Insect fungal pathogens secrete a cell wall-associated glucanase that acts to help avoid recognition by the host immune system

Fungal insect pathogens have evolved diverse mechanisms to evade host immune recognition and defense responses. However, identification of fungal factors involved in host immune evasion during cuticular penetration and subsequent hemocoel colonization remains limited. Here, we report that the entomopathogenic fungus Beauveria bassiana expresses an endo-β-1,3-glucanase (BbEng1) that functions in helping cells evade insect immune recognition/ responses. BbEng1 was specifically expressed during infection, in response to host cuticle and hemolymph, and in the presence of osmotic or oxidative stress. BbEng1 was localized to the fungal cell surface/ cell wall, where it acts to remodel the cell wall pathogen associated molecular patterns (PAMPs) that can trigger host defenses, thus facilitating fungal cell evasion of host immune defenses. BbEng1 was secreted where it could bind to fungal cells. Cell wall β-1,3-glucan levels were unchanged in ΔBbEng1 cells derived from in vitro growth media, but was elevated in hyphal bodies, whereas glucan levels were reduced in most cell types derived from the BbEng1 overexpressing strain (BbEng1OE). The BbEng1OE strain proliferated more rapidly in the host hemocoel and displayed higher virulence as compared to the wild type parent. Overexpression of their respective Eng1 homologs or of BbEng1 in the insect fungal pathogens, Metarhizium robertsii and M. acridum also resulted in increased virulence. Our data support a mechanism by which BbEng1 helps the fungal pathogen to evade host immune surveillance by decreasing cell wall glucan PAMPs, promoting successful fungal mycosis.

Introduction Microbial pathogens and their hosts engage in counter-selecting adaptations that define important elements in their co-evolutionary "arms race" [1]. Fungal pathogens have evolved the ability to invade and ultimately consume their hosts. In response, hosts have evolved defense systems that include immune pathways that function to suppress and kill invading parasites. Fungal parasites develop strategies to overcome host defense barriers, which can, in turn, select for additional host defenses that the pathogen must then overcome [1][2][3]. Insect pathogenic fungi have evolved a wide range of adaptations to overcome the formidable defenses of hosts. These adaptations include both the production of (insect) toxins and the ability to detoxify host microbe-targeting toxins and antimicrobial compounds, the production of a battery of enzymes capable of degrading host structures and tissues, and mechanisms of evading host immune responses that has even included the incorporation (via horizontal gene-transfer) of insect genes into the fungal genome that increases the virulence of the fungus [4][5][6].
Insects are considered to lack the canonical acquired immune pathways seen in vertebrates but have evolved innate immune systems that include cellular and humoral pathways that function to block microbial infections [3]. Entomopathogenic fungi typically infect host insects via cuticular penetration, which is initiated after attachment of the conidial (spores) to the host epicuticle, followed by germination and formation of infective structures, such as appressoria or penetration pegs, that function to breach the cuticle using a combination of mechanical pressure and the actions of secreted cuticle-degrading enzymes [7]. Similar to some plant and animal fungal pathogens, once penetrating invasive hyphae reach the hemocoel, they undergo a dimorphic transition, elaborate free floating yeast-like cells, called in vivo blastospores or hyphal bodies, that are able to evade immune responses [8,9]. During infection, fungal tactics used to evade insect immune defense response are diverse, and include a variety of means to evade, inhibit, detoxify, and/or eliminate host immune responses [3,10]. Mechanism used by members of Metarhizium genera of insect pathogens includes the formation of a collagenous protective coat on the hyphal bodies that acts to mask pathogen associated molecular patterns (PAMPs) that are recognized by the host immune systems in order to activate antimicrobial defenses [11]. However, this mechanism does not appear to be present in the related Beauveria genera of insect fungal pathogens. Beauveria bassiana secretes a laccase that inhibits the production of host reactive oxygen species (ROS) as well as oxidizes various (potentially toxic) insect phenolic substrates [12]. In addition, entomopathogenic fungi produce an array of metabolites, toxins, effectors, and other small soluble molecules that deactivate or repress host immune cascades [13][14][15][16] and/ or help overcome insect-generated cytotoxic intermediates [4,12]. Various entomopathogenic fungi have also evolved species specific offensive capabilities, e.g., the secreted catalase-peroxidase produced by M. acridum that acts on the surface of the host during initial stages of infection [17].
One major host defense mechanism is the recognition of fungal cell wall β-1,3-glucans (that acts as a PAMP) to initiate host immune defenses via the well characterized Toll signaling pathway [18]. Within this context, the ability to mask and/or reduce cell wall glucan levels could be considered as a potential mechanism for reducing host immune recognition and subsequent activation. As a broad host range insect pathogen, B. bassiana is considered a more environmentally and human health friendly product as compared to chemical pesticides, and has been commercialized for use in the biological control of a wide range of arthropod pests [4,19,20]. Previously, we have shown that a secreted protein containing a conserved domain of the GH16 glucanase family was found in the cell surface and secreted protein pools of B. bassiana hyphal bodies during hemocoel infection of the silkworm Bombyx mori [21]. However, the function of the protein was not determined. Here, we have genetically and biochemically characterized this (endo-β-1,3-) glucanase (named BbEng1). Our data show that BbEng1 is expressed in response to signals from insect cuticles and hemolymph as well under conditions of osmotic and oxidative stresses. BbEng1 was shown to localize to the fungal (hyphal body) cell wall/surface, decreasing cell wall β-1,3-glucans levels, thus presumably resulting in decreased host recognition and subsequent failure of the host to initiate aspects of its (antimicrobial) immune response. Overexpression of BbEng1 resulted in greater levels of glucan reduction on fungal hyphal bodies and increased virulence as compared to the wild type parent. Orthologs of BbEng1 in Metarhizium appear to perform a similar function, indicating this to be a general mechanism used by entomopathogenic fungi to help evade host immune recognition.

