, *P* < 0.0001) during the topical infection of *T. molitor* larvae. For the topical infection of silkworm larvae, the LT50 value of Δ*MrHex1* (4.02 ± 0.14 days) was also higher than those of the WT (3.48 ± 0.09 days; χ2 = 11.04, *P* < 0.001) and Comp (3.62 ± 0.10 days; χ2 = 7.0, *P* < 0.01). However, survival dynamics were similar between the WT and mutant strains during the injection assays (χ<sup>2</sup> < 2.0, *P* > 0.1) of the silkworm larvae (Figure 5D). These results confirmed that the deletion of *MrHex1* impaired the fungal ability to penetrate host cuticles due to the null mutant defect in appressorium formation and or the generation of turgor pressure. Considering that the sporulation ability of Δ*MrHex1* was impaired, the mycosis of insect cadavers killed by either topical infection or injection might also be negatively affected for Δ*MrHex1* when compared with the WT and Comp strains.

Similar to our observations, the defects in appressorium formation and therefore virulence reduction were also observed in the Δ*Hex1* of *M. oryzae* [12]. Likewise, the failure of trap formation was evident for the Δ*AoHex1* of *A. oligospora* and the mutant lost its ability to capture nematodes [21]. Both the deletion and overexpression of *FgHex1* in *F. graminearum* reduced fungal infectivity [19]. However, intriguingly, the deletion of *CoHex1* in the cucumber anthracnose fungus *Colletotrichum orbiculare* did not produce any detectable defects in appressorium formation and infectivity [42]. This kind of species-dependent phenotypic diversity of *Hex1* deletion mutants indicates again the functional alterations of this conserved gene in different fungi.
