Essential Role of WetA, but No Role of VosA, in Asexual Development, Conidial Maturation and Insect Pathogenicity of Metarhizium robertsii

WetA and VosA regulate conidiation and conidial maturation required for the life cycle of Beauveria bassiana, like they do in Aspergillus, but remain functionally unexplored in Metarhizium robertsii, another hypocrealean pathogen considered to have evolved insect pathogenicity ~130 million years later than B. bassiana. This study reveals a similar role of WetA ortholog in asexual development, conidial maturation, and insect pathogenicity, and also its distinctive role in mediating some other conidial maturation-related cellular events, but has functional redundancy of VosA in M. robertsii. ABSTRACT Conidial maturation, which is crucial for conidial quality, is controlled by the asexual development activator WetA and the downstream, velvety protein VosA in Aspergillus. Their orthologs have proved functional in conidial quality control of Beauveria bassiana, as seen in Aspergillus, but are functionally unexplored, in Metarhizium robertsii, another hypocrealean insect pathogen. Here, WetA and VosA prove essential and nonessential for M. robertsii's life cycle, respectively. Disruption of wetA increased hyphal sensitivity to oxidative stress and Congo red-induced cell wall stress, but had little impact on radial growth. The ΔwetA mutant was severely compromised in conidiation capacity and conidial quality, which was featured by slower germination, decreased UV resistance, reduced hydrophobicity, and deformed hydrophobin rodlet bundles that were assembled onto conidial coat. The mutant's virulence was greatly attenuated via normal infection due to a blockage of infection-required cellular processes. All examined phenotypes were unaffected for the ΔvosA mutant. Intriguingly, mannitol was much less accumulated in the 7- and 15-day-old cultures of ΔwetA and ΔvosA than of control strains, while accumulated trehalose was not detectable at all, revealing little a link of intracellular polyol accumulation to conidial maturation. Transcriptomic analysis revealed differential regulation of 160 genes (up/down ratio: 72:88) in ΔwetA. These genes were mostly involved in cellular component, biological process, and molecular function but rarely associated with asexual development. Conclusively, WetA plays a relatively conserved role in M. robertsii’s spore surface structure, and also a differentiated role in some other cellular processes associated with conidial maturation. VosA is functionally redundant in M. robertsii unlike its ortholog in B. bassiana. IMPORTANCE WetA and VosA regulate conidiation and conidial maturation required for the life cycle of Beauveria bassiana, like they do in Aspergillus, but remain functionally unexplored in Metarhizium robertsii, another hypocrealean pathogen considered to have evolved insect pathogenicity ~130 million years later than B. bassiana. This study reveals a similar role of WetA ortholog in asexual development, conidial maturation, and insect pathogenicity, and also its distinctive role in mediating some other conidial maturation-related cellular events, but has functional redundancy of VosA in M. robertsii. The maturation process vital for conidial quality proves dependent on a role of WetA in spore wall assembly but is independent of its role in intracellular polyol accumulation. Transcriptomic analysis reveals a link of WetA to 160 genes involved in cellular component, biological process, and molecular function. Our study unveils that M. robertsii WetA or VosA is functionally differential or different from those learned in B. bassiana and other ascomycetes.

structural integrity, and asexual development, and play interdependent, overlapping or distinct roles in morphological development and metabolic remodeling in A. nidulans (45). In Fusarium graminearum, WetA was characterized as the production of longer conidia with fewer septa, the elevation of conidial sensitivity to oxidative and heat stresses, and the reduced survival rate of aged conidia in the absence of its coding gene (46). Likewise, abnormal conidia showing loose cell walls, delayed germination, and altered sensitivities to stress cues were observed in the DwetA mutants of Penicillium spp. (47,48). Exceptionally, deletion of wetA in Monascus ruber had no effect on conidial morphology, size, number, structure, and germination (49). The previous studies have revealed an array of important, but differential, cellular processes associated with conidial maturation in different ascomycetes. The role of VosA in conidial maturation remains unclear in the fungi other than Aspergillus.
The wetA and vosA orthologs in fungal insect pathogens have been functionally elucidated only in B. bassiana, in which wetA plays a greater role than vosA in sustaining conidial yield, morphology, size and density, hydrophobicity, cell wall integrity, and virulence but vice versa in trehalose accumulation into mature conidia (50). The different conidiation modes of M. robertsii and B. bassiana implicate that their conidial maturation may involve certain subtly different cellular events. This study seeks to characterize functions of wetA and vosA in M. robertsii. Our emphasis is placed upon the effects of targeted gene disruptions on asexual development, conidial maturation, stress tolerance, and insect pathogenicity. Unveiling the roles of wetA and vosA in M. robertsii will deepen an insight into the genetic control of conidial maturation in insect-pathogenic Hypocreales, and help to improve large-scale production technology and management of high-quality conidia as active ingredients of fungal pesticides.