BbEng1 is induced by insect-derived nutrients in B. bassiana
BbEng1 (BBA_04753) was previously isolated from the cell surface of B. bassiana insect hemocoel derived hyphal bodies as well as found (secreted) in the hemolymph from fungus infected silkworm larvae [12]. The protein consists of 588 amino acids, and features a putative N-terminal signal peptide (21 amino acids) and a putative conserved GH16 family glucanase domain (residues 22 to 343), which contains the catalytic motif (E-[ILV]-D-[IVAF]-[VILMF](0,1)-E) typically found in this family of proteins, and which includes two critical glutamic acids (Glu141 and Glu146) involved in catalysis as well as a conserved aspartic acid residue (Asp143) [22][23][24] (S1 Fig). In addition, BbEng1 contains 19 cysteine residues with the potential to make disulfide bridges. Phylogenetic analyses revealed broad distribution for BbEng1, that clustered to three distinct clades that mirrored host associations, i.e., into mammalian, plant and insect pathogens (S1 Fig).
The expression pattern of BbEng1 was examined via RT-qPCR and construction of a BbEng1 promoter-GFP (PBbEng1::GFP) reporter strain (Fig 1A and 1B). These data showed that BbEng1 was highly expressed in hemolymph derived hyphal bodies, but absent in in vitro cells, including conidia, submerged blastospores, and aerial and submerged hyphae grown under standard conditions in mycological media. BbEng1 expression increased 20-to 60-fold in basal salt solution (BS) or Czapek-Dox broth (CZB) containing insect-derived extracts, including insect cuticle and hemolymph produced as detailed in the Methods section. BbEng1 expression was also slightly induced by NaCl-and sorbitol-induced osmotic stress (~2.0-fold), and by oxidative stress (1.8-2.1-fold) as induced by addition of either tert-Butyl hydroperoxide, H 2 O 2 or menadione to CZB. However, no obvious changes in BbEng1 expression was seen

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction Cuticle and hemolymph, BS + 0.167 g / L silkworm cuticle or 5 mL / L hemolymph. Chitin, glucan or trehalose, used as sole carbon source in CZB as a substitute for sucrose at 2%. NaCl, sorbitol, H 2 O 2 , MND, t-BHP or lactic acid, CZB containing 0.5 M NaCl, 1.0 M sorbitol, 5.76 mM H 2 O 2 , 37 μM menadione or 0.78 mM tert-Butyl hydroperoxide or 2% (v/v) lactic acid. LO, low oxygen condition (with~6% of initial O 2 concentration). All fungal cells were cultured 6 h in different media and under different stress conditions. Error bars denote standard deviations (SDs) from three biological replicates. Different letters (a-f) indicate statistically significant differences (P < 0.01 in LSD test). B. Microscopic images (scale: 5 μm) of GFP fluorescence expressed in fungal cells of PBbEng1::GFP strain. GFP was driven by BbEng1 promoter. C. Microscopic images (scale: 5 μm) of GFP fluorescence in BbEng1::GFP hyphal bodies and cells cultured in insect-derived nutrients. Hyphal bodies were examined in hemolymph samples at the indicated time points after intra-hemocoel injection of conidia into G. mellonella larvae (left). Fungal cells grown in insectderived nutrients (5 mL / L hemolymph or 0.167 g / L cuticle from silkworm) were collected at the indicated time points (right) of incubation. D. Microscopic images (scale: 5 μm) of GFP fluorescence and FM4-64 (stained cell membrane) in BbEng1::GFP hyphal bodies. https://doi.org/10.1371/journal.ppat.1011578.g001

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction under low oxygen stress (LO) or lactic acid-stress growth conditions. BbEng1 expression was slightly (1.5-2 fold) elevated in cultures grown in 1/4 SDY and media supplemented with chitin, glucan or trehalose, however, no obvious differences in cells containing the GFP promoter fusion fluorescence construct (PBbEng1::GFP) in the same media.

BbEng1 localizes to the cell wall/cell surface and is secreted during infection and in media containing insect cuticle or hemolymph
To examine localization of BbEng1, wild type B. bassiana was transformed to express a BbEng1::GFP fusion protein as detailed in the Methods section. No fluorescent signal was observed in conidia prior to injection into the insect hemocoel (0 h), with fluorescent signals appearing in hyphal bodies formed 36 h post-injection ( Fig 1C). No fluorescence was seen in BbEng1::GFP cells at the initial period (at 0 h) after inoculation in insect cuticle or hemolymph-contained broth, but apparent in the fungal cells after 6 h of inoculation, particularly along/ within the cell wall and on the cell surface of germ tubes ( Fig 1C). Moreover, signals could be clearly distinguished as occurring at the outer layer of the cellular membranes via counter staining with FM4-64 ( Fig 1D).
To characterize the function of the gene product, targeted gene knockout (ΔBbEng1), complementation (ΔBbEng1::BbEng1, Comp) and overexpression (BbEng1 OE ,~200-fold upregulated in 1/4 SDY) strains were generated as detailed in the Methods section (S2 Fig). Using an indirect immunofluorescence (IFL) assay to probe cells with a polyclonal antibody generated towards BbEng1 (anti-BbEng1, see Methods) coupled to goat anti-rabbit IgG-FITC labeling, BbEng1 signals were seen as distributed as punctate spots on the outside edge of/ in the calcoflour white (CFW) counter stained cell (walls) in B. bassiana hyphal bodies (Fig 2A). The signal was significantly increased in BbEng1 OE hyphal bodies but was absent in ΔBbEng1 cells (Fig 2A). Immuno-transmission electron micrographs revealed a localization of BbEng1 on the cell wall and its distribution on the cell surface of BbEng1 OE hyphal bodies ( Fig 2B). Experiments using the BbEng1::GFP fusion construct bearing strain, indicated significantly decreased fluorescence signals (46%, P < 0.01) in hyphal bodies treated for 12 h with 2 mM dithiothreitol (DTT), with the fluorescence intensity in the surrounding solution (i.e., released into the media) increased by 2.5-fold in comparison to controls (not treated with DTT) ( Fig  2C). To further measure the release of BbEng1::GFP from cell walls after DTT treatment, samples from the cell cultures treated without or with DTT were probed with anti-GFP antibody. The results confirmed "release" of BbEng1::GFP from cell wall fractions after treatment with DTT ( Fig 2C). To determine whether BbEng1 was secreted in response to host cues, Western blots probed with an anti-BbEng1 antibody revealed the presence of secreted BbEng1 protein in B. bassiana grown in media supplemented with silkworm cuticle or hemolymph, but not when the fungus was grown in nutrient-rich 1/4 SDY. A secreted pool of BbEng1 was also detected in vivo from the hemolymph of Galleria mellonella 48 h after injection of wild type and BbEng1 OE conidia. Quantification of in vivo secreted signals indicated~0.1 μg and 0.15 μg of BbEng1 per 100 μg of protein extract in hemolymph samples of wild type and BbEng1 OE infected insects, respectively ( Fig 2D).

