Bombyx mori (Nistari), Helicoverpa armigera larvae, and Drosophila melanogaster adults (w¹¹¹⁸, PPO1^Δ, PPO2^Δ and PPO1^Δ, PPO2^Δ) were reared as described (15–17). The strains of fungal pathogens B. bassiana ARSEF 2860, M. robertsii ARSEF2575, and its mutant ΔMBZ1 were routinely maintained for the preparations of spore (conidia), hyphae, and blastospore as described (18). The yeast strains of Saccharomyces cerevisiae and Candida albicans were also routinely cultured for spore preparations (19). Spores (Chlamydospores) of C. albicans were collected (19). Escherichia coli (BL21) was cultured in lab. Micrococcus luteus (Sigma) was suspended in buffer for pull-down assay.

Spores of different fungal species were suspended in Tween 20 (0.05%, v/v) and centrifuged at 13,000× g for 2 min. Spores were then suspended in a binding buffer (50 mM Tris, 100 mM NaCl, pH 8.0) and washed three times. After last centrifuge, the spores were resuspended in binding buffer and the cell density was adjusted to 1 × 10⁹ cells/mL. The cells of Micrococcus were suspended in binding buffer to 100 µg/µL and the optical density (OD₆₀₀) was determined after proper dilution. The cells of E. coli were collected, washed, and resuspended in binding buffer to the same optical density (OD₆₀₀) as Micrococcus. Approximately, 1 µg of purified PPOs (BmPPO, HaPPO, rPPO1, rPPO2, and rPPO1-GFP) or 20 µL molting fluid (20 µg/µL with phenylthiourea (PTU) added to inhibit melanization) was mixed with approximately 1 × 10⁷ spores of different species or 1 × 10⁷ Micrococcus or E. coli in a 50-µL system (adjusted using the binding buffer). The above mixture was incubated by rotating at 4°C for 15 min. After incubation, fungal spores and bacteria were centrifuged at 13,000× g at 4°C for 2 min. The supernatant was used as the flow-through for Western blot assay or phenoloxidase (PO) activity assay. The cell pellets of different microorganisms were suspended in 100 µL washing buffer (100 mM Tris, 500 mM NaCl, pH 8.0), centrifuged (13,000× g at 4°C for 2 min) and washed for three times. During the second wash, the mixtures were transferred into a new tube to minimize protein contamination on the tubes. The washed cells were prepared for various purposes. The pelleted cells were suspended in 10 µL 1×SDS loading buffer and boiled for 10 min as elution. After centrifugation, the supernatant containing the eluted proteins was loaded for Western blot assay. Unless otherwise indicated, after binding assay using rPPO1 or rPPO2, approximately 8% and 16% of flow-through and elution solution treated as described above were loaded for each lane, respectively, in this work. To observe rPPO1-GFP fluorescence, pellets were suspended again in binding buffer and the fluorescence of spore was observed under microscope (Olympus BX51). To determine PO activity, the pellets were suspended in 200 µL of 10 mM dopamine prepared in 10 mM Tris buffer (pH 7.0) and incubated for 10 min. The same concentration of dopamine was added into the flow-through for PO activity assay. After the reaction, the mixture was then added with 10 µL of saturated PTU to stop PO activity and PMSF to a final concentration of 1 mM to inhibit protease activity. After centrifugation at 13,000× g for 2 min, the value of OD₄₉₀ of the supernatant was recorded. The exact amounts of PPO and/or PO cannot be determined in the flow-through and on the spores here. Thus, in this case, one unit of PO activity was defined as ΔOD₄₉₀/min = 1. To detect PO activity after being activated by the purified B. bassiana protease, one unit of PO activity was defined as ΔOD₄₉₀/min/μg = 0.001.

To determine the binding potential of insect PPOs with the chitin and β-1,3-glucan, commercial chitin (NEB) and β-1,3-glucan, laminarin (Sigma) from Laminaria digitata usually used for invertebrate PPO activation (20, 21), were constituted in a binding buffer at room temperature for 15 min, and washed twice with the binding buffer. An amount of 0.1 µg of rPPO1 or rPPO2 was mixed with 2 mg chitin or β-1,3-glucan, and the samples were rotated at 4°C for 15 min. As a control, an amount of HaPPO (1 µg), BmPPO (1 µg), or GFP (1 µg) was mixed with 1 mg chitin or β-1,3-glucan. Chitin and β-1,3-glucan were washed, respectively, following the procedure used for spores. The bound protein was eluted exactly as spores were for Western blot assay. The flow-through was loaded to monitor the unbounded PPOs.