RESULTS
Domain architecture of WetA and VosA in M. robertsii. Orthologous WetA and VosA were located in M. robertsii genome (4) via BLASTp search with the query sequences of WetA and VosA in A. nidulans (XP_659541, 555 amino acids [aa]; Q5BBX1, 430 aa) or B. bassiana (EJP64951, 639 aa; EJP70154, 273 aa). The located WetA (EFZ01616, 645 aa) share higher sequence identity with the query of A. nidulans (71.2% with evalue 7 Â 10 28 , total score 53.9 and coverage 10%) than of B. bassiana (56.6% with evalue 1 Â 10 228 , total score 215 and coverage 90%). The located VosA sequence (EFY98723, 397 aa) is more identical to the query of B. bassiana (77.04% with e-value 1 Â 10 2111 , total score 326 and coverage 71%) than of A. nidulans (35.71% with e-value 4 Â 10 228 , total score 99.4 and coverage 35.71%). Conventional (SMART or NCBI) domain analysis revealed a lack of any known domain in WetA orthologs (Fig. 1A), although A. nidulans WetA was predicted to act as a DNA-binding factor at http://www.pantherdb .org/ and take part in asexual spore wall assembly, conidium formation, and pigment biosynthesis. Orthologous VosA features 2 partially overlapping Velvety domains at Cterminus in B. bassiana or M. robertsii but at N-terminus in A. nidulans. A nuclear localization signal (NLS) motif was predicted from the C termini of 3 WetA orthologs at high probabilities of 0.727 to 0.814 and the VosA C termini of B. bassiana and M. robertsii at lower probabilities (#0.368). The N-terminal NLS motif of A. nidulans VosA was predicted a probability of 0.498. Obviously, either WetA or VosA orthologs in the mentioned fungi are similar in domain architecture but different in molecular size. In phylogeny, WetA and VosA orthologs found in the genomes of selected ascomycetes are clustered to the clades of their host lineages (Fig. S1).
Transcriptional profiles and subcellular localization of WetA and VosA. Both wetA and vosA were expressed in the wild-type strain M. robertsii ARSEF 2575 (WT hereafter) during a 7-day incubation on potato dextrose agar (PDA) at the optimal regime of 25°C and 12 h/12 h light/dark photoperiod (Fig. 1B); wetA was consistently expressed at higher level than vosA, despite their similar upregulation trends over the time of incubation. Green fluorescence-tagged WetA fusion protein (WetA-GFP) expressed in WT were localized in both nuclei and cytoplasm of hyphal cells, but accumulated more in the nuclei (Fig. 1C). The VosA-GFP fusion protein was more evenly distributed in the nuclei and cytoplasm. These observations are in accordance with the NLS motif of WetA predicted at much higher probability than of VosA in M. robertsii.
Roles of WetA and VosA in radial growth and stress tolerance. The disruption mutants (DM) and complement mutants (CM) of wetA and vosA generated in the WT background ( Fig. S2 and Table S1) (detailed in Methods) were incubated 7 days on different media at the optimal regime after initiation of radial growth with ;10 3 conidia (in 1 mL of a 10 6 conidia/mL suspension). Unlike fluffy colony morphology in B. bassiana (50), the DM strains displayed invisible defects in radial growth on the rich media PDA, Sabouraud dextrose agar plus yeast extract (SDAY) and ¼ SDAY (amended with ¼ of each SDAY nutrient) or the minimal Czapek-Dox agar (CDA) (Fig. 2A). The diameters of almost all 7-day-old colonies differed insignificantly between the DM strains and their control (WT and CM) strains (P . 0.05 in Tukey's test) on the rich media or the minimal CDA and CDAs amended with different carbon or inorganic nitrogen sources (Fig. 2B) or with organic nitrogen sources of different amino acids (Fig. 2C). The wetA DM showed significant, but moderate, growth defect (P , 0.05 in Tukey's test) only on the carbon sources of glucose, fructose, and maltose.