Involvement of BbEng1 in cell wall remodeling
The ΔBbEng1, BbEng1 OE and control strains were examined under different growth and developmental conditions. The BbEng1 OE strain showed a slight to moderate increase in colony growth (6-24%, P < 0.01) on nutrient-rich 1/4 SDAY, minimal medium CZA, and CZA modified with different carbon sources (S3 Fig), as well as moderate increases (20-65%, P < 0.01) in conidial production in various media (S3 Fig). BbEng1 OE conidia also more readily

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction dissociated from conidiophores than the wild type strain (S3 Fig). BbEng1 OE colonies displayed denser mycelial edges than control strains (wild type and complemented strains), with an abnormal curled hyphal morphology apparent ( Fig 3A). Ratios of polymorphic conidia (swollen and oval with section areas > 6 μm 2 ) were significantly higher in the BbEng1 OE strain (34%, P < 0.01) than those of the wild-type conidia (5.1%) (Fig 3B). Examination of dimorphic transition showed that similar to conidiation, the BbEng1 OE strain also produced more blastospores as compared to the wild type strains (S4 Fig). BbEng1 OE conidia also germinated slightly more rapidly than wild type conidia, with a median germination time (GT 50 ) shortened by 11-13% (P < 0.01) as compared to the control strains' conidia (S3 Fig). In addition, all germinated

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction BbEng1 OE conidia produced singular germ tubes (unidirectional hyphal growth) with fewer septum, contrasting with high incidences of bidirectional germination and hyphal growth (~63%) and higher septal density in the control strains (Fig 3C and 3D). No obvious differences were seen in colony growth, conidiation and germination as well as septation patterns between ΔBbEng1 and control strains (Figs 3 and S3).
Cell wall morphology was examined using transmission electron microscopy (TEM) as detailed in the Methods section. These data indicated an electron-dense brush-like structure present on the surface (edge) of BbEng1 OE hyphal bodies but absent in control strains, resulting in an increase in cell wall thickness for BbEng1 OE (0.092 ± 0.005 μm) as compared to wild type (0.051 ± 0.005 μm, n = 30, P < 0.01) (Fig 4A and 4B). Increased cell wall thickness was also noted in TEM images of BbEng1 OE conidia and blastospores (S5 Fig). In contrast, the ΔBbEng1 mutant showed a reduction in surface brush-like structures coupled to decreased cell wall thickness (to 0.039 ± 0.005 μm, n = 30, P < 0.01) of hyphal bodies (Fig 4A and 4B). Measurement of cell wall β-1,3-glucan contents from cells grown in 1/4 SDY (for 48 h) using an aniline blue assay revealed that overexpression of BbEng1 resulted in a significant reduction in β-1,3-glucan level (38%). No obvious differences were seen with respect to β-1,3-glucan content between the ΔBbEng1 and wild-type strains (S6 Fig). Similar results were obtained via detection of fluorescence intensities after probing with a monoclonal β-1,3-glucan specific antibody followed by application of a goat anti-mouse IgG-FITC secondary antibody in examinations of conidia and blastopores after fungal cells were fixed in 3% (v/v) formaldehyde as detailed in Methods section. Cell wall chitin contents assessed using HCl (6 N) hydrolysis (as detailed in Methods section) also indicated no obvious differences in chitin content for any of the strains tested. Similarly, no differences were seen in fluorescence labeling of the cell wall using FITC-conjugated wheat germ agglutinin lectin (WGA, specificity for N-acetylglucosamine and N-acetylneuraminic acid residues) or in cell wall labeling using calcofluor white (CFW, S6 Fig). Cell wall glucan detection using a monoclonal β-1,3-glucan specific antibody and a goat anti-mouse IgG-FITC antibody revealed that β-1,3-glucan levels were dramatically decreased in the cell walls of BbEng1 OE hyphal bodies (40.7%, P < 0.01), but increased in ΔBbEng1 hyphal bodies (36. 7%, P < 0.01) as compared to wild type (n = 200). However, no obvious difference was observed in chitin content of hyphal bodies from any of the tested strains as seen after labeling with either WGA or CFW (Fig 4C and 4D).
Hydrolyzed products > pentasaccharides were seen by TLC analyses using purified fungal cell wall alkali-soluble fractions (AS) and water-soluble fraction (WSDN) as substrates, which were derived from nitrous deamination of fully de-N-acetylated alkali-insoluble fractions (AI) as detailed in the Methods section ( Fig 5B). Similar hydrolyzed products were also seen after BbEng1 incubation with the AS and WSDN fractions as measured by HPLC (peaks eluted at 6.60 and 7.31 min, respectively) ( Fig 5C). No products were seen after BbEng1 treatment of additional substrates including fully de-N-acetylated fungal AS (ASDN) or AI (AIDN) fractions (Fig 5B and 5C).

BbEng1 alters cell wall carbohydrate epitopes on hyphal bodies
Carbohydrate cell wall epitopes of B. bassiana hyphal bodies derived from the wild type, ΔBbEng1, and BbEng1 OE strains were further examined via use of three fluorescence-labeled lectins including: concanavalin A (ConA, specific to α-mannopyranosyl and α-glucopyranosyl

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction fluorescence intensities of all three lectins labeled BbEng1 OE strain were significantly lower than other strains. Although little to no labeling of conidia or blastospores was seen using the β-1,3-glucan monoclonal antibody for any of the strain, the fluorescence intensities on BbEn-g1 OE strain were significantly lower than those on the other strain (S9 Fig).

