Glyoxal oxidase-mediated detoxification of reactive carbonyl species contributes to virulence, stress tolerance, and development in a pathogenic fungus

Reactive carbonyl and oxygen species (RCS/ROS), often generated as metabolic byproducts, particularly under conditions of pathology, can cause direct damage to proteins, lipids, and nucleic acids. Glyoxal oxidases (Gloxs) oxidize aldehydes to carboxylic acids, generating hydrogen peroxide (H2O2). Although best characterized for their roles in lignin degradation, Glox in plant fungal pathogens are known to contribute to virulence, however, the mechanism underlying such effects are unclear. Here, we show that Glox in the insect pathogenic fungus, Metarhizium acridum, is highly expressed in mycelia and during formation of infection structures (appressoria), with the enzyme localizing to the cell membrane. MaGlox targeted gene disruption mutants showed RCS and ROS accumulation, resulting in cell toxicity, induction of apoptosis and increased autophagy, inhibiting normal fungal growth and development. The ability of the MaGlox mutant to scavenge RCS was significantly reduced, and the mutant exhibited increased susceptibility to aldehydes, oxidative and cell wall perturbing agents but not toward osmotic stress, with altered cell wall contents. The ΔMaGlox mutant was impaired in its ability to penetrate the host cuticle and evade host immune defense resulting in attenuated pathogenicity. Overexpression of MaGlox promoted fungal growth and conidial germination, increased tolerance to H2O2, but had little to other phenotypic effects. Transcriptomic analyses revealed downregulation of genes related to cell wall synthesis, conidiation, stress tolerance, and host cuticle penetration in the ΔMaGlox mutant. These findings demonstrate that MaGlox-mediated scavenging of RCS is required for virulence, and contributes to normal fungal growth and development, stress resistance.


Introduction
Glyoxal oxidases (Glox) are copper-containing extracellular or membrane bound enzymes with broad substrate specificities capable of oxidizing aldehydes into carboxylic acids coupled to the conversion of oxygen to the production of hydrogen peroxide (H 2 O 2 ) [1,2].Glox belongs to the AA5_1 subfamily, composed of Glox and glyoxal-like enzymes termed radical copper oxidases, and is related in structure and reactive chemistry to galactose oxidases, although the latter are considered as a separate subfamily (AA5_2) [3][4][5].Gloxs contain a glyoxal oxidase domain and from zero to five cell wall integrity and stress component domains (WSC domains) [6,7].Glox was first characterized in the cellulose-degrading (white rot) fungus Phanerodontia chrysosporium [8] where it is a critical component involved in the breakdown of lignin [1,7,9,10].The enzyme converts aldehydes produced from the pretreatment of lignocellulose, e.g., furaldehyde and 5-(hydroxymethyl)-2-furaldehyde (HMF), producing H 2 O 2 [11], and the biochemical (and hence biological) roles of Gloxs have mainly focused on their contributions to substrate degradation in lignocellulose-degrading fungi white-rot fungi and plant associated Trichoderma sp., as well as its potential biotechnological application in bioethanol production (in yeasts) [11][12][13].Several Glox enzymes have also been characterized in plant pathogenic fungi.Glox has been shown to contribute to virulence and mycotoxin production in the wheat and barley blight pathogen, Fusarium graminearum [14], and is involved in mycelial development and pathogenicity in the corn smut causing fungus, Ustilago maydis [15].Downregulation of Glox1 in Trichoderma virens, a soil dwelling saprophyte currently being used as a biological control agent against plant microbial pathogens such as Pythium and Rhizoctonia sp., resulted in delayed hyphal growth and reduced hydrophobicity of conidia, but had no effect on its biocontrol ability [13].Interestingly, Glox enzymes are also found in plants, and a Glox gene in grapevine (Vitus pseudoreticulata) appears to help defend the plant against the powdery mildew fungal pathogen, Erysiphe necator, by producing toxic H 2 O 2 [16].However, the functions of Glox genes in fungal animal or insect pathogens has yet to be examined.
Entomopathogenic fungi are increasingly being utilized as safe and environmentallyfriendly alternatives to chemical insecticides for pest control, as agents that can be compatible with a variety of other pest management strategies as well as "green" farming practices [17,18].The infection process of these fungi, e.g.those from the Metarhizium genera, towards host insects typically involves spore adhesion, appressorium formation, cuticle penetration, colonization in the hemocoel, and sporulation on the host cadavers [19,20].During infection, entomopathogenic fungi are challenged by various biotic and abiotic stresses, including oxidative and osmotic stresses, unfavorable temperature and moisture, poor nutrition, antagonistic microbes on the cuticle surface, and the host innate immunity [21,22].These factors can induce the generation of reactive oxygen species (ROS), which can damage macromolecules and promote autophagy and apoptosis [23,24].Endogenous factors, which encompass the accumulated intermediate products in metabolism involving specific oxidases, also have the potential to induce the production of ROS [24][25][26].Aldehydes, classified as carbonyl compounds, hold a significant status as reactive carbonyl species (RCS).They can cause extensive damage to cellular biomolecules and exhibit various biological effects, ranging from disrupting normal gene regulation to toxicity [27].As Glox enzymes mediate aldehyde oxidation we hypothesized that (in non-lignin degrading fungi), these enzymes may be primarily used for detoxification of RCS while potentially producing ROS that then contributes to targeting host defenses and hence is required for virulence and normal fungal growth.
Although many Metarhizium species are broad host range insect pathogens (e.g., M. robertsii), M. acridum has a narrower host specificity towards acridids, and has been commercialized for use in killing locusts and grasshoppers [28,29].Several unique adaptions have been reported for M. acridum including expression of a specialized extracellular catalase-peroxidase implicated in scavenging host antimicrobial oxidative stress defenses [22].In order to further probe fungal detoxification mechanisms that might contribute to pathogenicity and development, we investigated the function of the Glox gene in M. acridum via genetic and biochemical approaches.Our results indicate that MaGlox scavenges RCS by targeting aldehyde accumulation, influencing numerous downstream physiological and biological pathways including, ROS accumulation, fungal growth and development, apoptosis, stress tolerance, and pathogenicity.

