The wild type (WT) of B. bassiana ARSEF2860 (Bb2860) was obtained from the U.S. Department of Agriculture Entomopathogenic Fungus Collection (Ithaca, NY, USA) (14). Sabouraud dextrose agar with yeast extract (SDAY) (4% glucose, 1% peptone, and 1.5% agar plus 1% yeast extract) was used to routinely maintain fungal strains at 25°C. Escherichia coli DH5α (Invitrogen, Waltham, MA, USA) was cultured in a Luria-Bertani medium with necessary antibiotics for plasmid propagation. Yeast extract broth (YEB) medium (w/v: 0.5% sucrose, 1% peptone, 0.1% yeast extract, and 0.05% MgSO₄) was used to culture Agrobacterium tumefaciens AGL-1, and the resultant bacteria were used as donor cells in fungal transformation. Czapek-Dox agar (CzA) (3% glucose, 0.3% NaNO₃, 0.1% K₂HPO₄, 0.05% KCl, 0.05% MgSO₄, and 0.001% FeSO₄ plus 1.5% agar) was used as the chemically synthetic medium in screening transformants and phenotypic assays.

Domain annotation was conducted on the predicted proteome of B. bassiana (14) through the online portal SMART (http://smart.embl-heidelberg.de). All resultant proteins were named as B. bassiana LTA, and their similarities with Escherichia coli LTAs were performed with multiple- alignment analysis using MEGA version 7 as described previously (15).

Plasmid construction was performed as described previously (16). Briefly, the coding sequence of each gene was amplified with the respective primers (PxGFPF/PxGFPR) (X: BbLTAX) using Complementary DNA (cDNA) as the template. The resultant DNA fragments were fused to the 5′-terminus of green fluorescence protein gene in the vector pBbTEF-MCS-Gfp-sur (pBMGS) (16). The resultant p0380-X-gfp-sur was integrated into the WT strain via Agrobacterium-mediated transformation. Putative transformants were screened on the selection plates with chlorimuron ethyl (10 μg/mL) and confirmed under a fluorescence microscope. The conidia of transformants were inoculated in SDB (SDAY without agar) and incubated at 25°C for 2 days. Hyphal samples were collected from each culture and stained with 7-amino-4-chloromethylcoumarin (CMAC). Fluorescent signals were visualized under a laser scanning confocal microscope (LSCM).

Disruption mutants were generated using a method of homologous replacement coupled with a fluorescence reporter (17). The primer pairs PxUF/PxUR and PxDF/PxDR were used to amplify 5′- and 3′-fragments of the indicated gene, respectively. The resulting fragments were sub-cloned into the restriction enzyme sites (XmaI/BamHI and XbaI/HpaI) in p0380-GTB using the ClonExpress II One Step Cloning Kit (Vazyme Biotech, Nanjing, China), generating gene disruption vector. The full-length gene together with promoter region was amplified for each gene with primer pair PxHF/PxHR, and the obtained fragment was ligated into the plasmid p0380-sur, generating the complementation vector. The resulting vector was transformed into the gene disruption mutant with the Agrobacterium-based transformation method. Putative gene disruption and complementation strains were grown on plates included with phosphinothricin (200 μg/mL) and chlorimuron ethyl (10 μg/mL), respectively, and further identified via PCR analyses with primer pair PxJF/PxJR. The candidate strains were confirmed for their fluorescent signals under a LSCM. All primers are included in Supplementary Table 1 .

Fungal phenotypes, including vegetative growth, stress response, development, and virulence, were evaluated among the WT, gene disruption, and complemented mutant strains as described previously (18). Fungal strains were cultured on SDAY plates for 7 days at 25°C, and conidia were used as initial inocula in following assays. All experiments were repeated three times.

Vegetative growth: Conidial suspension (1 µL, 10⁶ conidia/mL) was inoculated on the CzA plates modified with various carbon and nitrogen sources. Colony diameter was examined at 7 days after incubation at 25°C. Carbon sources (final concentration, w/v) included glucose (3%), sucrose (3%), fructose (3%), trehalose (3%), olive oil (0.5%) linoleic acid (0.01%), and oleic acid (0.2%). Nitrogen sources (final concentration, w/v) included NH₄NO₃ (0.5%) and urea (0.5%).

Stress responses: Fungal responses to oxidative stress were determined on a CzA plate supplemented with 1 mM H₂O₂, 2 mM H₂O₂, Congo red (3 μg/mL), and Calcofluor white (CFW; 1 μg/mL). A droplet (1 µL) of conidial suspension (10⁶ conidia/mL) was placed on the plate and incubated at 25°C. The colony diameter was measured at 7 days after incubation. CzA plates without stress chemicals were used as control.