BbEng1 expression facilitates fungal infection
TLC and HPLC analyses revealed the release of detectible carbohydrate products from insect cuticle fractions (AS, ASDN and WSDN) after treatment with purified BbEng1 (Fig 5B and  5C), suggesting a potential role of the secreted BbEng1 in participation in degradation and utilization of host carbohydrate containing substrates during infection. Insect bioassays using last-instar larvae of the greater wax moth, G. mellonella, via topical application of the fungal strains revealed a calculated median lethal time to kill 50% of hosts (LT 50 ) for the BbEng1 OE strain = 109.9 ± 1.2 h, which was significantly lower (i.e., showed increased virulence, P < 0.01) as compared to the wild type and complemented strains (LT 50 = 127.0 ± 1.6 h and 125.8 ± 1.2 h, respectively) which were in turn significantly lower (P < 0.01) than the ΔBbEng1 strain, LT 50 = 147.3 ± 3.5 h (Fig 7A). Similarly, bioassays performed in which fungal cells were directly injected into host larvae (i.e., via cuticle-bypassing intrahemoceol injection) revealed a decreased LT 50 for the BbEng1 OE strain (69.1 ± 3.9 h, P < 0.01) as compared to control wild type and complemented strains (89.3 ± 2.7 h and 88.9 ± 1.4 h), which were lower than the ΔBbEng1 LT 50 which was = 100.3 ± 2.7 h (P < 0.01) ( Fig 7B). In addition, the BbEng1 OE strain proliferated more rapidly out of the cadavers as compared to the other strains ( Fig 7C).
Melanization reactions/spots, hallmarks of one aspect of the insect immune system, were more apparent on ΔBbEng1 infected larvae, 48 h after topical inoculation, and reduced on BbEng1 OE infected larvae, as compared to controls (Fig 7C). Larvae infected by ΔBbEng1 turned black 120 h after topical inoculation and 60 h post-injection, retaining some of this color even during "mumification", i.e., fungal out growth at 240 / 180 h for topical / injection applications as contrasted with the BbEng1 OE and control strains (Fig 7C). To examine host hemocyte-fungal interactions, conidia injected into the hemocoel were examined microscopically over a time-course of the infection. Unlike as seen in controls, fungal cells from ΔBbEng1 encapsulated by hemocytes at 24 h post-injection did not escape encapsulation until 36-48 h post-injection, whereas BbEng1 OE conidia / fungal cells were not encapsulated by insect hemocytes even at the early stage post-injection, and were able to germinate and proliferate directly within the hemolymph without any delay ( Fig 7D). Total fungal load analyses by qPCR from hemolymph samples of injected larvae revealed more rapid proliferation of BbEng1 OE than of wild type in the hemocoel, with the latter forming 50% and 71% more cells at 48 and 60 h postinjection, respectively. In contrast, proliferation of ΔBbEng1 was reduced by 25-35% at the same time points in comparison to controls (Fig 7E).
To examine host immune responses, phenoloxidase (PO) activity, reactive oxygen species (ROS) production, and expression levels of host Toll pathway genes were measured. For wild type, PO activity increased during the first 9 h period post-injection and then decreased ( Fig  8A). In contrast, larval PO activity was significantly higher (P < 0.01) after injection of ΔBbEng1, and lower (P < 0.01) after injection of the BbEng1 OE strain (Fig 8A). ROS levels increased from 3 to 24 h post-challenge by the wild type, a response which was dampened for BbEng1 OE but higher for the ΔBbEng1 strain (P < 0.01) (Fig 8B). Transcriptional profiling targeting 4 Toll pathway genes [25,26], 8 antimicrobial peptide (AMP) genes [16] and 2 proPO system-associated genes [27,28] were performed using larvae 12 h post-injection with the various strains as detailed in the Methods section. Expression levels for the Toll pathway genes (28-393%), 6 AMP genes (12-65%) and 2 proPO system-associated genes (56% and 68%) were higher for the ΔBbEng1 versus wild type injected larvae (Fig 8C), whereas expression of all analyzed genes was lower in larvae injected with BbEng1 OE (14-99.8%, P < 0.01) (Fig 8C).

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction repressed by inclusion of BbEng1 (55%, P < 0. 01). However, the UKC-activated PO activity was not repressed by injection with inactivated / denatured BbEng1 (iBbEng1) (Fig 8D). Transcriptional profiling revealed that the inclusion of BbEng1 also repressed UKC-activated expression of several host immune genes with the exception of PGRP, a Toll pathway gene whose expression was also decreased with the addition of iBbEng1 (Fig 8E). BbEng1 (FITClabeled) was able to bind the cell walls of UV-killed as well as wild type (normal) conidia, whereas FITC-iBbEng1 only weakly bound to the conidial cell walls (Fig 8F and 8G).  Expression levels of genes involved in the Toll pathway, antimicrobial peptide (AMP) and proPO system in the larvae 12 h post-injection (β-actin used as a reference for normalized fold expression, NFE). D. Effects of added 0.5 μM BbEng1 or iBbEng1 (heated-inactivated BbEng1) on fungal cell wall (UKC, UV-killed B. bassiana conidia, 5 × 10 6 conidia / mL)-activated PO activity (total volume: 2 μL). Hemolymph samples taken at 12 h after injection of UKC, UKC + BbEng1 or UKC + iBbEng1 were used for PO assay. Larvae injected with 2 μL of 0.85% NaCl or UKC+ BSA (bovine serum albumin, 0.5 μM) were used as controls. E. Expression levels of genes in the Toll pathway components, AMP proPO-system-associated genes in the larvae at 12 h after injection of UKC or UKC + proteins (β-actin used as the reference). F, G. Microscopic images and fluorescence intensities of UKC and conidia (CO) (n > 100) labeled with FITC-labeled BbEng1. Error bars: SDs. Different letters indicate significant differences (P < 0.01 in LSD test).

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction overexpressing BbEng1 (Mr-BbEng1 OE ) or MrEng1 (MrEng1 OE ) and (2) M. acridum strains overexpressing BbEng1 (Ma-BbEng1 OE ) and MaEng1 (MaEng1 OE ) were constructed (S11 Fig). Similar to the BbEng1 OE strain, the various Eng1 overexpression strains showed faster colony growth, more biomass accumulation, higher conidial yield, and accelerated germination as compared to their respective parental wild type strains (S12 Fig). In addition, insect bioassays using G. mellonella, indicated significantly increased virulence for the Mr-BbEng1 OE and MrEng1 OE strains as compared to the wild type parent. The resultant LT 50 values via topical application and intrahemocoel injection were 128.5 ± 2.3 and 68.7 ± 1.0 h for Mr-BbEng1 OE , 122.0 ± 3.4 and 65.2 ± 1.0 h for MrEng1 OE and 155.9 ± 8.1 and 77.3 ± 1.0 h for wild type M. robertsii (Fig 9A and 9B). As M. acridum is a more acridid specific pathogen (although it can also infect G. mellonella), these strains were assayed using 5th instar locusts (Locusta migratoria manilensis) by placing 5 μL of a 10 7 conidia / mL suspension on the host pronotum. LT 50 values for Ma-BbEng1 OE , MaEng1 OE and wild type M. acridum were 123.8 ± 5.0, 121.6 ± 5.2 and 136.6 ± 1.8 h, respectively (Fig 9C)

Discussion
Glucanases produced by some plant and animal fungal pathogens during infection and/or in response to host signals, such as carbon sources, oxygen limitation or hypoxia, and iron, can