Stress assays on CDA supplemented with chemicals were conducted to reveal whether wetA and vosA were involved in some stress responses of M. robertsii, as seen in B. bassiana Relative transcript (RT) levels of wetA and vosA in the M r WT strain during a 7-day incubation on PDA at the optimal regime of 25°C and 12:12 (L:D) with respect to a standard on day 2. Error bars: standard deviations (SDs) of the means from 3 independent cDNA samples analyzed via qPCR. (C) Laser scanning confocal microscopic (LSCM) images (scale: 5 mm) for subcellular localization of WetA-GFP and VosA-GFP fusion proteins expressed in the hyphae of M r WT. Arrows indicate nuclei. Rows 1 to 4 are bright, expressed, DAPI-stained, and merged views of the same microscopic field, respectively. (50). Only the DwetA mutant was significantly more sensitive than the control strains to oxidative stress induced by H 2 O 2 (2 mM) or menadione (0.02 mM), and cell wall stress induced by Congo red (1 mg/mL), while the tested strains were equally responsive to either the other cell wall antagonist calcofluor white (50 mg/mL) or osmotic stress induced by NaCl (0.8 M) or sorbitol (1 M) (Fig. 2D). In DwetA, increased sensitivity to oxidative stress correlated with reduced activities of superoxide dismutases (SOD) and catalases (Fig. 2E), which scavenge intracellular superoxide anions and H 2 O 2 , respectively (51). Interestingly, the DvosA mutant without antioxidant response showed increased SOD activity and decreased catalase activity. Additionally, impaired cell wall integrity, suggested by increased DwetA sensitivity to Congo red, was further shown with significantly more release of protoplasts from hyphal cells of DwetA than of the remaining strains after 3 or 6 h of cell wall lysing (Fig. 2F). These data demonstrated significant role of wetA in M. robertsii's responses to 2 oxidants and cell wall perturbing Congo red, despite its dispensability for radial growth under normal culture conditions and response to osmotic agents, and the other cell wall antagonist. There was no role for vosA in all tested phenotypes except for the counteracting effect on SOD and catalase activities.
Essential versus null role of WetA versus VosA in asexual development and conidial quality control. Conidial yields were assessed daily from the cultures of all tested strains during a 15-day incubation at the optimal regime after 100 mL suspension (10 7 conidia/mL) was spread per plate of 1/4 SDAY for culture initiation. The DwetA mutant showed a 1 -day delay in conidiation, and a conidial yield was reduced by 72% to 92% in the 4-to 15-day-old cultures, whereas little change in conidial yield was observed in DvosA (Fig. 3A). Biomass levels assessed from the 7-day-old plate cultures showed no variability (P = 0.49) among the DM and control strains (Fig. 3B), suggesting no link of the DwetA mutant's conidiation defect to hyphal growth. Submerged blastospore production in 3-day-old SDBY (agar-free SDAY) cultures was reduced by 90% in DwetA but not affected in DvosA relative to the control strains (Fig. 3C), accompanied by an insignificant biomass variation (P = 0.18) in the same cultures of all tested strains (Fig. 3D). These data indicated an essentiality of wetA, but a dispensability of vosA, for aerial conidiation and submerged blastospore production.
As an index of conidial viability, median germination time (GT 50 ) at 25°C was significantly prolonged by 28% in DwetA versus WT (Fig. 3E), accompanied by a 60% reduction in median lethal dose (LD 50 ) of UVB irradiation (Fig. 3F). Conidial hydrophobicity required for adhesion to insect cuticle (14) was largely lowered by 44% in DwetA (Fig. 3G). However, none of the properties were significantly affected in DvosA compared to the control strains, suggesting no role for vosA in conidial quality control of M. robertsii.
Conidial hydrophobicity reduced in DwetA was further revealed with scanning electron microscopic (SEM) images of conidial surfaces. Hydrophobin rodlet bundles were well defined and intact on the conidial surfaces of the control strains, but impaired or deformed on the conidial surfaces of DwetA (Fig. 4A), indicating, again, an impairment  (Table S2). As a result, hyd2 and hyd3 were sharply downregulated in the 5-and 6-day-old cultures of DwetA versus the control strains ( Fig. 4B). In addition, half of 19 other genes involved in cell wall composition and integrity were also downregulated in the mutant's 5-dayold cultures (Fig. 4C). The repressed genes encode the MAPK Slt2 and the MAPK kinase Mkk1 required for the regulation of cell wall integrity, 2 chitin synthases (chs4 and chs9), 2 GPI-anchored cell wall proteins (but2 and ecm33), b-1,3-glucan synthase catalytic submit (fks1), b-1,6-glucanase precursor (bglC), and Concanavalin A-like lectin glucanase (lamG). These data suggest a role for wetA in transcriptional mediation of those genes involved in cell wall composition