The bioinsecticide Botanigard (Botanigard^® ES: Arysta LifeScience, Tokyo, Japan) was stored at 4°C before use according to the instruction (1-year shelf life). For treatments, Botanigard was diluted with distilled water to the desired test concentrations and allowed to stand at room temperature for 0.5 to 1 h and then used for experiments according to the manufacturer’s instruction. This formulation contains petroleum distillates in the inert ingredients and B. bassiana strain GHA as an active ingredient.

Potted seedlings of cucumber (Cucumis sativus L. cv. Sharp 1; Saitama Gensyu Ikuseikai, Saitama, Japan) were grown in a growth chamber at 25°C with a 16-h light/8-h dark photoperiod. Inoculum of cucumber powdery mildew pathogen P. xanthii strain pxB (Fukino et al., 2008), kindly provided by Dr. Koichiro Shimomura (National Agriculture Research Organization, Japan), was maintained on potted cucumber plants grown under the same conditions. Four to five leaves of 3- to 4-week-old cucumber plants (N = 10) were sprayed with a 1,000-fold dilution of Botanigard (1 × 10⁷ spores/mL, ca. 2 mL per plant) from 3 days before to 3 days after plants, and were inoculated with a spore suspension of the P. xanthii inoculum (1 × 10⁴ spores/mL, ca. 2 mL per plant) with three independent replications. The upper surface of all leaves of each plant was covered thoroughly with a spray of Botanigard, the inoculum, or distilled water (mock control).

Disease severity of leaves was rated 12 days after inoculation of powdery mildew on a scale of 0–4: 0, no symptoms; 1, less than 5% of leaf area diseased; 2, 5%–24% diseased area; 3, 25%–49% of leaf area diseased; and 4, more than 50% of leaf area diseased based on the recommendations in the Fungicide Evaluation Manual of the Japan Plant Protection Association for field trials (www.jppa.or.jp/test/04.html). The severity ratings for individual leaves were then used to calculate disease severity as 100[(1n1 + 2n2 + 3n3 + 4n4)/4N], where N = total number of tested leaves and n1 to n4 = number of leaves scored with each respective score (1–4). The mean disease severity on the mock plants was set as 100, and the relative disease severity was calculated on the basis of the mean (± SD) (more than three independent replications).

For microscopic observations of the fungus after treatment, as described above, the cucumber leaves were sprayed with a 1,000-fold dilution of Botanigard, followed with the powdery mildew spore suspension but with only ca. 0.5 mL per leaf. After 3 days, leaves were cut into 1-cm squares and stained with 0.01% aniline blue (w/v) mixed with lactophenol solution (1:1:1:1 lactic acid, TE-saturated neutral phenol, glycerin, and distilled water). The leaves were boiled for a few seconds and then destained with 5% chloral hydrate solution (w/v). At least 200 spores of P. xanthii at five sites on each of three leaves (15 sites total) were observed for germination and hyphal growth using a microscope IX73 (Olympus, Tokyo, Japan) equipped with a U-MWU filter (Olympus). The percentage of spore germination and hyphal growth was calculated on the basis of the area of stained hyphae in 50 mm² using ImageJ version 1.51 (imagej.nih.gov/ij).

Cucumber (C. sativus), melon (Cucumis melo L.), tomato (Solanum lycopersicum L.), and strawberry (Fragaria × ananassa Duch.) were grown in the greenhouse, and eggplant (Solanum melongena L.) was grown in an open field at different locations as described in Supplemental Table 1 (N = 6 for cucumber, N = 3 for tomato and strawberry, and N = 2 for melon and eggplant; three independent replicates in each trial). Leaves and stems of 3- to 6-month-old plants of each species above grown in the greenhouse or the field were sprayed with a 1,000-fold dilution of Botanigard 3 to 10 times at about 1-week intervals. Untreated plants served as the control. Plants were inoculated with the powdery mildew species of the respective vegetables (P. xanthii for cucumber and melon, Pseudoidium neolycopersici for tomato, Podosphaera aphanis for strawberry, and Sphaerotheca fuliginea for eggplant) either naturally or by planting plants infected with the powdery mildew in the same area. Disease severity was scored 6–13 days after the last application of Botanigard as described above.

In trials using whiteflies, 60–72 female adult whiteflies (Bemisia tabaci, B biotype) were released in the greenhouse where 1-month-old tomatoes were growing. Leaves and stems of tomato plants were sprayed with a 1,000-fold dilution of Botanigard 10 times at about 1-week intervals (detailed in Supplemental Table 1 ). The number of young and old larvae and of adult whiteflies on 16 small leaves of four plants was counted with a loupe, and was compared between Botanigard-treated and untreated tomato plants (six independent replicates).