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction remodel the cell wall to reduce or mask exposed PMAPs that comprise β-glucans, so as to evade host immune recognition [29][30][31][32]. The present study adds the existence of similar mechanisms in insect-pathogenic fungi including B. bassiana, and both generalist and more specialized Metarhizium species. Our data show these pathogens express a specific virulenceassociated endo-β-1,3-glucanase (Eng1) during entry into the hemocoel as part of the infection process that can be mimicked in vitro by addition of insect-derived substrates including cuticle and hemolymph. In addition, a low level of induction can occur in vitro under osmotic or oxidative stress. Both of these occur during the infection process, both during initial stages of infection on the insect cuticle as well as during growth within the host hemocoel, and hence these may act minor cues for the fungus that they are interacting with a potential host [22,33]. Our hypothesis was that the main function of BbEng1 is to facilitate evasion of host immune activation during fungal infection by decreasing surface glucan that are recognized by host immune surveillance systems. Our data show that the protein is distributed on both cell wall and surface fractions. BbEng1 is secreted and can bind to the fungal cell wall. BbEng1 was previously found as a cell surface protein on hyphal bodies that could be released from cells by treatment with 2 mM DTT [21]. Here we show using a green fluorescence fusion protein BbEng1::GFP, distribution of the protein on the surface of hyphal bodies that was also labile to DTT treatment. These results suggest important disulfide linkages that may mediate BbEng1 interaction with other cell wall proteins and/or structures. Immunofluorescence (IFL) assays, however, showed punctate BbEng1 signals that distributed on the outside edge of/ in calcofluor white staining in hyphal bodies. This discrepancy is likely due to technical differences between the assays in which IFL experiments involve extensive washing as well as potential steric effects due to use of antibody/secondary antibody visualization.
To determine the role of BbEng1, genetic characterization including construction of a targeted gene knockout mutant (ΔBbEng1) and overexpression strains (BbEng1 OE ) were performed. Our data revealed that the ΔBbEng1 strain was compromised in virulence whereas the BbEng1 OE showed enhanced virulence, using insect bioassays performed either via topical (cuticle) or intrahemocoel injection (cuticle-bypassing) methodologies. In larvae infected by BbEng1 OE , cuticle melanization spots were reduced indicating a failure of the host to encapsulate the fungal cells by aggregated hemocytes. In contrast, infection by ΔBbEng1 saw increased levels of cuticle melanization and prolonged encapsulation of the fungal cells by aggregated hemocytes, which was consistent with these cells having a reduced ability to escape these responses. Overall immune responses followed a similar pattern, with host PO activity and expression of suite of immune genes higher after ΔBbEng1 infection but lower after BbEng1 OE infection as compared to the wild type. Insect glucan-binding proteins (β-1,3-glucan recognition proteins, βGRPs) can recognize fungal PAMPs, i.e., β-1,3-glucan and chitin [34,35] to activate innate immune response including activation of the prophenoloxidase (proPO) cascade [36,37] and the Toll receptor pathway to initiate production of antimicrobial compounds and other immune responses [38]. As mentioned, expression of the host β-GRP gene as well as other Toll pathway genes and proPO cascade genes was lower during BbEng1 OE infection but higher during ΔBbEng1 infection as compared to wild type, correlating with the altered insect immune responses observed in the bioassays. All of these responses are hypothesized to be mainly due to decreased levels of immune recognized cell wall glucan in the BbEng1 OE as compared to increased levels seen for the ΔBbEng1 strain, when compared to the wild type parent. However, additional altered cell wall epitopes as detected using lectins (ConA and WGA) and anti-β-1,3-glucan antibody were noted. These data revealed exposure of other potential PAMPs in hyphal bodies, including α-mannopyranosyl / α-glucopyranosyl residues, GlcNAc and GlcNAc acid residues (chitin) and β-1,3-glucan were significantly increased in the ΔBbEng1 strain, and conversely decreased in the BbEng1 OE strain as compared to wild type.
Thus, in addition to affecting β-1,3-glucan levels, modification of other surface PAMPs may occur in the mutant strains due to changes in surface glucan levels. Purified BbEng1 was shown to be able to bind to the cell walls of both UV-killed (UKC, with our data showing that these cells can still activate host immune responses) and normal conidia, with the UKC-activation of host immune responses blocked by co-injection with purified BbEng1 but not by denatured BbEng1. These data support the idea that BbEng1 removes surface glucans rendering the cells more "stealth-like" in terms of being recognized by host immune surveillance systems. These findings are similar to previous observations in mammalian fungal pathogens H. capsulatum and C. albicans, which secrete endo-β-1,3-glucanase (Eng1) and an exo-β-1,3-glucanase (Xog1), respectively, during infection or in response to various host-derived signals that act to remove cell wall β-glucans facilitating evasion of the fungal cells from host immune recognition [29][30][31][32].
Aside from evasion, assimilation of host-derived nutrients is crucial for successful fungal infection. Phytopathogenic fungi are known to secrete β-1,3-glucanases and other plant cell wall hydrolases and glycosidases, including polygalacturonases, cellulases, xylanases, laccases, and hemicellulases, to break down plant cell wall structures for acquisition of host nutrients and facilitation of fungal colonization on plant tissues during infection [39,40]. Similarly, fungal insect pathogens express a variety of hydrolytic enzymes, such as proteases, chitinases, lipases and peptidases, to breach the insect exoskeleton for uptake of host nutrients [4,41,42]. Different, but overlapping, sets of enzymes are presumably important for breaching the cuticle and the fungal growth in the hemocoel. It is unclear the extent to which BbEng1 may act in terms of nutrient acquisition and its main function is likely to be to act on the surface of the fungal cell. Indeed, cell wall visualization of hyphal bodies revealed that BbEng1 OE cells show increased cell wall thickness (and decreased β-1,3-glucan) whereas the ΔBbEng1 mutant displayed the opposite phenotype (decreased cell wall thickness and increased β-1,3-glucan levels) when compared to the wild type. It is well known that β-1,3-linked glucan is a main component of fungal cell walls, and β-1,3-glucanases function in cell wall glucan metabolism, impacting fungal growth and development, including morphogenesis during spore germination, hyphal elongation and hyphal branching in filamentous fungi [43][44][45][46][47]. Thus, it is not too surprising that the BbEng1 OE strain showed faster growth, abnormal hyphal morphologies, increased conidiation levels, faster germination, mono-directional hyphal growth as compared to the wild type. As expected, BbEn-g1 OE conidia and blastospores had thicker cell walls and lower β-1,3-glucan content. However, cell wall composition was not significantly altered for in vitro cells (as opposed to the in vivo hyphal bodies) when comparing ΔBbEng1 to the wild-type strain. No activity for BbEng1 was seen using insect cuticle or chitin (colloidal chitin) as substrates, however, release of carbohydrate products from insect cuticle fractions (AS, ASDN and WSDN) after treatment with BbEng1 were detected by TLC and HPLC analysis. These data suggest a potential role for secreted BbEng1 in degrading insect β-glucan substrates, however the extent to which this occurs and its importance in the infection process remains to be further characterized.
Our data shows that BbEng1 helps fungal cells to evade host immune responses, including hemocyte encapsulation and melanization likely by depleting the fungal cell surface of a crucial PAMPs (β-1,3-glucan) that is a critical step in the transition from hyphal invasion into the hemocoel to the formation of free floating yeast-like hyphal bodies that must evade host detection and/or any subsequent immune responses. Our data show that this process occurs in Metarhizium and likely in most if not all fungal entomopathogens. Overall, our results determine that the endo-β-1,3-glucanase BbEng1 secreted during the infection process of B. bassiana and other fungal entomopathogens acts to render fungal cells less able to be recognized by host immune systems. BbEng1 functions as a cell wall remodeler enabling removing cell wall glucans, allowing for fungal growth in host hemocoel.
Last instar larvae of G. mellonella (greater wax moth) from Huiyude Technology (Tianjin, China) and the 5th instar nymphs of L. migratoria manilensis provided by School of Life Sciences, Chongqing University (China), were used in virulence bioassays and insect immune responses.