To identify the targeted binding components, B. bassiana spores were treated in the following ways, and then used for binding assays. To remove spore surface proteins, spores were suspended and incubated in the binding buffer containing 0.5 µg/µL Protease K (Roche) at room temperature for 12 h. High concentration of HCl can hydrolyzate proteins (22). In order to remove all proteins, spores were also incubated in 37% HCl at room temperature for 12 h. To remove chitin component, spores were suspended and incubated in the binding buffer (200 µL) containing 1 µg/µL chitinase (Roche) and 1-mM phenylmethylsulfonyl fluoride (PMSF) (to inhibit protease contamination in this product according to our assay). To remove β-1,3-glucan component, spores were suspended and incubated in 200 µL of 0.2 M of citric acid–sodium citrate buffer (pH 5) containing 10 µg/µL glucanase (Sigma) and 1 mM PMSF at 37°C for 2 h. After these separate treatments (37% HCl, chitinase, and glucanase), spore samples were separated from chemical and washed in binding buffer for three times. Subsequently, the samples were incubated with rPPO1 or rPPO1-GFP as described above. In order to examine the enzyme hydrolysis effect, the dye Calcofluor White (Sigma) was applied for chitin staining (18), and β-1,3-glucan was detected by using Aniline Blue Fluorochrome (Biosupplies) (23).

Competition between the fragment PPO1-IP (PPO2-IP) and rPPO1 (rPPO2) to bind to spores was performed as follows. Approximately, 5 µg of the peptide fragments PPO1-IP (⁵¹⁸NQALNLEEQR⁵²⁷) and PPO2-IP (⁶²²CSDAASYCGVR⁶³²) were suspended in the binding buffer and separately preincubated with 1 × 10⁷ spores in a 50-µL system by rotating at 4°C for 15 min. After that, rPPO1 (0.06 µg) and rPPO2 (0.5 µg) were separately added to each tube containing the corresponding competitor peptide for another 15 min. As control, the peptide fragments PPO1-IP and PPO2-IP were not preincubated with spores. The above mixtures were centrifuged to obtain the flow-through solution, and the spores were washed as described above and proteins were eluted for Western blot assay.

To identify the spore-secreted proteases, 1 × 10⁷ spore were suspended and agitated in 50-µL binding buffer at room temperature for 12 h. The supernatant was collected and concentrated for analysis with an 8% native gel containing gelatin (1 mg/mL) (Gelatin Zymography) as described (24). The gel was then incubated in a binding buffer at 37°C for 2 h followed by Coomassie blue staining and destaining. Another native gel without gelatin was run in parallel for separating protease. The corresponding position of the white band on the gelatin gel was cut from the native gel (without gelatin) and treated for liquid chromatography–mass spectrometry (LC–MS)/MS assay as described (25). The genome database of B. bassiana was used for the search of fungal proteins (26).

Full-length cDNA of BPS8 (BBA_00319) (26) was amplified from a cDNA library and cloned into a pET28 vector and transformed into E. coli BL21 (see Table S1 in Supplementary Material for primer information). BPS8 was induced to express and purified as described (27). Purified BPS8 was reconstituted in Tris buffer (10 mM, pH 7.0) and the concentration was determined. Different amount of BPS8 was incubated with 1 µg recombinant PPO1 (rPPO1) in a 20-µL system prepared in the binding buffer for 10 min, then the mixture was added with PTU to stop PO activity and PMSF (1 mM) to inhibit protease activity. PO activities were assayed by addition of 200 µL of 10 mM dopamine and incubating for 10 min before recording the OD₄₉₀ values. The reaction sample was also loaded for Western blot assay to demonstrate rPPO1 cleavage.

To determine the contribution of BPS8 to activate insect PPOs or fungal virulence, the gene was deleted in B. bassiana by homologous replacement using the Agrobacterium-mediated transformation (18). To complement the gene deletion, full open reading frame of the gene together with its promoter region was amplified and cloned the binary vector to transform ΔBPS8 (18). Drug resistant mutants were selected and verified by polymerase chain reaction.