and integrity to promote conidial maturation in M. robertsii. Marked roles of WetA and VosA in intracellular mannitol accumulation. High performance liquid chromatography (HPLC) was performed to assess contents of trehalose, mannitol, and glycerol in the extracts isolated from the 7-and 15-day-old cultures of all DM and control strains. Mannitol was consistently detected in all samples of the extracts isolated from the young (Fig. 5A) or old cultures (Fig. 5B). However, neither trehalose nor glycerol was detectable from the samples of any strains compared to the standard curve of either. The mannitol contents of DwetA and DvosA were reduced respectively by 77% and 63% in the young samples, and 80% and 58% in the old samples in comparison to the control strains' counterparts (Fig. 5C). These data indicate greater role of wetA than of vosA in mannitol accumulation. The similar mannitol contents observed in the young and old cultures of either mutant implicated little role for either WetA-or VosA-reliant mannitol accumulation in conidial maturation. Essential versus null role of WetA versus VosA in fungal infection and virulence. The virulence of each strain against Galleria mellonella larvae (4th instar) was assayed by topical application (immersion) of a 10 7 conidia/mL suspension for normal cuticle infection (NCI) or intrahemocoel injection of ;500 conidia (in 5 mL of a 10 5 conidia/ mL suspension) per larva for cuticle-bypassing infection (CBI). The means (6SD) of median lethal time (LT 50 ) estimated by modeling analyses of time-survival trends (Fig. 6A) were 10.9 (60.6) and 3.1 (60.2) days (n = 9) for 3 control strains against the model insect via NCI and CBI, respectively (Fig. 6B). The DvosA mutant's LT 50 s, namely, 11.9 (60.3) and 3.4 (60.2) days via NCI and CBI, differed insignificantly (P . 0.05 in Tukey's test) from the control strains' estimates. In contrast, the mean LT 50 of DwetA was prolonged to 21.9 (60.9) days via NCI and to 3.8 (60.1) days via CBI. Apparently, the DwetA virulence greatly attenuated via NCI indicated an essential role for WetA in the fungal NCI.
NCI starts from conidial adherence to insect surface, followed by conidial germination and hyphal invasion into insect hemocoel under the actions of cuticle-degrading enzymes, including extracellular enzymes (ECEs), with proteolytic, chitinolytic, and lipolytic activities and subtilisin-like Pr1 family proteases (52,53). For better insight into essential role of WetA in NCI, conidial adherence was assayed on locust hind wings, and showed a 54% reduction in DwetA compared to the WT strain (Fig. 6C). Next, total activities of ECEs and Pr1 proteases were quantified from the supernatants of submerged cultures in CDB (agar-free CDA) amended with 3% bovine serum albumin (BSA) as sole nitrogen source for enzyme induction. After the 3-day shaking incubation of a 10 4 conidia/mL suspension in CDB-BSA, the ECEs and Pr1 activities were lowered respectively by 67% and 77% in the supernatants of the DwetA versus WT cultures, although the mutant's biomass was twice of the WT biomass in the cultures (Fig. 6D). Both enzyme activities and biomass levels differed insignificantly between the control strains. The reduced conidial adherence and the decreased ECEs/Pr1 activities accorded well with the DwetA mutant's virulence largely attenuated via NCI, reinforcing a requirement of WetA for successful NCI and insect pathogenicity of M. robertsii.
Intriguingly, none of those genes analyzed via qPCR were found in the list of DEGs. Neither did the list include any known genes required for asexual development or involved in cell wall composition and hydrophobicity, providing little clue to the DwetA mutant's severe defects in aerial conidiation, conidial adhesion, and hydrophobicity. Three DEGs (2:1) enriched to the GO term conidiophore development virtually encode glucose-methanolcholine oxidoreductase (MAA_07744), C6 zinc finger protein (MAA_07622), and hypothetical protein (MAA_08878), respectively. Four DEGs (3:1) in the GO term obsolete pathogenesis encode fatty acid synthase subunit alpha reductase (MAA_01642), cell surface protein Mas1 (MAA_08289), iron permease FTR1 (MAA_08848), and hypothetical protein (MAA_01659).  Surprisingly, the 3 transcriptomes comprised 64 co-dysregulated genes, including 44 codownregulated, 8 co-upregulated, and 12 differentially regulated ( Fig. 6D and Table S4). Despite few functionally known in M. robertsii, the co-downregulated genes could be causative of the DwetA mutant's severe defects in aerial conidiation and submerged blastospore production since both phenotypes were abolished in the absence of brlA or abaA (34). The 2 subtilisin-like protease genes, aforementioned, were also drastically downregulated in the previous DbrlA and DabaA mutants incapable of hyphal invasion into insect body, suggesting their possible roles in blocking the fungal NCI in the absence of each CDP gene.