MaGlox sequence analyses, cellular localization and gene expression pattern
A single Glox homolog, with a full-length nucleotide sequence of 3762 bp, containing three introns was identified in the M. acridum genome.The open reading frame of MaGlox was 3093 bp, encoding a protein of 1030 amino acids with an estimated molecular weight of 108.16 kDa and an isoelectric point of 4.97 (https://web.expasy.org/protparam/).Sequence analyses indicated that the protein contained eight transmembrane helices (at the C-terminus) as predicted by TMpred, an N-terminal signal peptide of 23 amino acids, four WSC domains at the N-terminus, a conserved glyoxal oxidase N-terminal domain (Glyoxal_oxid_N) (from Val-583 to Gly-776) and a C-terminal early "set" domain (E_set_GO_C) (from Gly-911 to Ile-1009, S1A Fig) .The Glyoxal_oxid_N and E_set_GO_C domains grouped together (referred to as the AA5_1 domain) and are conserved in all Glox proteins.Sequence and domain alignments confirmed the absence of WSC domains in plant Gloxs, and in several white-rot fungi and human pathogenic fungi, whereas three to five WSC domains were found in phytopathogenic fungi including Trichodema spp., Thermothelomyces thermophilus, and in all Ascomycete entomopathogenic fungi examined (S1B To examine subcellular localization of the protein, a MaGlox-eGFP expression fusion construct was transformed into wild type M. acridum as detailed in the Methods section (S2 Fig and S1 Table ).These data showed green fluorescence signals at the cell membrane, particularly

PLOS PATHOGENS
at the septa in hyphae, and in the membranes of conidia (Fig 1A).To provide a quantitative analysis of MaGlox expression, RT-qPCR analyses were performed with RNA extracted from various fungal cells.These data indicated the highest expression in growing hyphae/mycelia, which was ~5-13 times greater than that seen in conidia (Fig 1B).MaGlox expression during appressorium formation on locust wings was twice that seen in conidia, but MaGlox expression was low in in vivo-derived fungal cells (termed hyphal bodies) growing inside the host hemolymph after cuticle penetration (Fig 1B).Addition of exogenous methylglyoxal at low concentrations (0.15 mg/mL and 0.3 mg/mL) induced expression of MaGlox by 40-50% (P < 0.01), however, increased expression was not observed at a higher methylglyoxal concentration tested (0.6 mg/mL) (Fig 1C ).

MaGlox contributes to growth and conidiation
To investigate functional roles for MaGlox, we utilized homologous recombination to construct a targeted gene deletion mutant via replacement of a portion of the MaGlox gene with the bar gene resistance marker cassette as well as constructing a corresponding complementation strain (CP).We also constructed a constitutive MaGlox (over) expression strain (MaGlox-OE), in which MaGlox expression was driven by the glyceraldehyde-3-phosphate dehydrogenase gene promoter (PgpdM) (S2A Fig) .The integrity of the MaGlox disruption mutant, MaGlox-OE and complementation strains were confirmed by PCR and Southern blot analyses (S2B-S2E Fig) .As expected, expression of MaGlox was undetectable in the ΔMaGlox mutant and recovered in the CP strain.MaGlox expression was 7 times higher in the MaGlox-OE strain as compared to the wild type parent (S2F Fig).
The ΔMaGlox mutant showed a moderate growth inhibition phenotype (12.3-25.7%decrease, P < 0.05) when cultured on ¼-SDAY, with the colony color changed from normal yellow green to grey-green and conidial concentric rings more evident as compared to the wild type strain.MaGlox-OE strain had a similar phenotype as that of the wild type (Fig 2A).ΔMaGlox hyphae were irregular and curled as compared to longer and straighter hyphae seen for the wild type and MaGlox-OE strain (Fig 2B).The wild type strain initiated conidiation at ~18 h on ¼-SDAY, whereas the hyphae of ΔMaGlox only began to produced conidiophores at ~24 h (Fig 2C).Few conidia were apparent for the mutant strain even at 36 h, whereas extensive conidial production was evident for the control strains.MaGlox-OE strain also showed delayed conidiation as compared to wild type, with hyphae longer than wild type (Fig 2C).Measurement of the hyphal apical cell length revealed an ~70% reduction in the ΔMaGlox and ~25% increase in MaGlox-OE strain as compared to wild type and complemented controls (P < 0.01, Fig 2D).Quantification of conidial yield indicated a 70-90% decrease in conidial production for the ΔMaGlox strain over the entire time course examined (5-15 d on ¼-SDAY, P < 0.01, Fig 2E).However, no significant differences was found in conidial yield between the MaGlox-OE strain and wild type (Fig 2E).

Loss of MaGlox induces autophagy and cellular apoptosis
To examine whether the observed growth phenotypes were related to disruption in normal autophagy and/or cell apoptosis, these processes were examined at different time points in fungal cells grown in liquid cultural by M52 staining (i.e. the fluorescent probe perylene-3,4-dicarboxylic anhydride) and TUNEL assays, respectively, as detailed in the Methods section.Autophagosome staining of fungal cells grown for 14-h in media showed large (single) autophagosomes within each septal compartment of wild type and MaGlox-OE hyphae.However, the ΔMaGlox exhibited smaller, diffuse, yet still punctuate staining with an overall stronger fluorescence as compared to the wild type (Fig 2F).Similar results were observed at the 20 h growth time point (Fig 2F).TUNEL staining of 14 h cells revealed no differences between the three strains (Fig 2G).However, at 20 h, TUNEL staining was present in only a small fraction of growing hyphae in the wild type and MaGlox-OE strains, while extensive staining of nuclei was observed in the hyphae of the ΔMaGlox strain (Fig 2H).Additionally, live/dead staining using propidium iodide (PI) showed strong and extensive staining of nuclei in ΔMaGlox hyphae, which was not seen in the wild type and MaGlox-OE strains (Fig 2H).To determine whether the apparent increase in autophagosomes and subsequent apoptosis correlated with increased expression of autophagy and cell death genes, RT-qPCR analysis of the autophagyrelated genes (atg1, atg3, atg4, atg5, atg8, atg11, atg12, atg13, and atg17) and apoptosis-related genes (Cas, CasA1, CasA2) were performed on fungal cells grown in media.Results showed significantly (P < 0.01) increased expression of the atg genes (atg4, atg8, atg11 atg12, atg13, atg17 and atg101) and all the three apoptosis-related genes in the ΔMaGlox mutant as compared to the wild type (Fig 2I ).