To determine the effect of PPO-mediated melanization on fungal spore viability, the spores of B. bassiana were first treated with commercial melanin (Sigma, 10 µg/µL) suspended in a minimal medium at 28°C for 12 h and rotated at 200 rpm. After centrifugation at 1,000× g for 5 min, the saturated supernatant was transferred to a new tube and labeled as 100%. Different concentration of melanin was diluted in minimal medium to 25% and 50%, respectively. B. bassiana spores were suspended in different concentrations of melanin solution and rotated at 25°C for 12 h to count the germinating spores.

To mimic the melanization process of molting fluids upon contacting fungal spores, 2.5 µg rPPO1 or BSA (control) was incubated with 1.0 × 10⁷ spores suspended in binding buffer (total volume 50 µL) at room temperature for 15 min. Then, one aliquot was added with 150 µL of saturated l-3,4-dihydroxyphenylalanine (l-DOPA) (dissolved in a minimal medium) and 10 µL water, and the other aliquot was added with 160 µL minimal medium as control, and the third aliquot was added with 10 µL saturated PTU (to inhibit melanization) and 150 µL saturated l-DOPA. The spores were incubated by rotation at 25°C, and the germinating rates were determined 12 h post-incubation.

To determine the effect of PPO binding on spore hydrophobicity, the spores of B. bassiana (1.0 × 10⁷) were incubated with 2.5 µg of rPPO1 or rPPO2 at room temperature for 15 min. Hydrophobicity index was determined as using hexadecane described (18). To assay the spore adhesion ability after being bound by rPPO1 or rPPO2, spores were washed in a Tris buffer (50 mM, pH 8.0) three times and suspended in the same buffer to adjust the spore density to 1.0 × 10⁵ spore/mL. Aliquots of 100 µL spore suspension were loaded into a six-well plastic plate (Corning) for 1 h and then washed to determine the ratio of spore binding to the hydrophobic surface (26).

B. bassiana spores (1.0 × 10⁸) were incubated with 2.5 µg of rPPO1 or rPPO2 in 0.05% Tween 20 (v/v) at room temperature for 15 min. Topical infection of silkworm larvae (on day 2 of 5th larval stage, V-2, 15 insects for each replicate) or wild type (WT) and mutant (PPO1^Δ and PPO2^Δ) Drosophila female adults (3 days after eclosion; 20 insects for each replicate) were performed by immersion in the spore suspension for 30 s (28). In another experiment, at 40 h post the topic infection by spores preincubated with rPPO1 or rPPO2 or nothing as described above, the melanized larvae were observed and counted. Hemolymph of each larva was sampled, and then 25 µL hemolymph was diluted for smearing one Potato Dextrose Agar plate. The numbers of hyphae colonies were counted and compared. To determine the importance of molting-fluid residues in inhibiting B. bassiana infection, silkworm larvae were washed immediately with sterile water for 2 min after the third and fourth ecdysis. The washed larvae were dried on paper, and the washed and unwashed larvae were placed side by side and sprayed with spore suspension (1.0 × 10⁷/mL). The insects were fed in a box containing wet paper to maintain moisture. To determine the contribution of BPS8 to fungal virulence, insect bioassays were conducted using the spores of WT, ΔBPS8 and gene-rescued mutant (1.0 × 10⁸, respectively) for topic infections of the silkworm larvae (V-2, 15 insects for each replicate). Approximately, 1.0 × 10⁵ spores of WT, ΔBPS8, and gene-rescued mutant were injected into pupae on day 1 (P-1). On day 6 after injection, the pupae were compared and imaged. For all bioassays, each experiment was repeated three times and mortality was recorded every 12 h. The median lethal time (LT₅₀) was calculated as described (18).

Rabbit polyclonal antibodies against the silkworm PPO (1: 5,000) (29), 30Kc19 (1: 3,000) (30), M. sexta βGRP-2 (1: 2,000) (31), Drosophila PPO1 (1:5,000) (27), and mouse polyclonal antibody against Drosophila PPO2 (1: 1,000) were used as the primary antibodies for Western blot analyses. Antibody binding was visualized by a color reaction catalyzed by alkaline phosphatase (AP) conjugated goat anti-rabbit/mouse IgG (1:10,000; Chemicon), or visualized by enhanced chemiluminescence catalyzed by horseradish peroxidase (HRP) using Pierce ECL Western blot substrate (Thermo).