Targeted gene manipulation
All primers used for gene manipulation are given in S2 Table. To examine BbEng1 gene expression, the enhanced green fluorescent protein (eGFP) gene was fused to the promoter sequence of BbEng1. Briefly, the BbEng1 promotor (1093 bp) was amplified from B. bassiana using primer pair C1 / C2 and inserted into the HindIII / EcoRV restriction sites of pK 2 surPB3-GFP [21] to form vector PBbEng1::GFP. The resultant vector was introduced into wild type B. bassiana using Agrobacterium-mediated transformation of conidia and selection of 40 μg / mL sulfonylurea [48]. Transformants were verified using PCR with primer C1 / GFP2 (1823 bp). For protein localization, a strain expressing the BbEng1::GFP fusion protein was constructed by fusing eGFP to 3'-terminus of BbEng1. Briefly, BbEng1 (4168 bp containing 2194 bp promotor sequence) was cloned into the NdeI / EcoRV restriction sites of pBARG-PE1-GFP [21] replacing gpdA promoter to yield vector construct BbEng1::GFP, which was then transformed into wild type B. bassiana as described with selection of 200 μg / mL phosphinothricin [49]. Transformants were verified via PCR with primer pair RT1 / GFP2 (1160 bp).
For construction of a targeted gene knockout of BbEng1, its partial coding region (197 bp) was replaced with the bar marker (resistant to phosphinothricin) through A. tumefaciensmediated homologous recombination as described previously [50]. Putative gene disruption mutants were screened on selective media (200 μg / mL phosphinothricin) and examined via PCR with primer pair S1 / S2. Integrity of mutants were then verified using reverse transcription-PCR (RT-PCR) with primer pair RT1 / RT2 (to check transcript levels) and Southern blotting (single insertion in correct site). Southern blots were performed using a digoxinlabeled bar fragment as probe. Complementation strains of BbEng1 were generated by ectopic integration of BbEng1 into the ΔBbEng1 mutant. Briefly, the coding region and upstream promoter regions of BbEng1 (4118 bp) was amplified from the wild-type DNA using primer pair RC1 and RC2 and inserted into the BamHI / XbaI restriction sites of modified pCB1536 containing the sur (resistant to sulfonylurea) marker [51]. The resultant vector was introduced into the ΔBbEng1 strain as described above. Transformants were selected on plates amended with 40 μg / mL sulfonylurea, screened by PCR with primer pair S1 / S2, and verified by RT-PCR with primers RT1 / RT2. Southern blotting was performed to confirm the integration of BbEng1. A BbEng1 overexpression strain was constructed by an introduction of its coding sequence under control of the Aspergillus nidulans gpd promoter. Briefly, the BbEng1 coding region (1947 bp) was amplified with primer pair OE1 / OE2 from the wild-type DNA and cloned into the BamHI / EcoRI restriction sites of pBARGPE1. The resultant vector was then transformed into the wild type strain. Transformants were selected on plates containing 200 μg / mL phosphinothricin, and verified by PCR using primer pair S3 / S4 (partial PgpdA:: BbEng1 fragment, 652 bp) and RT-qPCR with the primer pair RT1 / RT2. To

Gene expression and blotting analyses
RT-PCR and RT-qPCR analyses and/or analysis of fluorescence signals in the PBbEng1::GFP strain were performed to examine the expression patterns of BbEng1 and its' its homologous genes, and insect immune genes. RT-PCR and RT-qPCR analyses were performed toward target genes and the reference of 18S rRNA in B. bassiana (XM_008603008.1), gpd in M. robertsii (XM_007825673.1) and M. acridum (XM_007817733.1), and β-actin (XM_026904349) in G. mellonella. Primers for detection of all genes examined are given in S2 Table. Transcript levels of BbEng1 in the G. mellonella hemolymph-derived fungal cells and hyphal bodies of the disruption, complementation, and overexpression strains were examined via RT-PCR and the used cDNA samples were prepared as described previously [52]. RT-qPCR and / or GFP fluorescence signals were used for transcriptional profiling of BbEng1 and its homologues in the wild type and overexpression strains of B. bassiana, M. robertsii and M. acridum. Expression levels of host (G. mellonella) immune genes were also examined via RT-qPCR with primers (S2 Table). Conidia, blastospores, hyphal bodies, aerial hyphae and submerged hyphae were collected as described previously [21,52]. Media used included basic salt broth (BS, 0.3 g / L MgSO 4 .7H 2 O, 0.3 g / L K 2 HPO 4 , 0.3 g /L NaCl) [12], 1/4 SDY, BS containing insect-derived nutrients (i.e., 5 mL / L silkworm hemolymph or 0.167 g / L silkworm cuticle), and CZB amended with the carbon source of 1% chitin, glucan or trehalose. For response to stress cues, fungal growth was initiated on CZB supplemented with 0.5 M NaCl, 1.0 M sorbitol, 5.76 mM H 2 O 2 , 37 μM menadione, 0.78 mM tert-Butyl hydroperoxide or lactic acid (2%). For fungal growth under low oxygen conditions, fungal cells were inoculated on CZB broth sealed with liquid paraffin for static cultivation (with~6% of initial O 2 concentration) [53]. Insect hemolymph and cuticle were isolated and prepared from the 5th instar larvae of Bombyx mori as described previously [52]. To assess transcript levels of BbEng1 and its homologues in the overexpression strains, RNA was isolated from 1/4 SDY (48 h) cultures. Insect immune genes (S2 Table) were examined at 12 h post-injection with 2 μL suspension aliquots of (i) conidia (10 7
For Western blot analyses, two predicted BbEng1 antigenic regions, (RGTIPDRPDDNWSWV, residues 43-56 of the N-terminal domain of GH16 family glucanases; TAAPEPESANQIC, residues 574-586 of the C-terminal domain) were synthesized and conjugated with keyhole limpet hemocyanin. Targeted antibodies were raised in New Zealand White rabbits and designated as anti-BbEng1 antibody. Western blot analyses were performed according to the manufacturer's instructions (Bio-Rad). Samples tested included BbEng1 secreted by the wild-type and overexpression strains grown in either insect-derived nutrient cuticle or hemolymph containing minimal media for 6 h and (iii) insect hemolymph for 48 h after injection of 2 μL 10 7 spores / mL. Blots were probed with the anti-BbEng1 antibody as described previously [52]. Protein concentrations were quantified using a BCA (bicinchoninic acid) assay kit (Generay Biotech, Shanghai, China). BbEng1 concentrations were densitometrically quantified using the yeast expressed, purified BbEng1 as a standard and the ImageJ software [55]. Secreted BbEng1::GFP fusion protein was detected using similar protocols and blots probed with an anti-GFP antibody (Thermo-Fischer Science).
Southern blot analysis was performed using standard protocols with a DIG-High Prime DNA labeling and detection starter kit I (Roche). Probes corresponding to the bar and sur fragments were prepared as described previously [56] and used for verification of ΔBbEng1 mutant (bar) and complementation strain (sur), respectively.