MaGlox is essential for maintaining intracellular balance of ROS
Enzyme activity assays revealed that the ΔMaGlox strains exhibited an ~80% reduction in Glox activity compared to the wild type (P < 0.001) (Fig 3A).Additionally, tolerance to methylglyoxal and H 2 O 2, which are substrates and reaction products of the enzyme, respectively, were significantly decreased for the ΔMaGlox mutant (Fig 3B and 3C).However, similar to the wild type, the growth of the MaGlox-OE strain was inhibited at high concentrations of methylglyoxal, although MaGlox-OE colonies appeared slightly fluffier.In contrast, the MaGlox-OE strain demonstrated significantly increased tolerance to H 2 O 2 compared to the wild type (Fig 3B and 3C).Staining for cellular ROS levels with the peroxide indicator dihydroethidium (DHE) revealed strong fluorescence in the ΔMaGlox strain, whereas only a weak signal was observed in the hyphae of the wild type and MaGlox-OE strains for cells grown for 20 h, and fluorescence was strongly increased at the 36 h growth time point for all three strains (Fig 3D).Enzyme assays also revealed a significant reduction in total catalase (CAT) activity, coupled with increased superoxide dismutase (SOD) activity in the ΔMaGlox strain as compared to the wild type (P < 0.01, Fig 3E and 3F).Furthermore, alterations in the expression of genes related to oxidative stress tolerance were observed, including elevated levels of the expression of SOD and glutathione-S-transferase (GST) genes (P < 0.01) in the ΔMaGlox strain, while three out of four tested CAT genes were significantly downregulated (P < 0.01,

Loss of MaGlox results in accumulation of cellular RCS
The substrates for Glox are small chain aldehydes (e.g., glyoxal and methylgloxal), which represent an important class of RCS that are often highly toxic to cells and can lead to generation of ROS [30][31][32][33][34].To explore if loss of MaGlox affected intracellular levels of RCS, UHPLC-QTOF-MS was applied for measurement of fungal metabolites as detailed in the Methods section.A total of sixteen RCS species were detected and their concentrations were quantified (Fig 4A and S2 Table).Concentrations of seven RCS compounds: 2,3-butanedione, methylglyoxal, glyoxal, p-tolualdehyde, benzaldehyde, 3-hydroxybenzaldehyde, trans-2-pentenal, and trans, trans-2,4-hexadienal were significantly increased in the ΔMaGlox mutant as compared to wild type (P < 0.01).Of note among these, concentrations of methylglyoxal and glyoxal, the optimal substrates of Glox [9], were 6.7 and 10.3 times higher in the mutant than that seen for wild type, respectively.

PLOS PATHOGENS
RCS scavengers, which comprise thiol-, amino-and imidazole groups, are capable of scavenging RCS [33].To determine if the accumulation of RCS play a role in the growth inhibition of M. acridum, three types of RCS scavengers, 2-hydroxybenzylamine (2-HOBA), carnosine (CAR) and sodium 2-mercaptoethanesulfonate (MESNA), were used to test potential phenotypic rescue on fungal growth.The results showed that CAR (10 mmol/L) and MESNA (125 μg/mL) significantly recovered the colonial growth defects seen in the ΔMaGlox mutant compared to wild type.However, exogenous addition of 2-HOBA at 400 μg/mL did not rescue the growth phenotype of the ΔMaGlox mutant (Fig 4B and 4C), likely due to the observation that 2-HOBA has a relatively short half-life of 62 minutes, resulting in a relatively poor performance in phenotype recovery [35].
Conidial yield was also determined after RCS scavengers were included in the media.Results showed that CAR and MESNA supplementation in the media resulted in an ~130-150% increase in conidial yield for ΔMaGlox as compared to an ~30% increase for the wild type when CAR was added and no changes when MESNA was included (Fig 4D).In addition, cellular ROS levels were significantly decreased in the ΔMaGlox mutant when CAR and MESNA were added (P < 0.01, Fig 4E).Expression of autophagy-related (atg1, atg3, atg4, atg8, atg11, atg12, atg13, atg17) and apoptosis-related genes (Cas, CasA1, CasA2) were also significantly decreased when CAR and MESNA were added during growth of the ΔMaGlox mutant (S Fig 3A and 3B).