We showed before that the proteins identified in insect molting fluids contribute to the antimicrobial activities and ecdysis regulation (5). After ecdysis, molting fluid dries quickly on the cuticle surface of newly molted silkworm B. mori, and we suspected that immunity-related proteins of B. mori like PPO (BmPPO) might remain bioactive on the insect cuticles for a certain period. To test this, we applied dopamine on the silkworm integument at different time points and found that PO activity could be detected up to 72 h after silkworm larvae emerged into the fifth instar (Figure S1A in Supplementary Material). In addition, the proteins of BmPPOs were detected in the molting fluids throughout this period by Western blotting analysis (Figure S1B in Supplementary Material). BmβGRP-3 was also detected in the washed solution (data not shown). After further washing, no BmPPOs could be detected (Figure S1C in Supplementary Material), and the PO activity was significantly reduced (Figure S1D in Supplementary Material).

B. mori like PPOs in the molting fluids were significantly removed when the newly molted larvae were washed using water within 2 h after ecdysis (Figures S1C,D in Supplementary Material). To test the protective effects of molting-fluid residues on fungal infection, silkworm larvae (within 2 h post the third time of ecdysis) were either washed or unwashed before the topical infection with the spores of B. bassiana by spraying method. We found that the larvae washed of molting-fluid residue became heavily melanized two days after fungal infection as compared with unwashed or uninfected insects (Figure 1A). If the numbers of visible melanized spots are compared, the differences are significant among the washed and unwashed treatments (Figure 1B). Similar results were obtained when using the silkworm larvae after the 4th ecdysis (Figures 1C,D). Usually there are not so many fungal spores in the insect living environment, from a mean of 0.2 CFU/m³ during 10 months per year in an Europe country to a range from 0 to 3.1 × 10³ CFU/m²/day in a forest in Japan (32), which is far below high spore concentrations used in this study (1.0 × 10⁷/mL) and in other labs. These data demonstrate that the molting-fluid residue is potently biologically active and likely functionally essential to protect larvae from fungal infection after ecdysis.

To identify the protein(s) functioning in the molting fluid, residue was incubated with either B. bassiana spore, E. coli (G⁻), or M. luteus (G⁺). Western blotting analysis indicated that BmPPOs could be eluted specifically from fungal spore but not from bacterial cells (Figure 2A). As a control, we found that a low molecular-weight protein (30 kDa) named Bm30Kc19 (NCBI Accession: ADQ89805) (30), previously identified in the molting fluids, could not bind to fungal spore (Figure 2B). To further compare the binding activity of insect PPOs to fungal spores, plasma PPOs were purified from the silkworm B. mori (BmPPO) and cotton bollworm H. armigera (HaPPO) (33), and used to treat the spore of B. bassiana. The results indicated that both plasma HaPPO and BmPPO could bind to spore (Figures 2C,D). Likewise, the plasma PPO1 and PPO2 from D. melanogaster were able to bind to fungal spores (Figures S2A,B in Supplementary Material).

Next, we produced recombinant PPO proteins in E. coli to examine their fungal binding activity. Consistent with a previous observation in Manduca sexta (34), recombinant BmPPO had no PO activity whereas Drosophila PPO1 and PPO2 expressed in E. coli exhibited PO activities (27). Thus, rPPO1 and rPPO2 of Drosophila were purified and incubated separately with the spore of B. bassiana and the cells of E. coli and Micrococcus. Confirmatively, rPPO1 and rPPO2 could bind to fungal spores but not to bacterial cells (Figures 2E,F). The addition of protease cocktail to inhibit the potential fungal proteases did not disturb rPPO1 and rPPO2 to bind to spores (data not shown). Our previous study showed that the GFP-fused rPPO1 (rPPO1-GFP) could exhibit both the PO activity (after being activated) and GFP fluorescence (35). The incubation of purified rPPO1-GFP with the cells of different microorganisms revealed no fluorescent signals associated with the cells of E. coli and Micrococcus (Figures 3A,D) but strong fluorescence associated with fungal spore surfaces (Figures 3E,F). Interestingly, rPPO1-GFP did not bind to the germinated hyphae and blastospores of B. bassiana (Figures 3G,H; Figure S3A in Supplementary Material). Recombinant rPPO1-GFP could also bind to the heat- or NaN₃-killed spores of B. bassiana (Figure S3B in Supplementary Material). The rGFP protein alone as a control did not bind to spore (Figures 3M,N; Figure S3B in Supplementary Material). We also tested the binding activities of rPPO1 to other fungal spore types and detected binding to the spores of insect pathogen M. robertsii but only weak binding to the sores of S. cerevisiae and the chlamydospores [also called as spores (19)] of C. albicans (Figures S4A,B in Supplementary Material).