B. bassiana strain GHA, the active ingredient in Botanigard, was grown on sabouraud dextrose yeast extract agar (glucose, 20 g; peptone, 2 g; yeast extract, 2 g; agar, 15 g/L) for 2 weeks at 25°C. Conidia were collected in sterile distilled water containing 0.05% (v/v) Tween 20 and filtered through sterile cheesecloth to remove hyphae. The suspensions were washed twice by centrifugation for 5 min at 2,500 × g. Concentrations of conidia were determined using a hemocytometer.

We assessed the disease severity of powdery mildew on the leaves treated with GHA. Four to five leaves of 3- to 4-week-old cucumber plants were sprayed with a spore suspension of GHA at 1 × 10⁷, 1 × 10⁸, or 1 × 10⁹ spores/mL, ca. 0.5 mL per leaf, and then with a spore suspension of P. xanthii (1 × 10⁴ spores/mL, ca. 0.5 mL per leaf) to cover the upper surface of all leaves as described above. Plants were grown in a growth chamber as described above. Plants treated with distilled water or a 1,000-fold dilution of Botanigard served as the control. Disease severity was assessed 12 days after inoculation of powdery mildew as described above.

At 24 h after inoculation, leaves were stained with 0.01% aniline blue (w/v) mixed with lactophenol solution and destained as described above. The number of hypersensitive response (HR)–like cell death was counted under 100 spores or germ tubes on the stained leaves using a fluorescence microscopy IX73 optical microscope equipped with UV light and a U-MWU filter (Olympus). The percentage of spore germination was calculated for 100 spores.

To clarify whether systemic resistance was induced by B. bassiana GHA, one-half of the each detached leaf surface was sprayed with a spore suspension of B. bassiana (1 × 10⁹ spores/mL, ca. 0.2 mL per leaf) and the other with water, and, then, the entire leaf was inoculated with a spore suspension of P. xanthii (1 × 10⁴ spores/mL, ca. 0.5 mL per leaf). Ten leaves were tested for each treatment. Disease severity relative to that on mock control plants was calculated on the basis of the mean (± SD) for three independent replications as described above.

To elucidate the function of B. bassiana GHA as a PGPF, roots of 3- to 4-week-old cucumber plants (N = 10) were dipped into a spore suspension of GHA (1 × 10⁹ spores/mL) or sterile distilled water for 30 s and then planted into sterile soil in pots. Above ground parts and root biomass (fresh mass) were weighed after 2 weeks.

All data were analyzed using a Mann–Whitney U-test in the program R version 4.0.3 (R Core Team, 2020).

Cucumber plants were sprayed with a spore suspension of GHA (1 × 10⁹ spores/mL) and then inoculated with a spore suspension of P. xanthii (1 × 10⁴ spores/mL) as described above (three independent replications). At 24 h after inoculation, total RNA was extracted from leaves using the NucleoMag RNA kit (MACHEREY-NAGE, Duren, Germany) according to the manufacturer’s instructions. cDNA libraries (150-bp paired-end reads) were prepared using the NEBNext Poly(A) mRNA Magnetic Isolation Module kit (for PolyA selection) and NEB NEXT Directional Ultra RNA Library Prep Kit for Illumina (for strand-specific library) (New England Biolabs, MA, USA) and sequenced using a NovaSeq 6000 platform by Rhelixa (Tokyo, Japan). Sequence data were deposited into the DNA Data Bank of Japan (DDBJ) database (accession PRJDB15569).

Paired-end RNA-seq reads were trimmed using Trimmomatic version 0.38 (Bolger et al., 2014) and then mapped to the cucumber (Chinese Long) v3 genome (cucurbitgenomics.org/organism/20) using HISAT2 version 2.1.0 (Kim et al., 2015). Read counts, fragments per kilobase of exon per million reads, and transcripts per million were calculated using featureCounts version 1.6.3 (Liao et al., 2013). Differentially expressed genes (DEGs) were determined using DESeq2 version 1.24.0 (Anders and Huber, 2010) at p <  0.05 after Benjamini–Hochberg adjustment and log₂|fold change| > 1. The DEGs were then annotated for Gene Ontology (GO) enrichment using TBtools and p <  0.05 (Chen et al., 2020a).