Fungal development: Conidiation capacity was determined on SDAY plates. Aliquots (100 μL of 10⁷ conidia/mL) were inoculated on SDAY plates and grown for 7 days at 25°C. Conidia in mycelial discs (5 mm in diameter) were rinsed into 0.02% Tween 80 solution. Cell concentration was quantified and used to calculate conidial yield as conidial number per square centimeter. In addition, the mycelia of the WT and gene disruption mutants were sampled at 4 days after incubation, respectively. The conidium-producing structures were stained with CFW and examined under a LSCM. Fungal development under submerged conditions was evaluated in SDB medium (SDAY without agar). Conidia were inoculated into SDB at the final concentration of 10⁵ conidia/mL and incubated for 3 days at 25°C with shaking. The blastospore concentration was determined, and blastospore yield was displayed as the blastospore number per milliliter of broth.

To examine fungal virulence, the last instar Galleria mellonella larvae were used as the bioassay insects, and each treatment involved 30–35 larvae. Conidial concentrations in suspension were adjusted to concentrations of 10⁵ and 10⁷ spore/mL, which were used in the intra-hemocoel injection and cuticle infection methods, respectively. For injection assay, a conidial suspension (5 μL, 10⁵ spores/mL) was inoculated into the host hemocoel with a syringe. In cuticle infection assay, the larvae were immersed in suspension for 15 s. All infected insects were reared at 25°C, and the mortality was recorded daily. The median lethal time (LT₅₀) was calculated by Kaplan–Meier method with log-rank test for determining the difference between the paired survival trends. In addition, fungal development in the host hemocoel was determined by quantifying the concentration of hyphal bodies in hemolymph at 3 days after infection.

Hemocyte encapsulation assay was performed as described previously (19). Briefly, aliquots of 5-μL conidial suspension were injected into the host hemocoel, and the infected hosts were reared at 25°C. The hemolymph was sampled everyday within 2 days after injection and then every 6 h after 2 days. Fungal cell interactions with host blood cells were recorded under a microscope.

Conidial lectin-binding traits were examined as previously described (20). The Alexa Fluor 488–labeled lectins included concanavalin A (ConA), Galanthus nivalis lectin (GNL), peanut agglutinin (PNA), and wheat germ agglutinin (WGA) (Vector Laboratories Inc., California, USA). Briefly, conidia were suspended in 0.02% Tween 80 and fixed in 3% volume per volume (v/v) formaldehyde for 30 min. The resultant conidia were re-suspended in buffer and labeled with lectin for 1 h in darkness. Unbound lectins were removed by washing. Fluorescent signals were quantified on a CytoFLEX LX flow cytometer (Beckman Coulter Life Sciences, Indianapolis, USA) at the excitation/emission wave lengths of 488/530 nm.

Real-time PCR analysis was used to reveal the responses of host hemocytes and performed as described previously (19). Conidia (500 cells) of the gene disruption mutant strains were injected into the G. mellonella larvae and then were reared at 25°C. Three days later, the larvae were bled at the hind legs, and the haemolyph was collected into 0.2 mL of anticoagulant solution [0.14 M NaCl, 0.1 M glucose, 26 mM citric acid, 30 mM trisodium citrate, and 10 mM EDTA (pH 4.6)]. All manipulations, including bleeding, centrifugation, and washing, were performed at 4°C. The hemocytes were separated from hemolymph by centrifuging and used to extract total RNAs using TRIzol^® Reagent (Sigma-Aldrich, Missouri, USA). cDNA was prepared from total RNA with a PrimeScript™ RT Reagent Kit (TaKaRa, Dalian, China). Real-time PCR was used to analyze the expression levels of genes involved in insect immune response. The 18S rRNA gene was used as an endogenous reference, and the relative expression levels were calculated by the 2^−ΔΔC_T method.

All measurements for the WT, gene disruption, and complementation strains were subjected to Student’s t-test, and the significance was confirmed at p < 0.05. Statistical analyses were performed with the software of GraphPad Prism 8 (GraphPad Software, Boston, MA, USA).