DISCUSSION
Our data confirm the essential role of wetA, but no role of vosA, in asexual development, conidial maturation, insect pathogenicity, and cell tolerance to oxidative and cell wall stresses in M. robertsii. Previously, the DwetA and DvosA mutants of B. bassiana Counts of genes upregulated (log 2 ratio $ 1) and downregulated (log 2 ratio # 21) at a significance of q , 0.05. The transcriptome was based on three 4-day-old PDA cultures (replicates) of DwetA and WT grown at 25°C and 12:12 (L:D). (B and C) Counts of differentially regulated genes significantly enriched (P , 0.05) to GO terms and KEGG pathways (map00071, fatty acid degradation; map00053, ascorbate and aldarate metabolism; map00780, biotin metabolism; map00350, tyrosine metabolism), respectively. (D) Distribution of log 2 ratios for all genes (detailed in Table S4) co-upregulated, co-downregulated, and differentially regulated (marked by dashed lines) in the DbrlA, DabaA, and DwetA mutants of M. robertsii.

Characterization of WetA and VosA in M. robertsii
Microbiology Spectrum showed more cottony and thicker colonies, greatly (98% and 88%) reduced conidial yields, similarly increased sensitivity to NaCl-induced osmotic stress, and much more attenuated virulence via CBI than NCI, aside from moderate decreases of DwetA in conidial hydrophobicity (only ;7%), and tolerance to cell wall stress and increased sensitivity of DvosA to menadione (50). Apparently, the wetA orthologs in B. bassiana and M. robertsii play similar roles in asexual development but differential roles in conidial quality control and NCI. Their contributions to aerial conidiation are much greater than those reported with a 35% decrease in conidial yield of F. graminearum DwetA (46), and little effect on Penicillium DwetA mutants' conidial yields (47,48). In M. robertsii, WetA was more involved in hydrophobin assembly onto conidial surfaces, which determine conidial hydrophobicity and adhesion essential for initial NCI (14), and also in the secretion of cuticle-degrading enzymes crucial for successful NCI (52,53). Thus, the DwetA mutant's virulence was largely attenuated via NCI. Likewise, the reduced catalase and SOD activities were ascribed to the mutant's higher sensitivity to oxidative stress. For the DwetA mutant, impaired cell wall integrity was also shown by increased sensitivity to Congo red, easier cell wall lysing, and repressed expressions of hyd2, hyd3, and 10 other genes required for, or involved in, cell wall integrity. These observations suggest a pivotal role for WetA in mediating the cell wall composition and maturation of conidia in M. robertsii, and are close to the previous observations in the DwetA mutants of Penicillium spp. (47,48) and F. graminearum (46). Notably, NCI and virulence were much more compromised in the DwetA mutant of M. robertsii than of B. bassiana (50). This is very different from unaffected virulence of phytopathogenic fungal DwetA mutants (46)(47)(48). The previous and present studies demonstrate that WetA othologs play relatively conserved roles in cell wall composition and maturation of conidia, but are functionally differential or even different in the lifecycles in vitro and in vivo of different insect and plant mycopathogens. Unexpectedly, deletion of vosA resulted in little change in all examined phenotypes, indicating its dispensability for the asexual cycle and host infection of M. robertsii. This is very different from a significant role of its B. bassiana ortholog in conidiation capacity, conidial quality control, stress tolerance, and virulence (50), presenting 1 more big difference between the 2 insect pathogens. In DvosA, increased SOD activity could be counteracted by decreased catalase activity, leading to its null response to either H 2 O 2 or menadione, a compound generating superoxide anions degraded by SOD into water, and H 2 O 2 to be further decomposed by catalases (51).
Aside from spore surface structure, intracellular polyol accumulation is also a cellular event involved in conidial maturation and stress tolerance (38)(39)(40). In A. nidulans, the VelB-VosA complex was shown to mediate trehalose biogenesis and spore wall integrity (43,44). In B. bassiana, single-gene deletions of mannitol-1-phosphate dehydrogenase (MPD) and mannitol dehydrogenase (MTD) resulted in reduced mannitol versus increased trehalose accumulation (56), while reduced trehalose versus increased mannitol accumulation occurred when trehalose synthesis was partially or completely inhibited by knockout mutation of the trehalose-6-phosphate synthase paralog TpsA, TpsB or both (57), suggesting reversed accumulation trends of the 2 small-molecule metabolites. Previously, trehalose contents in fresh and mature conidia from the 7and 15-day-old cultures of B. bassiana DwetA and DvosA mutants grown at the optimal regime were differentially increased and decreased, respectively (50). In M. robertsii, trehalose content was undetectable in the extracts isolated from the young and old cultures of all DM and control strains. Instead, mannitol contents of either DwetA or DvosA decreased similarly in the young and old cultures. For the DvosA mutant, however, a decrease of ;60% in mannitol content was not influential on any examined phenotypes. The null effect of so reduced mannitol accumulation on biological aspects of M. robertsii agrees with the little role of mannitol as a conidial reserve carbon source in Aspergillus niger (39), and also with an inference that the metabolism of mannitol may not exist as a cycle to support its roles speculated in earlier studies (40).