Contribution of MaGlox to fungal stress tolerance
In terms of growth in the presence of various cellular stress inducing agents, the ΔMaGlox strain was less tolerance to the membrane perturbing agent, calcofluor white (CFW), but showed increased resistance to detergent (SDS), and osmotic stress causing agents, sorbitol (SOR) and NaCl, as compared to the wild type parental strain.The MaGlox-OE strain also showed increased resistance to NaCl and SOR (Fig 5A and 5B).Except for the CFW phenotype, the complemented mutant phenotype was restored to wild type.Immunofluorescence and flow cytometry (FCM) were used to measure binding of (fluorescent) lectins to visualize cell surface carbohydrate epitopes (i.e, mannan/glucan and chitin).Results indicated no differences in ConA binding between any of the strains tested (Fig 5C ), however, binding of wheat germ agglutinin (WGA, chitin) revealed two apparent populations for the ΔMaGlox mutant, with one showing stronger and the other weaker relative fluorescence as compared to controls (Fig 5C).This observation was further confirmed by flow cytometry analyses, in which two close but distinct peaks for ΔMaGlox conidia could be discerned (Fig 5C).The ΔMaGlox strain also displayed improved tolerance to UV-B radiation with increased germination under these conditions (P < 0.01, Fig 5D ), and an overall significantly increased length of time of exposure to UV-B required to inhibit germination to 50% of untreated cells (IT 50 , P < 0.01, i.e., the mutant showed greater resistance to UV-B exposure of conidia in terms of germination as compared to the wild type and the complemented strains, Fig 5E).However, the ΔMaGlox strain had a significantly impaired tolerance to heat shock, showing a 66% decreased IT 50 compared to the control strain (P < 0.01, Fig 5F and 5G).MaGlox-OE strain showed no difference in tolerance to UV-B or heat stress as compared to the wild type strain (Fig 5D -5G).

Loss of MaGlox results in decreased virulence
To explore the effects of disruption of MaGlox on virulence in M. acridum: (1) topical inoculation and (2) intrahemocoel injection bioassays were conducted using 5 th instar locust nymphs as the host.In topical inoculation, the wild type and MaGlox-OE strains killed all locusts within 9.5 d, while similar mortality was seen on day 10.To assess whether effects seen with respect to virulence were influenced by germination and/or appressoria (infection structure) formation, these processes were monitored on dissected locust wings.Conidial germination for wild type (10%) and MaGlox-OE strain (33%) initiated 6 h post-inoculation onto the wings, whereas little to no germination had yet occurred for the ΔMaGlox strain (Fig 6D).At 9 hours post-inoculation, the germination level for the ΔMaGlox strain reached 27.7 ± 1.2%, whereas 74.3% of the wild type had already germinated (representing a 62.7% decrease, P < 0.01).Germination of MaGlox-OE strain showed no difference compared to wild type at 26 h post-inoculation (Fig 6D).Overall, the calculated GT 50 of the ΔMaGlox strain was 14.82 ± 0.16 h, which as significantly increased (P < 0.001) compared to the wild type strain (GT 50 = 7.90 ± 0.17 h, Fig 6E), while the GT 50 of MaGlox-OE strain (6.69 ± 0.14 h) was significantly shorter (P < 0.01) than that of wild type (Fig 6E).

Transcriptomic analysis
To explore the consequences on global gene expression resulting from loss of MaGlox, a comparative transcriptomic analysis was performed between wild type and the ΔMaGlox strain grown on ¼-SDAY as detailed in the Methods section (S4A Fig) .These data generated 1554 differentially expressed genes (DEGs), of which 669 showed higher expression and 885 showed lower expression in the ΔMaGlox mutant as compared to the wild type parent (S4B Fig) .The DEGs with per kilobase of transcript per million mapped reads (FPKM) > 10 are listed in S3 Table .Eighteen DEGs associated with growth, stress tolerance and virulence were analyzed by RT-qPCR to verify the transcriptome data.Results showed a high correlation coefficient (R = 0.966) between these two methods, suggesting that the RNA-Seq data are reliably robust (S4C Fig).
To identify virulence-related genes among DEGs (S4 Table ), a BLAST analysis was conducted against the pathogen-host interaction (PHI) gene database.There were 270 DEGs involved in virulence, with 38 genes assigned to the hypothetical protein category.Of the remaining 232 genes, 92 were upregulated and 140 downregulated in the mutant as compared to the wild type.Some cuticle degrading enzyme and other factors involved in penetration showed decreased expression, including: aspartyl protease (MAC_01391, MAC_05521) [41], glycosyltransferase (MAC_01796, MAC_01312) [42,43] and glycerol-3-phosphate dehydrogenase (MAC_07043) [44].Genes showing higher expression levels included two Zn(II)2Cys6