The cell wall of B. bassiana spore consists mainly of chitin, glucans, and proteins (36), so we sought to determine the component of fungal cell wall targeted by PPOs. First, protease K was incubated with spores overnight to digest spore cell wall proteins (22, 37). We found that the resulting spores could be effectively bound by rPPO1-GFP (Figures 3I,J). When HCl was incubated with spores overnight to hydrolyze proteins as described (22), the cell wall debris was still able to bind rPPO1-GFP (Figures 3K,L). Thus, the proteins on spore cell wall are likely not the main binding targets for PPO.

We next tested the binding potential of PPOs against the cell wall components glucan and chitin. Both rPPO1 (Figures 4A,D) and rPPO2 (Figures 4B,E) bound to β-1,3-glucan and chitin, respectively. Likewise, the purified plasma HaPPO (see Lane P1 in Figure 2C without any treatment) bound to both β-1,3-glucan and chitin (Figures 4C,F). In contrast, when GFP was incubated with β-1,3-glucan and chitin, respectively, no binding was detected (Figure S5 in Supplementary Material). We then treated fungal spore with either chitinase or glucanase, and performed rPPO1-GFP binding assays and counterstaining with the fluorescence dyes Calcofluor White and Aniline Blue, which detect remaining chitin and β-1,3-glucan components, respectively. The results indicated that while untreated spores bound tightly to rPPO1-GFP (Figure 4G, a1,a2), spores treated with the chitinase or glucanase had strongly reduced GFP-signals (Figure 4G, d1,d2,g1,g2). Confirming the activity of glucanase, Aniline Blue staining could barely detect the β-1,3-glucan signals (Figure 4G, i1,i2), and the chitinase-treated spores had barely detectable staining-signals of both the chitin and β-1,3-glucans (Figure 4G, e1,e2,f1,f2), which could be due to the fact that chitin is localized inside the cell walls (36). When we performed the binding assay on a null mutant of M. robertsii (ΔMBZ1), which has enhanced accumulation of chitin on the cell wall (18), we found that rPPO1-GFP bound more strongly than the WT spores (Figures S4B, b1,b2 in Supplementary Material). Notably, the presence of carbohydrate like sugars (Figures S6A,B in Supplementary Material) or peptidoglycan from Gram-positive and Gram-negative bacteria (Figures S7A,B in Supplementary Material) in the mixture did not significantly influence rPPO1 or rPPO2 binding to spores. We also notice that Lipopolysaccharide did not disturb the binding (data not shown). rPPO1 or rPPO2 was pretreated with either chitin or β-1,3-glucan. rPPO1 and rPPO2 in the flow-through after removing chitin or β-1,3-glucan still bound to spores without significant difference (Figures S8A,B in Supplementary Material). In addition, rPPO1 or rPPO2 did not significantly disturb rPPO1-GFP to binding to spores (Figures S9A,B in Supplementary Material). These data indicate that rPPO1 and rPPO2 can bind to the spores via cell wall components chitin and β-1,3-glucan.

We next sought to identify the domain(s) in the rPPO1 and rPPO2 proteins that mediate interactions with fungal cell wall components. To determine the binding site, the proteins rPPO1-GFP, rPPO1, and rPPO2 were individually incubated with fungal spores in buffer for 4 h and then treated with trypsin for 4 h (Figure S10 in Supplementary Material). When bound fragments were then eluted for LC–MS/MS assay, the trypsin fragments ⁵¹⁸NQALNLEEQR⁵²⁷ from PPO1 (named as PPO1-IP) (Figure S11A in Supplementary Material), and ⁶²²CSDAASYCGVR⁶³² (named as PPO2-IP) from PPO2 were identified (Figure S11B in Supplementary Material). The corresponding DNA sequences encoding these two fragments were synthesized and fused in frame with a GFP protein gene at either the N-terminus or C-terminus. The resulting fusion proteins r(PPO1-IP)-GFP, rGFP-(PPO1-IP), r(PPO2-IP)-GFP, and rGFP-(PPO2-IP) were expressed, purified, and then incubated with the spore of B. bassiana, respectively. Both Western blotting assays (Figures 5A,B) and fluorescence microscopy (Figure 5C, b1–e2) analysis confirmed that, relative to the controls, these GFP-fused fragments bound to spores more strongly than rGFP, but with a weaker signal than observed for full-length rPPO1-GFP (Figure 5C, g1,g2). When the peptide fragments PPO1-IP and PPO2-IP were preincubated with the spores, these peptides inhibited the binding activity of rPPO1 and rPPO2 (Figures 5D,E). We also performed an in silico molecular docking analysis of rPPO1 and rPPO2 based on the conserved structure of insect PPOs (38, 39). Consistently, both chitin and β-1,3-glucan could be docked predictively into cavities composed of the above identified fragments. Notably, these two cavities are not in the same surface areas of PPO1 and PPO2 (Figure S12 in Supplementary Material). Therefore, rPPO1 and rPPO2 bind to spores through specific peptide fragments. However, the integrality of PPO structure should be advantageous to binding.