Heterologous expression and characterization of BbEng1
To determine BbEng1 activity, the protein was expressed in P. pastoris and purified from the yeast culture. Briefly, BbEng1 cDNA lacking the endogenous signal peptide was amplified by PCR with primer pair P1 / P2 and cloned into the EcoRI / NotI restriction sites of plasmid pPIC9K (Invitrogen), and the integrity of the clone was verified by sequencing. The resultant vector was introduced into the methylotrophic yeast strain P. pastoris GS115 (Invitrogen) according to the manufacturer's instructions. Selection of the recombinants, protein expression and purification were performed as described previously [52]. Protein was quantified using the BCA assay.
The ability of BbEng1 to bind various polysaccharides, insect cuticle and the surfaces of fungal conidia was examined in pull-down assays followed by Western blot analyses as well as by using fluorescein isothiocyanate (FITC)-labeled proteins as described previously [15,59]. Bound protein detected in either assay was densitometrically quantified using the ImageJ software [55].
Glucanase activity was assayed using a variety of substrates including barley β-glucan, yeast glucan, pachyman, carboxymethyl cellulose (CM-cellulose), dextran, pustulan, colloidal chitin, laminaritetraose and cellobiose. For the slightly soluble or insoluble barley β-glucan, pachyman, CM-cellulose and pustulan, 10 mg aliquots of each substrate were soaked in 100 μL ethanol (100%) and mixed with 800 μL dd-H 2 O. The mixtures were incubated in a water bath at 100˚C until the substrates were dissolved (barley β-glucan and pustulan) or formed sol like substance (CM-cellulose and pachyman) in 1 mL volumes of dd-H 2 O after cooling. The dinitrosalicylic acid (DNS) method [60] with slight modification [61] was used to quantify enzymatic activity. Briefly, 2 μg purified BbEng1 in 10 μL 0.02 M phosphate buffer saline (PBS, pH 7.2) was mixed with 90 μL mixture containing 10 mg / mL glucan substrate in 50 mM sodium acetate-acid acetate (NaAc-Ac) buffer (pH 6.0), and incubated at 37˚C for 1 h. The mixture was heated at 100˚C for 5 min after 200 μL DNS buffer (182.0 g / L C 4 H 4 KNaO 6 , 6.3 g / L C 7 H 4 N 2 O 7 , 21.0 g / L NaOH, 5.1 g / L C 6 H 5 OH, 5.1 g / L Na 2 SO 4 ) [60] was added to terminate the reaction, and then cooled down on an ice bath for 3 min. The absorbance at 540 nm (OD 540 ) of the supernatant was photometrically determined with glucose (reducing sugar) used as the standard as described [62]. Alternatively, a glucose oxidase method was used for assay of BbEng1 enzyme activity to laminaritetraose and cellobiose by measurement of produced glucose levels according to the manufacturer's guides (Applygen Technologies Inc, China). One unit of β-glucanase activity was defined as the amount of enzyme required for a release of 1 μmol reducing sugar (glucose) per minute per μg of protein under the tested conditions. To reveal optimal temperature for BbEng1 activity, reaction mixtures of BbEng1 (2 μg) and yeast glucan (10 mg / mL) were incubated over a range of 20˚C to 90˚C at 10˚C intervals. To determine enzymatic thermostability, 2 μg BbEng1 in 10 μL 0.02 M PBS (pH 7.2) was incubated for 1 h over a range of temperatures, followed by assessing enzymatic activity using yeast glucan at 30˚C. The optimal pH of BbEng1 activity was determined at 30˚C using yeast glucan as substrate in different buffers at pH 3.0 to 6.0 adjusted with 50 mM NaAc-Ac buffer and 7.0 to 9.0 adjusted with 50 mM Tris-HCl buffer. Relative enzyme activity was evaluated with respect to the maximal enzyme activity (100%) in a tested pH range. All the experiments were repeated three times with different batches of purified BbEng1 samples.
TLC and HPLC analyses were performed to quantify the products of substrates hydrolyzed by BbEng1. For TLC, 10 μL 0.02 M PBS (pH 7.2), 2 μg BbEng1 and 90 μL NaAc-Ac buffer (pH 6.0) containing 1% (w/v) of each substrate, namely yeast glucan, barley β-glucan, pachyman, CM-cellulose, pustulan, dextran, colloidal chitin, laminaritetraose or cellulose, were mixed and reacted at 30˚C for 1 h, and the reaction was terminated by heating at 100˚C for 5 min. Aliquots (5 μL) of reaction supernatants were examined on high temperature (110˚C)-activated TLC Silica gel 60 F 254 plates (Merck, Germany). The solvent used was n-butanol: methanol: water 8: 4: 3 (v/v/v). The plate was air-dried after resolution, and developed by spraying with 5% concentrated sulphuric acid (v/v), 0.5% vanillin (w/v) in ethanol, followed by heating at 95˚C for 10-min color development [63]. Glucose, maltose, maltotriose, maltotetraose and maltopentaose were used as standards on the TLC plates. For HPLC, reaction supernatants were first filtered through a 0.22 μm filter. Aliquots (20 μL) of samples were analyzed using a sugar-H column (4.6 mm × 250 mm, 5 m) (Welch Co. Ltd., Shanghai, China) that was maintained at 80˚C and coupled to an RI-201H detector. The mobile phase was composed of 5 mM H 2 SO 4 at a flow rate of 0.6 mL / min.
For determining endo-vs exo-glucanase activities, periodate-oxidized laminarin and 4-nitrophenyl-β-D-glucopyranoside (p-NPG, Sigma) were used as substrates, respectively. Periodate-oxidized laminarin substrate preparation was performed as described previously [64] and the assay was carried out using the DNS method as described above. The substrate p-NPG was used to test for exo-1,3-β-glucanase using exo-1,3-β-D-glucanase (Megazyme) as a control, as described previously [65]. One unit of enzyme activity was defined as the amount of enzyme required for a release of 1 μmol p-nitrophenol per hour under standard assay conditions.
Cell wall chitin was determined as described [69]. Briefly, SDS-extracted cell walls (~5 mg dry weigh) hydrolyzed with 6 N HCl at 100˚C for 17 h were evaporated at 65˚C and resuspended in water (1 mL). After 0.1 mL 1.5 N Na 2 CO 3 in 4% acetylacetone was added, the samples were incubated at 100˚C for 20 min and then mixed with 0.7 mL of 96% ethanol. The mixture was reacted 1 h with 0.1 mL of solution containing 1.6 g of p-dimethylaminobenzaldehyde in 30 mL of concentrated HCl and 30 mL of ethanol, followed by reading the absorbance at 520 nm. Glucosamine was used as a standard for estimation of chitin content. Cell wall chitin was also assayed by fluorescent labeled wheat germ agglutinin (WGA) labeled using cells fixed with 3% (v/v) formaldehyde, and by calcofluor (CFW) staining as described [68,70].
For TEM analysis, fungal cells (conidia, blastospores and hyphal bodies) were fixed, dehydrated, embedded (with Lowicryl K4M resin, Plano, Wetzlar, Germany) and sectioned (75 nm) as described previously [73]. Ultrathin sections were sequentially stained with 3% uranium acetate and lead citrate, followed by visualization for collection of TEM images. Cell wall thickness was measured using the software Image-Pro Plus (Olympus). At least 30 samples per strain were measured as described previously [50].
BbEng1 distribution in fungal cells was examined using immuno-transmission microscopy as described previously [74]. Fungal hyphal bodies isolated from insect hemolymph infected with indicated stains were used to prepare samples for immuno-transmission microscopy with anti-BbEng1 polyclonal antibody (diluted 1: 30) and anti-Rabbit IgG (Whole molecule)-Gold (Sigma-Aldrich).