Discussion
Glyoxal oxidases were initially characterized as fungal enzymes in white rot involved in lignin/ plant biomass turnover [1].Additional functions in some phytopathogenic fungi, e.g., U. maydis and F. graminearum, suggest a role in targeting plant tissues (13,14).However, as these enzymes exist in most fungi, many of whom are not directly involved in lignin degradation, it seems likely that they have broader functions in detoxification of reactive carbonyl species that could be present in the environment and/or generated endogenously as metabolic by products.Our analysis of Glox genes shows a clear divide in terms of structural motifs, with the Glox of plant white rot, and mammalian fungal pathogens lacking WSC domains, whereas the genes found in fungal plant and invertebrate pathogens contain from 3-5 such domains.As these domains are known to be involved in binding to chitin/glucans, the functional roles of these two groups of Gloxs are likely to have diverged.Indeed, whereas the white rot Glox is an extracellular enzyme, the M. acridum Glox is localized to the outer membrane and septa (the latter particularly rich in chitin) likely via the WSC domains.
In terms of function, our data show that the MaGlox contributes to detoxification of intracellular RCS and expression of MaGlox is induced by the addition of exogenous aldehyde substrates within a specific concentration range.Although the product of Glox is H 2 O 2 , loss of Glox still resulted in increased cellular H 2 O 2 levels, likely due to downstream generation as a result of accumulation of toxic RCS, which are markers of oxidative stress [49].Thus, despite generating H 2 O 2 , in terms of wider cellular metabolism, Glox contributes to decreased levels of both intracellular RCS and ROS, the former via direct enzymatic activity, and the latter as an indirect result, including by affecting (decreasing expression of) catalase activity.ROS are generated through various enzymatic and non-enzymatic processes, including the activities of NADPH oxidases (NOXs), mitochondria, the endoplasmic reticulum, peroxisomes, and external stimuli [50].Notably, the DEGs involved in maintaining mitochondrial function (11 genes), ribosome biogenesis (28 genes), DNA repair (8 genes), and translation (6 genes) all showed decreased expression in the MaGlox mutant (see S3 Table ).Additionally, genes associated with endoplasmic reticulum stress, signal transduction, protein/ion transport, and DNA, RNA, and ATP metabolism also showed decreased expression.Decreased expression of these genes likely contributes to the induction of cell death pathways, e.g., apoptosis, autophagy, and necrosis, that subsequently negatively affects cell growth, conidiation, resistance.In addition, activation of these pathways impairs the ability of the fungus to infect hosts, resulting in decreased virulence.Interestingly, our data indicate that MaGlox mutant cells try to compensate for the increased levels of RCS/ROS by increasing expression of potential scavenging enzymes including glyoxalase I, glutathione-S-transferase, superoxide dismutase, and cytochrome P450 genes, however, these are apparently unable to limit the increased levels of toxic RCS/ROS.As the mutant strain grows and metabolizes, the absence of MaGlox leads to the accumulation of RCS, which induces autophagy.With the accumulated cellular damage triggering apoptosis.The consequences of the increased RCS/ROS appear to be the induction of a stress responses including autophagosome proliferation, followed by the induction of cellular death pathways.RCS scavengers rescued the impaired growth and conidiation of mutant strain, down-regulated the expression of autophagy and apoptosis-related genes and decreased ROS levels in MaGlox mutant strain, supporting the model for MaGlox functioning.The induction of increased autophagy and apoptosis can account for the ΔMaGlox phenotypes of reduced vegetative growth, distorted hyphal morphology and elongation, delayed germination, and decreased conidial yield.Some of these phenotypes are consistent with Glox mutants of T. virens and U. maydis, which also showed impaired hyphal development [13,15], but differ from findings in Fusarium spp.[14].Genes related to fungal growth and development (MAC_03443, MAC_05865, and MAC_03916) were significantly downregulated in ΔMaGlox.Of note, expression of elongator complex protein gene (MAC_01832), involved in vegetative growth and conidiation, was reduced by ~75% in the ΔMaGlox mutant.Constitutive expression of MaGlox promoted germination, colony growth, and increases tolerances to H 2 O 2 and osmotic stress.This contrasts with the findings in U. maydis, where Glox overexpression does not result in any growth phenotype differences compared to the wild type [15].This discrepancy may be attributed to the presence of three Glox genes in U. maydis, which exhibit functional redundancy.Nonetheless, the overexpression of MaGlox exerts a significantly less pronounced effect on fungal phenotypes compared to its disruption.This reduced impact is likely due to the overexpression of MaGlox not substantially altering the overall metabolic flux, as other rate-limiting metabolic steps remain influential, leading to less pronounced phenotypic differences.
Entomopathogenic fungi are natural biological insecticides that have evolved a series of stress tolerance and detoxification systems to mitigate biotic and abiotic stresses including those that act as host (immune) defenses [28].The ΔMaGlox mutant exhibited decreased tolerances to heat-shock, cell wall stressors and oxidative stresses, but little no changes were seen with respect to osmotic stress.Exogenous H 2 O 2 has been shown to partially restore hyphal growth and conidiation in a T. virens ΔGlox mutant [13], however, this contrasts with our findings in M. acridum where addition of exogenous H 2 O 2 or methylglyoxal was toxic to the ΔMaGlox strain.In terms of the fungal cell wall (and septa), chitin is a major complex carbohydrate component of these structures [51].Changes in the distribution and content of chitin in the cell wall of ΔMaGlox can affect fungal cell wall integrity, thereby affecting tolerance to cell wall stresses, as well as normal growth and development.With respect to virulence, MaGlox affects the formation of infection structures, conidial penetration of the insect cuticle, and proliferation in the host/evasion of host immune responses.Such consequences can be due to either failure to detoxify metabolic by products mobilized during infection, e.g., lipids during appressoria formation, and/or failure to detoxify exogenous RCS produced by the host either as a defense response or via its own metabolic activities.Expression of M. acridum chitinase [52,53] and aspartyl protease (MAC_01391) [51], both of which function in hydrolyzing the insect cuticle, were decreased by >80% in the MaGlox mutant, thus helping to account for the decreased cuticle penetration phenotype.Intriguingly, expression of a KP4 killer toxin protein homolog (MAC_04104), involved in plant fungal pathogenicity, was decreased by 95% in the ΔMaGlox strain, although the consequence of this remains unknown.Impaired virulence was also marked by a decrease in the ability of the mutant to overcome a number of host immune defenses [54,55], resulting in an inability to suppress nodule formation and PO activity, to the extent that the wild type could.

Conclusion
Here, we characterized the Glox gene in the entomopathogenic fungus M. acridum.Our results demonstrate that MaGlox contributes to directly decreasing toxic levels of RCS and indirectly levels of ROS, contributing to oxidative stress tolerance, autophagy, apoptosis, and virulence.These data provide a mechanism by which Glox affects fungal growth, development, and virulence in pathogenic fungi.

Construction of targeted gene disruption, complementation and EGFPtagged MaGlox overexpression strains
Targeted gene disruption and complementation strains of the MaGlox gene were constructed with pK2-PB and pK2-Sur as the backbone vectors, which contain the phosphinothricin resistance (Bar) cassette and chlorimuron ethyl resistance gene, Sur, respectively [57].Transformants were screened by colony polymerase chain reaction (PCR), and the correct integration event verified by Southern blotting and loss of gene expression confirmed via reverse transcription quantitative PCR (RT-qPCR).Primers for vector construction, verification and probe preparation are listed in S1 Table .Briefly, 1.0-kb 5 0 -and 3 0 -flanking sequences of