Phenoloxidase activities were detected when spores were incubated with either rPPO1 or rPPO2 (Figures 6A,B). However, neither binding (Figures 2 and 3) nor PO activities were detected when the bacterial cells of E. coli or Micrococcus were incubated with rPPO1 or rPPO2. The spores of the insect pathogenic fungus Metarhizium anisopliae contain spore-secreted proteases that affect fungal virulence (40). We were curious whether a similar mechanism is involved in activation of PPOs by spore-secreted protease in B. bassiana. To examine this possibility, spores were suspended in a binding buffer for 12 h, and a native gelatin zymography analysis was performed. From this we identified one band containing protease activities (Lane 2 in Figure 6C), and a corresponding band in a clean gel (without gelatin but running at the same time) subjected to LC–MS/MS assay identified a member of the peptidase S8 family of proteases (BBA_00319, termed BPS8 for Beauveria peptidase S8) based on the eluted fragments (Table S2 in Supplementary Material). The BPS8 protease is similar to the subtilisin Pr1J of Metarhizium that contains an N-terminus I9 family of peptidase inhibitor domain (41). To determine the function of BPS8, the cDNA of this gene was cloned, expressed in E. coli, and the purified protein was verified to have protease activity (Lane 3 in Figure 6C). Deletion of BPS8 gene in B. bassiana using a gene-replacement protocol (18) resulted in the disappearance of the protease band in native gelatin zymography (Lane 1 in Figure 6C), and when rPPO1 was incubated with the spores of the WT and gene-deletion mutants (ΔBPS8), respectively, the PO activity became significantly lower (P < 0.0001) after incubation with the ΔBPS8 spores than with the WT spores (Figure 6D). The fact that the PO activity was significantly reduced but did not completely disappear in the ΔBPS8-treated sample suggested that additional unidentified protease(s) of B. bassiana (the bracket-labeled bands in lane in Figure 6C) could also contribute to PPO activation.

We next tested the effect of the protease BPS8 on PPO activity. To do this, purified protease was incubated with rPPO1 at different ratios for 10 min. While cleavage and degradation of rPPO1 were observed for the treatments with higher doses of BPS8 by Western blot assay (Figure 6F), surprisingly PO activity increased at lower doses of protease (Figure 6E). These data suggest that the spore-secreted protease(s) could be co-opted to activate PPOs for insect defense against fungal infection or deployed by pathogen to degrade host-produced PPOs to promote fungal infection. To determine the effect of BPS8 on fungal virulence, we performed insect bioassays using the spores of WT, ΔBPS8 and gene-rescued mutant (ΔBPS8:BPS8) for topic infections. The estimation of median lethal time (LT₅₀) values revealed that deletion of the gene could significantly (P = 0) reduce fungal virulence when comparing the ΔBPS8 (LT₅₀ = 3.5 ± 0.047 days) to the WT (LT₅₀ = 3.0 ± 0.061 days). No difference was observed between the WT and gene-rescued mutant (LT₅₀ = 3.0 ± 0.013 days) (Figure 6G). When identical numbers of spores were injected into each pupa (P-1) we found that there were fewer conidial spores on the pupal integuments from the gene-rescued mutant (ΔBPS8:BPS8) (inset in Figure 6G). Taken together, these data indicate that BPS8 is a virulence factor that can be either utilized by host to activate PPO at low concentration or by fungi to degrade PPO at the same time.