Also unexpectedly, none of significantly downregulated genes in the qPCR analysis appeared in the list of identified DEGs, indicating a discrepancy between the 2 methods used in transcriptional profiling. The discrpepancy might have arisen from a 1-or 2-day difference of culture ages and more strict standards for identification of DEGs from the transcriptome. Notably, only 160 DEGs were identified in the present DwetA transcriptome. This count is very small in comparison to 5,725 DEGs (3,076:2,649) identified in A. nidulans DwetA (45), and 2,018 and 1,053 DEGs (881:1,137 and 429:624) identified from the respective DwetA cultures of F. graminearum at 6 and 12 h after conidiophore induction (46). The present and previous analyses reveal a large variation in the gene expression networks regulated by wetA orthologs in different ascomycetes. In M. robertsii, 3 functionally unknown genes enriched to the GO term conidiophore development were helpless to interpret the DwetA mutant's severe defect in asexual development. This is different from 319 conidiation-related DEGs identified from the F. graminearum DwetA less impaired in conidiation capacity (46). More small GO terms enriched provided very limited clues to deeper insight into the DwetA mutant's defects in conidial viability, hydrophobicity, adhesion, infectivity, and stress tolerance. Instead, the majority of downregulated genes in the main GO terms hint at that blocked asexual development and impaired conidial quality in DwetA could have resulted from their comprehensive roles in cellular component, biological process and molecular function. Additionally, 44 genes were codownregulated in the DbrlA, DabaA and DwetA mutants of M. robertsii. Since both aerial conidiation and submerged blastospore production were abolished in DbrlA or DabaA (34), those co-downregulated genes, taking 27.5% of all DEGs identified from DwetA, could be important targets of WetA in M. robertsii. However, almost all of them remain functionally unknown like other 48 DEGs encoding hypothetical proteins in Metarhizium, warranting more studies.
Conclusively, WetA plays a relatively conserved role in regulating aerial conidiation and conidial maturation in M. robertsii as seen in B. bassiana. The process of conidial maturation regulated by WetA is linked mainly to hydrophobin assembly onto conidial coat determinant to conidial hydrophobicity and adhesion required for initial NCI, but not related to intracellular polyol accumulation. WetA is also involved in the secretion of cuticle-degrading enzymes essential for successful NCI, and, hence, plays an important role in the insect-pathogenic life cycle of M. robertsii. In contrast, VosA was proven to be functionally redundant in M. roberstii. These findings provide a scenario of WetA or VosA that is functionally differential or different from those learned in B. bassiana and other ascomycetes.

MATERIALS AND METHODS
Recognition and domain analysis of WetA and VosA in M. robertsii. The amino acid sequences of WetA and VosA in A. nidulans, and of their orthologs previously characterized in B. bassiana (50), were used as queries to search through the genomic databases of M. robertsii (4) and other entomopathogenic and nonentomopathogenic ascomycetes at http://blast.ncbi.nlm.nih.gov/blast.cgi. Conserved domains were predicted from the amino acid sequences of the used queries and their orthologs located in M. robertsii at http://smart.embl-heidelberg.de, followed by predicting an NLS motif from each sequence with a maximal probability at https://www.novopro.cn/tools/nls-signal-prediction. All WetA and VosA orthologs found in the fungal genomes examined were clustered via phylogenetic analysis with the maximum likelihood method in the online program MEGA11 (http://www.megasoftware.net/).
Transcriptional profiling and subcellular localization of WetA and VosA. The qPCR analysis with paired primers (Tables S1) was carried out to assess daily expression levels of wetA and vosA in the PDA cultures of the WT strain during a 7-day incubation at the optimal regime of 25°C and 12:12 (L:D) (detailed later). For subcellular localization, the open reading frame of wetA or vosA was amplified from the WT cDNA and ligated to 59-terminus of gfp at appropriate enzyme sites of the backbone plasmid pAN52-C-gfp-bar, in which C denotes a cassette (59-EcoRI-XmaI-BamHI-PstI-HindIII-39) driven by the endogenous promoter Ptef1 (34). The resultant vector was transformed into the WT strain via Agrobacterium-mediated transformation, followed by screening transgenic colonies by means of the bar resistance to phosphinothricin (200 mg/mL). A strong fluorescence colony selected from each transformation was incubated on PDA for conidiation at the optimal regime. The resultant conidia were incubated in SDBY (4% glucose, 1% peptone) and 1% yeast extract) for 3 days at 25°C. Hyphae from the liquid culture were washed in sterile water, stained with the nuclear dye 49,69-diamidine-29-phenylindole dihydrochloride ([DAPI]; Sigma-Aldrich) of 4.16 mM, and visualized under a laser scanning confocal microscope at the excitation/emission wavelengths of 358/460 and 488/ 507 nm to determine subcellular localization of WetA-GFP or VosA-GFP fusion protein.