PLOS PATHOGENS
MaGlox were separately amplified from wild type genomic DNA with primer pairs Glox_LF/ Glox_LR and Glox_RF/Glox_RR.The upstream PCR products were digested with XbaI/EcoRI and inserted into the pK2-PB vector with a Bar cassette (pK2-PB-LGlox).The downstream amplification products were inserted into the SpeI/EcoRV-digested pK2-PB-LGlox, and the construct was delivered into wild type by A. tumefaciens to obtain the MaGlox-disruption mutant (ΔMaGlox).Transformants were screened on Czapek-dox agar containing 500 μg/mL glufosinate ammonium (Sigma, St. Louis, MO, USA) and the desired insertion confirmed by PCR and Southern blotting with DIG-High Prime DNA Labeling and Detection Starter Kit for the latter (Roche, Basel, Switzerland).To generate the complemented ΔMaGlox::MaGlox (CP) strain, the full-length MaGlox sequence including the promoter region was amplified and cloned into vector pK2-Sur.The CP strain was selected on Czapek-dox agar supplemented with 20 μg/mL chlorimuron ethyl (Sigma, Bellefonte, PA, USA).
To examine the subcellular location of MaGlox, an MaGlox-EGFP fusion construct was constructed in which expression of the gene was driven by the glyceraldehyde-3-phosphate dehydrogenase gene promoter (PgpdM) (EGFP-MaGlox-OE).Briefly, the complete MaGlox gene (3762 bp) was amplified with primers OE-F/OE-R and fused with EGFP.The integrity of the final construct was confirmed by sequencing, and the construct was transformed into wild type M. acridum as above.Transformants were confirmed by PCR and RT-qPCR.

Microscopy
Green fluorescence signals in conidia and mycelia for the MaGlox-EGFP strain were observed using a Laser Scanning Confocal Microscope (LSCM) (TCS SP8, Leica, Germany).Briefly, fungal cells cultured on ¼-SDAY were collected and washed twice with sterile H 2 O. Sample were fixed using 4% paraformaldehyde for 20 min.Aliquots (5 μL) were loaded on the slide and then sealed after covered by coverslip, which were then placed under the microscope for observation.

Phenotypic characterizations
Growth, hyphal development and conidiation assays.Aliquots of 50 μL conidial suspensions (1×10 6 conidia/mL) were evenly spread on ¼-SDAY plates and incubated at 28˚C for 20 h for measurement of conidial germination as described previously [58].Mycelial development at 18 h, 24 h and 36 h was observed on ¼-SDAY plates under microscope.Conidial germination and conidial yield analyses was performed on ¼-SDAY as described previously [59].
Fungal stress tolerance to various chemical agents was determined on ¼-SDAY amended with 0.01% SDS, 0.5 mg/mL Congo red (CR), 0.5 M NaCl, 0.05 mg/mL calcofluor white (CFW), 1 M sorbitol (SOR) respectively.Colony growth was compared to control unamended plates after 5 d of incubation at 28˚C.Conidial tolerance to heat-shock (45˚C) and UV-B irradiation (1350 mW/m 2 ) was conducted as described previously [63].The IT 50 after heat-shock or UV-B exposure was compared between the wild type, ΔMaGlox and complemented (CP) strains.

Measurement of cell wall components.
To analyze the effect of loss of the MaGlox gene on cell wall components, fluorescence-labeled antibodies and flow cytometry were used to detect the distribution of chitin and mannan/glucan on the conidial cell wall surface as described previously [64].Fungal cell wall chitin was stained with fluorescein isothiocyanate (FITC)-wheat germ agglutinin (WGA) (Invitrogen, Carlsbad, CA, USA) and mannan/glucan was stained with FITC-concanavalin A (ConA, Vector Laboratories, Burlingame, CA, USA) according to the operation manual.Fluorescence was analyzed with BD FACSCalibur Flow Cytometer (Becton Dickinson, San Jose, CA, USA) and BD CellQuest Pro and FACS Express v3.Mannan/glucan bound with FITC-ConA was detected at excitation wavelength (Ex) of 488 nm and the emission wavelength (Em) of 530 nm.The chitin was examined at the Ex of 488 nm and Em of 630 nm.The analysis of each experiment was repeated three times.All phenotypic and cell wall experiments contained three technical replicates and the entire experiment repeated three times.

Glyoxal oxidase activity assays
Glox enzyme activity was assayed according to the protocol previously described with modifications [65].Culture of wild type and ΔMaGlox in ¼-SDY grown for 2 d at 28˚C were harvested and ground under liquid nitrogen.The resulting tissue powder was suspended in 1 mL of 10 mM PBS (pH 7.0).After centrifugation with 5000 rpm at 4˚C for 10 min, the supernatant was filtered using a 30-kDa ultra centrifugal filter (6000 rpm, 15 min, 4˚C).Proteins retained on the filter membrane were dissolved in 1 mL 50 mM sodium citrate buffer (pH 6.0) containing 5 mM EDTA.Glox enzyme activity was determined by measuring H 2 O 2 formation using a coupled assay with horseradish peroxidase (HRP, Sigma-Aldrich, China) and 2,2 0 -azino-bis (3-ethylben zothiazoline-6-sulfonic acid) (ABTS, Solarbio, China) according to the manufacturer's instructions.The reaction mixture contained 50 mM sodium citrate buffer (pH 6.0) containing 1 mM ABTS, 7 U HRP, 200 μg protein sample and 10 mM substrate methylglyoxal in a total reaction volume 1 mL.The reaction was initiated by the addition of methylglyoxal and the lag phase was eliminated by the addition of 5 μM H 2 O 2 as described previously [66].The absorbance of oxidized ABTS was measured at 420 nm for 3 min at 30˚C using a SPEC-TRAmax 190 Spectrometer (MDC, USA).One enzyme unit was defined as the amount of enzyme that converts 1 μmol of O 2 to H 2 O 2 per minute under the assay conditions.Assays were conducted with three technical replicates and using triplicate biological samples.