A domain annotation analysis indicated that there were 14 LTA proteins in B. bassiana genome, and each protein contained a domain of enterotoxin_A. The GenBank accession numbers for each protein were EJP68166, EJP61702, EJP67856, EJP67132, EJP67318, EJP62787, EJP68172, EJP63767, EJP61236, EJP70164, EJP64236, EJP61933, EJP66767, and EJP67352. Fourteen proteins were designated successively as BbLTA1 through BbLTA14. The full length and domain size varied among these proteins. In addition, BbLTA1, BbLTA5, BbLTA8, and BbLTA14 did not carry signal peptide, and the other proteins did. As shown in Figure 1A , these 14 proteins were grouped into different clusters, and no apparent relationship was noted between domain architecture and phylogeny.

On the basis of the characteristics of domain organization of BbLTA, five proteins (i.e., BbLTA1, BbLTA2, BbLTA3, BbLTA6, and BbLTA12) were selected as representative ones for functional studies. The N-terminus of the indicated protein was fused with GFP fluorescent protein, and the fluorescent signals indicated the distribution of the fusion proteins. As shown in the Figure 1B , the green fluorescence signals of BbLTA1, BbLTA3, and BbLTA12 were distributed on the cell periphery and vacuoles, whereas the green signals of BbLTA2 were exclusively observed in vacuoles and BbLTA6 were localized on hyphal septa. These results indicated that the BbLTA proteins exhibited diverse sub-cellular localizations.

To investigate the roles of BbLTAs, gene disruption and complementation strain were constructed as detailed in Materials and methods section and as shown in Supplementary Figures 1A, B . On SDAY plates (aerial condition), the conidia-forming structures of each strain were observed at 3.5 days after incubation. As shown in Figure 2A , the WT strain produced normal flask-shaped conidia-forming structures. The disruption mutants of ΔBbLta2 and ΔBbLta3 produced the conidia-forming structures with abnormal appearance. No noticeable changes were observed in the sporulating structures of ΔBbLta1, ΔBbLta6, and ΔBbLta12 strains, as compared with that of WT. However, these three disruption mutants displayed the significant reduction in conidial production [means ± standard deviation (SD)], when compared with that of the WT strain ( Figure 2A ). The conidial yield of ΔBbLta1, ΔBbLta6, and ΔBbLta12 were 3.58 ± 0.14, 5.61 ± 0.62, and 5.48 ± 0.44 (× 10⁷ conidia/cm²), respectively, as compared with that of the WT strain (8.59 ± 0.41 × 10⁷ conidia/cm²). The decrease degrees for these three mutants were 58.29%, 34.76%, and 36.22%, respectively. ΔBbLta2 and ΔBbLta3 produced 3.00 ± 0.11 and 4.56 ± 0.46 (× 10⁷ conidia/cm²), respectively. The complementation strains showed no significant difference in production, when compared with that of the WT strain ( Figure 2B ). These results indicated that all gene disruption mutants displayed the significant reduction in conidial production.

As shown in Figure 3A , gene disruption resulted in significant change in formation of the blastospore-producing structures; in particular, ΔBbLta1 barely formed structures for blastospore generation. The disruption of BbLTA genes had significant impacts on the blastospore production (means ± SD) ( Figure 3B ). The blastospore yield of WT strain was 1.36 ± 0.06 × 10⁸ blastospores/mL, and those of ΔBbLta1, ΔBbLta2, ΔBbLta3, ΔBbLta6, and ΔBbLta12 mutant strains were 0.57 ± 0.09, 1.08 ± 0.02, 1.04 ± 0.10, 0.93 ± 0.03, and 0.89 ± 0.09, respectively. The decrease degrees were 58.35%, 20.64%, 23.46%, 31.57%, and 34.77%, respectively. The complementation strains showed no significant difference with the WT strain.

To determine the BbLTA roles in nutrient utilization, the fungal growth was evaluated on different carbon or nitrogen sources. After an incubation of 7 days at 25°C, all disruption mutants displayed the reduced growth on a medium with linoleic acid as carbon source, when compared with the WT strain. On oleic acid plates, the disruption mutants of ΔBbLta2, ΔBbLta3, and ΔBbLta6 showed the slight reduction in colony diameter ( Figure 4A ). On media with glucose, sucrose, lactose, trehalose, fructose, and maltose as carbon sources as well as NH₄Cl, peptone, and gelatin as nitrogen sources, the disruption mutants did not displayed significant change in their colony diameter when compared with the WT ( Supplementary Table 2 ).

As shown in Figure 4B , on the CzA plates, the colony diameters of all disruption mutants were consistent with that of the WT strain. However, when adding 1 mM H₂O₂ into medium, the colony diameters of all mutants were significantly decreased when compared with that of the WT strain. As the H₂O₂ concentration increased to 2 mM, the colony diameter of ΔBbLta1 did not significantly differ with that of the WT strain; however, the other four disruptants still showed the reduced colony diameters when compared with the WT strain.