Melanin, including reactive intermediates of oxygen produced during the process of PO-mediated melanization, is toxic to both pathogens and hosts (42–45). To verify the toxicity of PO on spore viability, we performed spore germination assays using a commercial melanin. The results confirmed that the inhibition of spore germination occurred in a dose-dependent manner (Figure S13 in Supplementary Material). We next examined spore germination rates by incubation of fungal spore with either rPPO1 or rPPO2 with the addition of the PPO substrate l-DOPA and/or the PO inhibitor PTU. While the addition of l-DOPA to BSA (bovine serum albumin) treatment (control) reduced the rate of spore germination as compared with the mock control, an outcome that was likely due to the occurrence of l-DOPA self-oxidization after overnight incubation. Incubation with rPPO1 or rPPO2 could significantly (P < 0.05) inhibit spore germination caused by the addition of l-DOPA 12 h post-incubation (Figure 7A). The inhibitory effect of rPPO1 incubation was comparable to the commercial melanin treatment. Not surprisingly, the addition of PO inhibitor PTU could block the PO inhibition of spore germination (Figure 7A). rPPO1 had comparatively a higher inhibitory effect than rPPO2, consistent with the lower PO activity of rPPO2 after being activated (27, 46). When PPO was incubated with hyphae without being activated to induce melanization, there was no obvious influence on the hyphae growth (data not shown). These data demonstrate that the melanization induced by fungal infection can delay the spore germination.

We also performed assays to determine the effect of PPO pretreatments on spore characteristics and insecticidal properties. rPPO1 but not rPPO2 binding could significantly reduce spore hydrophobicity (Figures 7B,C), however pretreatments with both rPPO1 and rPPO2 significantly (P < 0.05) reduced spore adhesion ability to a hydrophobic surface (Figures 7D,E).

Based on these observations, we performed insect bioassays to compare PPO pretreatments of spore on fungal virulence using the silkworm larvae (V-2) and Drosophila adults (day 3) as described (18, 28). When spores were preincubated with buffer (control), the larvae melanized significantly at 40 h after topical infection (Figure 8A, a). However, if the spores were preincubated with rPPO1, melanization was reduced significantly (Figures 8A, b; Figure 8B). In contrast, the numbers of melanized larvae and hyphae in hemolymph in the rPPO2 treatment were significantly higher than in the rPPO1 treatment (Figures 8B,C). Topical infection by spores pretreated with rPPO2 induced melanization comparable to the control (Figure 8A, c; Figure 8B). At 40 h, the numbers of hyphae in the hemolymph of larvae infected by rPPO1-treated spores were significantly lower than that in the control (Figure 8C), but the differences were not significant between the control and rPPO2 treatment. It seems that insect PPOs can impair the virulence of B. bassiana upon contacting.

Bioassays using spores pretreated with rPPO1, rPPO2 or mock treated were performed on the silkworm larvae (Figures 8D,E). Based on the estimation of LT₅₀ values, preincubation of spores with rPPO1 (LT₅₀ = 3.933 ± 0.118 days) could significantly (χ² = 35.09, P < 0.0001) delay fungal infection compared with the usage of untreated spores (LT₅₀ = 2.944 ± 0.081 days) (Figure 8D). No significant delay was detected when rPPO2 was preincubated with spores (χ² = 0.2000, P = 0.6547) (Figure 8E). The different effects of rPPO1 and rPPO2 on the virulence of B. bassiana spores may be due to the low PO activity in rPPO2 (27, 46). However, in most insects, there are over two or even more PPO genes, and the corresponding PPO proteins can exist at the same time (47). When they work together, the influences should be stronger than each single one.

When the adults of WT Drosophila (w¹¹¹⁸) were used for infection, the difference in LT₅₀ values between the rPPO1-treated and untreated spores (LT₅₀ = 2.136 ± 0.075 and 4.203 ± 0.212 days, respectively) was significant (χ² = 78.63, P < 0.0001) (Figure 8F). For adults mutated for both PPO genes (PPO1^Δ and PPO2^Δ), LT₅₀ value for rPPO1-incubated spores (3.553 ± 0.159 days) was significantly different (χ² = 43.15, P < 0.0001) from the LT₅₀ value after treatment with unincubated spores (LT₅₀ = 2.225 ± 0.119 days) (Figure 8G). These data suggest that PPO-mediated ex vivo immunity is likely independent of hemolymph PPO and that pretreatment of fungal spores with PPOs impairs fungal virulence.