Insect bioassays
Insect bioassays were performed using fifth instar L. migratoria nymphs via (1) topical inoculation and (2) intrahemoceol injection as described previously [67].For topical bioassays, aliquots of 5 μL conidial suspensions (1×10 7 conidia/mL) prepared in paraffin oil and conidia were inoculated onto the pronotum of the locust.For intrahemoceol injection assays, aliquots of 5 μL conidial suspensions (1×10 6 conidia/mL H 2 O) were injected into the hemocoel using a microinjector.Respective treatments with paraffin oil or sterile ddH 2 O were used as the blank control for topical bioassay and intrahemocoel injection assays.For each bioassay, three groups of 30 locusts were treated for each test fungal strain and the experiment repeated three times.The number of dead locusts was recorded daily over a period of 11 days.LT 50 was estimated and compared between the wild type and ΔMaGlox strains.
To quantify the total number of spores on the surface and inside of the locusts, host cadavers were collected at 9 d after death, immersed in liquid nitrogen and then ground into a powder and suspended in 2 mL 1% Tween-80.The number of conidia was quantified using a hemocytometer.The experiments were performed in triplicate with three biological samples.

Germination, appressorium formation and penetration assay on locust wings
Conidial germination and appressoria formation on locust hind wings were determined as described previously [68].Briefly, ten sterilized locust hind wings were vortexed in 5-ml tube

PLOS PATHOGENS
Glox detoxifies RCS in a mycopathogen containing 4 ml conidial suspensions in 0.05% Tween-80 (1×10 7 conidia/mL) for 10 min and then spread on the glass slide gently.The glass slides were laid in petri dishes at 28˚C with wet filter paper to keep humidity.Conidial germination was quantified at 6 h, 9 h, 22 h and 26 h and appressorium formation was observed at 22 h and 30 h.Conidia were considered germinated when the germ tube was greater than or equal to the width of the conidia.Appressoria were scored visually.For examining fungal penetration of insect wings, sterilized wings were laid on ¼-SDAY plates.Aliquots of 2 μL spore suspension (1×10 7 conidia/mL) were inoculated in the middle of wings.After incubation for 42 h, the wings were carefully removed, and the plates were continuously incubated for an additional 3 days.Plates in which the spore suspension was directly spotted on ¼-SDAY and grown for 5 d were used as control.

Measurement of host innate immune responses to fungal infection
For examination of locust nodule formation, 5 μL conidial suspension (1×10 8 conidia/mL ddH 2 O) indicated fungal strains was injected into fifth instar locust nymphs.At 12 h post injection, a mid-dorsal cut was made along the full length of the body.The gut and fat bodies were removed to expose the inner ventral surface and then the number nodules were counted in all abdominal segments under a dissecting microscope as previously described [64].Each experiment examined 10 individuals and all experiments were repeated three times.
Prophenol oxidase (PO) activity in hemolymph was tested using L-dopamine (Aladdin, Shanghai, China) as the substrate [69].Briefly, 5 μL conidial suspension (1×10 6 conidia/mL H 2 O) was injected into insect hemolymph, and hemolymph was collected from ten fifth-instar locust nymphs at 12 h post inoculation.The hemolymph sample was centrifuged at 500 g at 4˚C and the cell free supernatant collected.PO reaction mixtures (200 μL) containing 20 μL cell free hemolymph, 50 mM PBS (pH 6.5), 150 mM NaCl, and 10 mM L-dopamine were incubated at 28˚C for 28 h after which the absorbance was recorded at 470 nm with a TriStar LB941 multifunctional microplate reader (Berthold, Bad Wildbad, Germany).The trial was performed in triplicate with 10 locusts/experiment.Protein concentration in cell free hemolymph was determined by the bicinchoninic acid method using a BCA Protein Assay Kit (CWBIO, Beijing, China).One unit (U) of PO was defined as the amount of enzyme that increases 0.001 of absorbance at A470 per minute per mg protein.

Autophagy and apoptosis assays
Autophagy was analyzed using the M52 probe kit (Bestbio, Shanghai, China) of autophagosomes according to the manufacture's protocols, which contains fluorescent probe perylene-3,4-dicarboxylic anhydride to detect mitophagy without mitochondria damage [70].Briefly, the samples washed with PBS was incubated at 37˚C with M52 diluted with detection buffer for 15 min.Fluorescence was observed via microscopy (Nikon Y-TV55, Tokyo, Japan) at an excitation wavelength of 460 nm.Apoptosis was assessed using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling (TUNEL) assay, as previously described [71] with a detection kit (Beyotime, China).Briefly, hyphae were washed with 10 mM PBS and fixed in 4% paraformaldehyde for 1 hour at 4˚C.Samples were then treated with 2 ml of 1 M sorbitol containing Snailase (5 mg/mL; Dingguo, China), cellulase (5 mg/mL; Dingguo, China), and lysing enzyme (5 mg/mL; Sigma-Aldrich, USA) and shaken at 180 rpm for 3 hours at 28˚C.The subsequent staining procedures were carried out following the manufacturer's protocol.TUNEL-positive cells were stained by 3 0 ,3 0 diaminobenzidine (DAB).Propidium iodide (PI; Bestbio, Shanghai, China) was used for live/dead cell staining to detect cell apoptosis.PI fluorescence intensities were visualized using a fluorescence microscope at excitation wavelengths of 488 nm.Gene expression levels of a set of autophagy related genes including atg1, atg8, atg9, atg13, atg17 genes and apoptosis-related genes (Cas, CasA1, CasA2) were determined at different fungal developmental stages by RT-qPCR.Primers used for these experiments are given in S1 Table.