As shown in Figure 4C , on the medium supplemented with Congo red, the colony diameters of ΔBbLta1, ΔBbLta2, ΔBbLta3, ΔBbLta6, and ΔBbLta12 were 0.6 cm, 0.67 cm, 0.8 cm, 0.7 cm, and 0.8 cm, respectively. The decrease degrees were 50.00%, 44.44%, 33.33%, 41.67%, and 33.33%, respectively. In addition, ΔBbLta1, ΔBbLta3, and ΔBbLta12 strains exhibited the reduced resistance to cell wall stress induced by CFW.

Fungal development in the host hemoceol was determined at 3 days after infection. The WT and complemented strains generated plenty of yeast-like hyphal bodies. The in vivo blastospore yield (means ± SD) of ΔBbLta1, ΔBbLta2, ΔBbLta3, and ΔBbLta6 mutant strains was 2.67 ± 0.41, 3.73 ± 0.90, 4.47 ± 0.52, and 3.80 ± 0.59 (× 10⁶ blastospores/mL), respectively ( Figure 5A ). The decrease degrees were 57.89%, 41.05%, 29.47%, and 40.00%, respectively, when compared with that of the WT strain (6.33 ± 0.57 × 10⁶ blastospores/mL). ΔBbLta12 displayed a slight reduction in blastospore production when compared with that of the WT, but the difference was not statistically significant. No significant difference in blastospore production was observed between the WT and complementation strains. These results indicated that the loss of five BbLTA proteins resulted in the impaired pathogenic growth in the host hemoceol.

Conidial virulence against the greater waxmoth was assessed using two methods: intrahemocoel injection and cuticle inoculation. As shown in Figures 5B, C , the survival percentage decreased over time. In both bioassays, the survival curve for each disruptant was significantly lower than that of the WT strain ( Supplementary Table 3 ). The LT₅₀ (median lethal time) for the WT strain was 3.58 days and 6.00 days in the injection and cuticle inoculation methods, respectively. In the intrahemocoel injection assay, the LT₅₀ values for the five disruptants (ΔBbLta1, ΔBbLta2, ΔBbLta3, ΔBbLta6, and ΔBbLta12) were 4.5 days, 4.2 days, 4.0 days, 4.0 days, and 4.0 days, respectively. In the cuticle inoculation bioassay, the LT₅₀ values were delayed by 1.7 days, 1.7 days, 0.7 days, 1.7 days, and 1.0 days, respectively, when compared with that of the WT strain Figures 5D . The results indicated that the loss of BbLTA genes significantly resulted in the reduced virulence of B. bassiana.

Following fungal invasion, host hemocytes congregated and formed melanic dots (21). As shown in Figure 6 , up to 24 h post-infection (HPI), there were no significant differences in the hemocyte response between each disruptant and the WT strain. However, at 48 HPI, the WT and complemented strains emerged from a cluster of hemocytes and started to produce hyphal bodies, whereas the gene disruption mutants evaded from the host hemocytes until 60 HPI. By 54 HPI, the WT and complemented strains produced numerous free-floating hyphal bodies, whereas the gene disruption mutants produced hyphal bodies until 66 HPI. Nonetheless, the hyphal bodies of the WT and complemented strains underwent differentiation into hyphae at the same time point.

Flow cytometry analyses revealed that the disruption of BbLTA genes altered the lectin-binding traits of conidia (19). Specifically, ΔBbLta1, ΔBbLta2, ΔBbLta3, ΔBbLta6, and ΔBbLta12 mutant strains exhibited a significantly increased ability to bind WGA, whereas the ΔBbLta1, ΔBbLta2, and ΔBbLta3 mutant strains showed an enhanced binding ability to PNA. In addition, ΔBbLta2 displayed an increased ability to bind GAL, and the ΔBbLta2, ΔBbLta3, ΔBbLta6, and ΔBbLta12 mutant strains showed an enhanced binding ability to ConA ( Figure 7A ).

Moreover, RT-qPCR analyses unraveled that the transcriptional responses of hemocytes challenged by fungal cells. At 3 days after infection, the expression of 12 antimicrobial peptide genes was significantly higher in hemocytes infected by disruptants when compared with those challenged by the WT strain ( Figures 7B, C ). These results indicated that the loss of BbLTA genes resulted in the altered transcriptional responses of host hemocytes.