Insect bioassays and examination of intrahemocoel fungal cells
Virulence bioassays were using G. mellonella larvae and /or L. migratoria manilensis nymphs as indicated. Cuticle infection (topical bioassays) and cuticle-bypassing infection (intrahemocoel injection bioassays) using G. mellonella were initiated by spraying 1 mL aliquots of a 10 7 conidia / mL suspension on the surfaces of larvae in a Potter Precision Laboratory Spray Tower (Burkard Manufacturing Co., Ltd., UK) or injecting 2 μL of a 5 × 10 6 conidia / mL suspension into the larvae (G. mellonlella) or second proleg (locusts) with a microinjector (SPLab01, Baoding Shenchen Precision Pump Co., Ltd., China), respectively. Controls were treated with buffer alone (physiological saline solution). Each assay per strain included three replicates of 30-40 insects with the entire experiment repeated three times with different batches of fungal cells and insects. For M. acridum strains cuticle infection was assayed on L. migratoria manilensis nymphs by spotting 5 μL of a 10 7 conidia / mL suspension (in liquid paraffin oil) on the pronotum of each nymph as previously described [68]. Three groups of 20 locusts were included in each assay per strain/repeated twice. Controls were treated with liquid paraffin oil only. Mortality was recorded every 12 h. The survival data were plotted as Kaplan-Meyer curves. The mean lethal time to kill 50% of targets (LT 50 ) was calculated using SPSS 17.0 program.
Fungal proliferation in the hemocoel of G. mellonella (i.e., production of hyphal bodies) was examined from a subset of larvae infected through either topical infection or intrahemocoel injection as indicated. Larvae were bled at 24, 36, 48 and 60 h post-treatment and fungal development was monitored microscopically as well as via quantification of fungal cells

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction through qPCR analysis of by amplification of 18S rRNA sequences with primer pair 18S1 / 18S2 (S2 Table), as described previously [12,53].

Statistical analyses
All data from three independent replicates of experiments were statistically evaluated based on one-way analysis of variance (ANOVAs) and LSD test. The Student's test (t-test) was carried out to compare means between two groups or treatments. The SPSS 17.0 program were used in the analysis. Insect bioassay replicates and differences between groups were analyzed using a log rank test with GraphPad Prim 8.0.2 program.  Fig. Cell wall β-1,3-glucan and chitin contents. A. β-1,3-glucan and chitin contents in mycelia cultured in 1/4 SDY for 48 h. β-1,3-glucan levels were determined using an aniline blue assay as detailed in Methods section. Cell wall chitin contents were assayed with a method using HCl (6 N) hydrolyzed the SDS-extracted cell wall as detailed in Methods section. B. Detection of relative levels of cell wall β-1,3-glucan and chitin in blastospores and conidia (Scale: 5 μm). Cell wall β-1,3-glucan was labeled with monoclonal β-1,3-glucan specific antibody and goat anti-mouse IgG-FITC (Proteintech) after fixing cells with 3% (v/v) formaldehyde. Chitin was labeled using FITC-wheat germ agglutinin (WGA) after fixing cells with 3% (v/v) formaldehyde, as well as CFW directly stained cells. The fluorescence of the stained cells was observed using a confocal microscopy. C. Mean fluorescence intensities in (B) were quantified densitometrically using the ImageJ software. Error bars denote st. dev. (SD) and values with different letters indicate statistically significant differences from different treatments (P < 0.01 in LSD test). against different substrates and its endo-and exo-β-1,3-glucanase activity assay. A. TLC analysis of the released carbohydrates from BbEng1 (2 μg) hydrolysis reactions with the indicated polysaccharides at pH 6.0 and 30˚C for 1 h. G1 to G5 refer to the standard maltooligosaccharides: glucose, maltose, maltotriose, maltotetraose and maltopentaose. B. HPLC analysis of the released carbohydrates from BbEng1 (2 μg) hydrolysis reactions at the same conditions as those in (A). Polysaccharides include yeast glucan, barley β-glucan, pachyma, pustulan, CM-cellulose, dextran, laminaritriose, cellobiose, cuticle and colloidal chitin (1%, w/v). C. Assay of endo-β-1,3-glucanase and exo-β-1,3-glucanase of BbEng1 using periodate-oxidized laminarin and 4-Nitrophenyl-β-D-glucopyranoside (p-NPG) as substrates, respectively. iBbEng1 and exo-1,3-β-D-glucanase (Megazyme) (EXG) were uses as controls. The enzyme activity assay was detailed in Methods

PLOS PATHOGENS
A secreted glucanase affects cell wall structure during the fungal-insect interaction