Construction of wetA and vosA mutants. The DM and CM strains of either wetA or vosA were constructed respectively by homologous recombination in the WT strain of p0380-59x-bar-39x (x: wetA or vosA) vectoring its bar-separated 59 and 39 coding/flanking fragments (Fig. S1A and B) and ectopic Characterization of WetA and VosA in M. robertsii Microbiology Spectrum integration into an identified Dx mutant of p0380-sur-x vectoring its full-length coding/flanking sequences by means of the mentioned method. Putative DM and CM colonies were screened respectively by the bar resistance to phosphinothricin (200 mg/mL) and the sur resistance to chlorimuron ethyl (10 mg/ mL), identified via PCR analysis ( Fig. S1C and D), and verified via qPCR analysis ( Fig. S1E and F). The deletion vector of either target gene was constructed by amplifying its 59 and 39 fragments from the WT DNA and inserting into appropriate enzyme sites (BamHI/PstI and XhoI/XmaI) of linearized p0380-bar. The complement vector of each gene was generated by amplifying its full-length coding sequence and flank regions from the WT DNA and ligated into linearized p0380-sur-gateway to substitute the gateway fragment under the action of Gateway BP Clonase II Enzyme Mix (Invitrogen). Paired primers used for vector construction and detection of each target gene are listed in Table S1. The identified DM and CM strains were tested in parallel with the WT strain in the experiments, including 3 independent replicates, as follows.
Assays for radial growth rate, stress tolerance, and spore yield and quality. Radial growth of each fungal strain was initiated by spotting 1 mL aliquots of a 10 6 conidia/mL suspension on the plates of PDA, SDAY, 1/4 SDAY, CDA (3% sucrose, 0.3% NaNO 3 , 0.1% K 2 HPO 4 , 0.05% KCl, 0.05% MgSO 4 , 0.001% FeSO 4 , and 1.5% agar), and CDAs amended with different carbon or inorganic/organic nitrogen sources (see Fig. 2 for details). After a 7-day incubation at the optimal regime, the diameter of each colony was measured perpendicularly to each other across the center.
To assess stress tolerance, radial growth was initiated as above on CDA alone (control) or supplemented with NaCl (0.8 M), sorbitol (1 M), H 2 O 2 (2 mM), menadione (0.02 mM), Congo red (1 mg/mL), and calcofluor white (50 mg/mL), respectively, followed by a 7-day incubation at the optimal regime and estimation of each colony diameter as above. Relative growth inhibition [(d c 2d s )/d c Â100] of each strain by each chemical stress was computed using the diameters of stressed colonies (d s ) and control colonies (d c ).
The cultures for assessment of conidiation capacity were initiated by spreading 100 mL aliquots of a 10 7 conidia/mL suspension on cellophane-overlaid plates (f = 9 cm) of 1/4 SDAY (a medium suitable for the fungal conidiation) and incubated for 15 days at the optimal regime. From day 3 onwards, a cork borer (f = 5 mm) was used to take plugs from each plate culture. Conidial yield in each plug was measured as the number of conidia per unit area (cm 2 ) of plate culture as described previously (16). Biomass level was also assessed from the 7-day-old plate cultures of each strain dried at 70°C for 6 h after collection. Moreover, 50 mL aliquots of a 10 6 conidia/mL suspension in SDBY was incubated on a shaking bed (150 rpm) for 3 days at 25°C, followed by assessment of blastospore concentration from each culture. Biomass level of each SDBY culture was assessed after the culture was collected via filtration and dried overnight at 70°C.
The quality properties of conidia collected from the 15-day-old cultures were assessed as indices of viability [GT 50 (h) at 25°C], hydrophobicity in an aqueous-organic system, and LD 50 (J/cm 2 ) for resistance to UVB irradiation (weighted wavelength: 312 nm), as described previously (16,17). SEM images of conidial surfaces were collected to show microstructures of hydrophobin rodlet bundles.