ROS assays
Intracellular ROS levels were detected by using DHE (Uelandy, China) according to the manufacture's protocols.Briefly, fungal cells were treated with 5 μM DHE in 10 mM PBS (pH 7.4) for 1 hour at 37˚C, where DHE reacts with ROS in the cell to form ethidium oxide, which incorporates into chromosomal DNA to produce red fluorescence [72].Red fluorescent signals were observed under a fluorescence microscope (Nikon Y-TV55, Tokyo, Japan) at 512 nm.For enzyme assay measurements, fungal cells were grown in ¼-SDY for 3 d before harvesting.SOD and CAT activities were measured in the extracts of fungal mycelia that had been collected by filtration, washed, and ground in liquid nitrogen, following the guidelines provided by the manufacturer using commercial kits available for each enzyme (BC0200, BC0170, Solarbio, Beijing, China).Intracellular H 2 O 2 levels in conidia were quantified utilizing a hydrogen peroxide test kit according to the manufacturer's protocol (BC3595, Solarbio, Beijing, China).The transcription levels of cytochrome P450 (CYP60, CYP64, CYP105, CYP_fungal), SOD and CAT genes were determined via RT-qPCR.Primers used in these experiments are given in S1 Table.

RCS determination
RCS were determined using UHPLC-QTOF-MS [30], with the raw data analyzed using Mas-sHunter Workstation software (Agilent Technologies, version10.1).Fungal strains were grown in ¼-SDY for 2 days before hyphae were harvested by centrifugation (8000 g, 5 min), washed with ddH 2 O. Hyphal samples were ground into a fine powder in liquid nitrogen and 50 mg samples were dissolved in 100 μL ice-cold perchloric acid (1 M) and 400 μL ddH 2 O.After vortexing for 2 min, the samples were centrifuged for 15 min at 4˚C, and a 100 μL aliquot of the supernatant was mixed with 20 μL of 25 mg/mL butylated hydroxytoluene (Solarbio, China) and 20 μL of 1 M perchloric acid.Samples were then vortexed for 2 min and centrifuged at 14000 rpm for 15 min at 4˚C.Supernatants (80 μL) were derivatized with 200 μL of 200 mM precleaned 2,4-dinitrophenylhydrazine (DNPH) (Shanghai Macklin Biochemical Technology, China) in the dark at 37˚C with shaking for 2 h.The precleaned DNPH was prepared by dissolving the reagent in 16 ml solution containing 12 mol/L HCl, ddH 2 O and acetonitrile (Honeywell, South Korea) in the ratio 2:5:1 (v/v/v).After extraction with hexane twice for derivatization, the samples were centrifuged at 14000 rpm for 10 min at 4˚C and the supernatant was used for further analysis.Acetonitrile was used as the control.Experiments were performed with three biological samples.
High resolution mass spectrometry (HRMS) analyses were carried out using a 6546 quadrupole time-of-flight MS (QTOF-MS; Agilent, UK) coupled with a Dual AJS ESI source operating in negative mode.Source ionization parameters were as follows: gas temperature, 320˚C; gas flow, 8 L/min; nebulizer, 35 psi; sheath gas temperature, 350˚C; sheath gas flow, 11 L/min; PLOS PATHOGENS capillary voltage, 3500 V; nozzle voltage, 1000V; capillary outlet voltage, 150 V; Fixed collision energies, 10V, 30V and 50V.The MS spectra were acquired in a mass range from 50 to 500 m/ z.In the MS-MS mode, the TOF analyzed ions from 50 to 500 m/z.The MS and MS-MS scan rate were 1 Hz and 2 Hz, respectively; and reference mass, m/z 119.0363, 1033.9881.

Gene expression profiling
Total RNA was extracted from 15 d conidia harvested from ¼-SDAY growing at 28˚C using the Ultrapure RNA Kit (CWBIO, Beijing, China) according to the manufacturer's instructions.Three biological replicate samples for each strain were used for sequence analyses.The first and second cDNA strands were prepared as described previously [57].The cDNA libraries were sequenced on DNBSEQ 500 platform (BGI, Shenzhen, China) and the raw data were filtered using SOAPnuke (BGI, Shenzhen, China).After clean reads were aligned to the reference genome (GCF_000187405.1).The genes with transcription level change at least 2-fold with a Q-value � 0.001 were defined as differentially expressed genes (DEGs).Annotation and classification of DEGs were performed using Gene Ontology (GO) and KEGG pathway analysis.The DEGs were searched against the PHI-base database (http://www.phi-base.org/)using DIAMOND to identify pathogenic-related genes.

RT-qPCR
Total RNA was reversed transcribed into cDNA using PrimeScript RT reagent Kit with gDNA Eraser (TaKaRa, Dalian, China).Gene expression levels of specific genes were determined using TB Green Premix Ex Taq II according to the manufacture's protocols.The target gene transcript level was calculated using the 2 -ΔΔCt method [73].Quantification of transcripts was normalized to expression levels of the glyceraldehyde dehydrogenase (Magpd, EFY84384) gene.Three (technical) replicates for each sample were included, and the experiment was repeated three times (biological replicates).Primers for RT-qPCR are listed in S1 Table.

Statistics analyses
GraphPad Prism 7.0 and SPSS 22.0 software (SPSS Inc., Chicago, IL, USA) were used for statistics analysis.One-way ANOVA analysis followed by Tukey's post hoc HSD tests was performed to separate means at P < 0.05, 0.01 or 0.001.The numerical data used in all figures are included in S1 Data.

Fig 9 .
Fig 9. MaGlox functions in fungal growth, conidiation, stress resistance and virulence in M. acridum.MaGlox detoxifies the intracellular RCS to maintain cellular RCS/ROS balance.The loss of MaGlox leads to an accumulation of intracellular RCS and ROS, causing the ER stress, mitochondrial dysfunction, and damage to the nucleus, resulted in autophagy and apoptosis.The DEGs encoded NOX and SOD were up-regulation, which were related the generation of ROS.The glyoxalase I (Glo I) directly detoxifying MG were also up-regulated.The red-blue circles represent the DEGs results related the according category.The DEGs included in the red were upregulated and the blue were downregulated, with the number in the circle representing the quantity of DEGs and its proportion in the items.Genes in the rectangle in red indicated up-regulated and blue indicated downregulated when MaGlox was deleted.https://doi.org/10.1371/journal.ppat.1012431.g009