For deeper insight into antioxidant activity of each strain, moreover, total catalase and SOD activities (U/mg) were quantified from the protein extracts of the 3-day-old PDA cultures using Catalase Activity Assays Kit (Jiancheng Biotech) and SOD Activity assay kit (Sigma-Aldrich), following the manufacturers' guides. The impaired cell wall integrity of a given DM strain was further revealed in a cell wall lysing experiment. Briefly, the DM and control strains' cell samples (0.1 g) collected from the 3-day-old SDBY cultures were rinsed in sterile water, resuspended in 2 mL aliquots of 1.0 M NaCl containing snailase and lysing enzyme (Sigma-Aldrich) of 10 mg/mL and incubated by shaking at 37°C and 150 rpm for 3 and 6 h of cell wall lysing, followed by assessing the concentration of protoplasts released from each sample with a hemocytometer.
Assays for contents of intracellular polyols. Fresh samples (conidia and hyphae) of 1 g were taken from the 7-and 15-day-old cultures grown on 1/4 SDAY at the optimal regime, homogenized in liquid nitrogen, and suspended in 1 mL dd-H 2 O. The suspensions were boiled in a water bath for 6 h and centrifuged for 30 min at 1.6 Â 10 4 g. Intracellular trehalose, mannitol, and glycerol contents (mg/g dry mass) in each of the supernatants were determined with respect to the respective readings of standard trehalose, mannitol, and glycerol (Sigma-Aldrich) in a HPLC system described previously (58).
Assays for fungal virulence, conidial adhesion to insect cuticle, and total activities of EXEs and Pr1 proteases. Standardized bioassays for NCI and CBI were initiated by immersing 3 groups of 30 to 40 G. mellonella larvae per strain for 10 s in 40 mL aliquots of a 10 7 conidia/mL suspension, and injecting 5 mL of a 10 5 conidia/mL suspension into the hemocoel of each larva in each group, respectively. All grouped larvae were held at 25°C after inoculation and monitored for survival/mortality records at a 12 h interval until the mortality was stabilized. LT 50 estimates were made by modeling analysis of the timemortality trend in each group.
Conidial adhesion of each mutant required for initiation of NCI was assayed on locust (Locusta migratoria manilensis) hind wings with respect to the WT strain, as described previously (14,17). Our previous protocols (59) were adopted to quantify total activities (U/mL) of cuticle-degrading ECEs and Pr1 family proteases from the supernatants of 3-day-old CDB-BSA cultures, which were prepared by shaking three 50 mL aliquots of a 10 4 conidia/mL CDB-BSA per strain at 25°C and 150 rpm. Biomass level in each culture was also measured, as aforementioned.
Transcriptional analysis. Three independent PDA cultures (replicates) of each strain were initiated by spreading 100 mL of a 10 7 conidia/mL suspension per cellophane-overlaid plate and incubated for up to 7 days at the optimal regime. Total RNAs were extracted daily from the 2-to 7-day-old cultures of the WT strain or from the 4-to 6-day-old cultures of the DM and control strains using an RNAiso Plus Kit (TaKaRa), and reversely transcribed into cDNAs using a PrimeScript RT reagent kit (TaKaRa), respectively. The qPCR analysis was performed under the action of SYBR Premix Ex Taq (TaKaRa) to assess: (i) transcript levels of wetA and vosA from the daily WT cDNA samples with paired primers in Table S1; (ii) transcript levels of wetA or vosA in the cDNA samples derived from the 4-day-old cultures of each DM and control strains for verification of expected recombination events; and (iii) transcript levels of 3 hyd genes and 19 cell wall integrity-related genes in the cDNA samples derived from the 4-, 5-or 6-day-old cultures of DwetA and its control strains with paired primers in Table S2. Normalized with the fungal 18S rRNA, the 2 2DDCT method was adopted to compute relative transcript levels of wetA and vosA in the WT cultures over the days of incubation with respect to a standard on day 2 and of other genes in the mutant strains with respect to the WT standard. A total of 1-fold transcript change was used as a standard of significant down-or upregulation.
Transcriptomic analysis. The wetA-specific transcriptome was generated in Lianchuan BioTech Co. (Hangzhou, China) based on the 4-day-old PDA cultures of the DwetA and WT strains (3 cultures per strain) grown at the optimal regime, as described previously (34). Briefly, clean tags from RNA-seq data set were mapped to M. robertsii genome (4). DEGs were identified at the significant levels of both log 2 ratio (fold change) # 21 or $ 1 and q , 0.05, and enriched to GO terms of 3 function categories (P , 0.05) at http://www.geneontology.org and KEGG pathways (P , 0.05) at http://www .genome.jp/kegg, respectively.
Statistical analysis. All data collected from the experiments of 3 independent replicates were subjected to one-way analysis of variance and Tukey's honestly significant difference (HSD) test to differentiate statistical differences of phenotypic parameters between the tested DM and control strains.
Data availability. All data generated or analyzed during this study are included in the paper and associated supplemental files. All RNA-seq data of this study are available at the NCBI's Gene Expression Omnibus under the accession PRJNA905185.