Cucumber plants sprayed with a spore suspension of GHA and P. xanthii were prepared as described above. After 24 h, leaves were cut into 1-cm squares and examined for germination of P. xanthii spores as described earlier using a stereomicroscope SZ61 (Olympus). SA and JA were extracted and quantified using liquid chromatography–mass spectrometry as described previously (Seo et al., 2007).

For ethylene (ET) analyses, 10-day-old cucumber seedlings were sprayed with a spore suspension of GHA and powdery mildew as described and planted in soil in 50-mL plastic tubes. After 24 h, a sample was withdrawn from the headspace using a syringe and injected into a gas chromatograph (GC-8A, SHIMADZU, Kyoto, Japan) equipped with an alumina column (Porapak Q 50/80; Shinwa, Kyoto, Japan) and a flame ionization detector (Seo et al., 2007). Data were analyzed using a Mann–Whitney U-test.

Seedlings of tomato Solanum lycopersici L. cv. Moneymaker (MM) and its transgenic line expressing the NahG gene (MM-NahG), a bacterial gene encoding salicylate hydroxylase that converts SA to catechol (Brading et al., 2000), which are equally susceptible for powdery mildew of tomato P. neolycopersici, were grown in a growth chamber. Seeds of MM-NahG seeds were kindly provided by Professor Tsutomu Arie (Tokyo University of Agriculture and Technology, Japan). Discs (8 mm in diameter) were excised from leaves of 2-week-old tomato plants, placed on water agar in petri dishes, and then sprayed with a spore suspension of GHA (1 × 10⁹ spores/mL, ca. 0.1 mL per leaf disc) and with tomato powdery mildew P. neolycopersici (approximately 1 × 10⁴ spores/mL). The dishes were then placed in a growth chamber for 10 days, and, then, total DNA was extracted from the leaf discs using the NucleoMag Plant kit (MACHEREY-NAGEL). Fungal biomass in tomato leaves was quantified on the basis of the amplification of the P. neolycopersici alpha-tubulin gene using primer set (PnTub_F, TAATTCCTCGGGACTGCAAC; PnTub_R, CATCATCGGGTGAAGAAGGT) (Pathak et al., 2020) in 100 ng of total genomic DNA. Tomato ribulose-1,5-bisphosphate carboxylase/oxygenase (rubisco) gene was amplified using a primer set (Sl-rubisco_F, GAACAGTTTCTCACTGTTGAC; Sl-rubisco_R, CGTGAGAACCATAAGTCACC) (Mesarich et al., 2014) as a calibration standard. Quantitative real-time PCR analysis (qRT-PCR) was performed using the LightCycler 480 (Roche, Basel, Switzerland) with the KAPA SYBR Fast qPCR Kit (Nippon Genetics, Tokyo, Japan) according to the manufacturer’s instructions. Results were analyzed using the E^{− ΔCt} method (Livak and Schmittgen, 2001) with an average of 11 biological replicates. Data were analyzed for significant differences using the Mann–Whitney U-test and R version 4.0.3.

Because only the 500- and 1,000-fold dilutions of Botanigard were suppressive against cucumber powdery mildew, for further tests, we used the 1,000-fold dilution (ca. 1 × 10⁷ GHA spores/mL), the concentration generally used to control insect pests ( Supplemental Figure 1 ). When we tested the efficacy of the bioinsecticide against cucumber powdery mildew in the growth chamber, Botanigard applied to the cucumber leaves before inoculation with P. xanthii completely suppressed symptoms in contrast to the controls, which had typical symptoms with slight yellowing by 10 days after inoculation ( Figure 1A ; Supplemental Figure 2A ). In our comparison of disease severities on the cucumber leaves treated with Botanigard at different times before or after inoculation with P. xanthii, symptom development was reduced by 75%–90% only when plants were pretreated 1 or 3 days before inoculation with P. xanthii, indicating the prophylactic effect ( Figure 1B ). Pretreatment with Botanigard also had reduced spore germination of P. xanthii by about half compared with the untreated control ( Figure 1C ). The hyphal network of P. xanthii covered the leaf surface by 3 days after inoculation, but the hyphal growth was reduced about 75% by application of Botanigard. The thick hyphae of P. xanthii rarely overlapped with the thinner hyphae of B. bassiana ( Supplemental Figure 2B ).

In the greenhouse and field trials of the efficacy of Botanigard against powdery mildews of five vegetables (cucumber, melon, tomato, eggplant, and strawberry) ( Supplemental Table 1 ), disease severity after natural infection or inoculation with the respective powdery mildews was suppressed between 50% and 100% by pretreatment of Botanigard ( Figure 1D ; Supplemental Table 2 ). As expected, this bioinsecticide reduced the number of whitefly larvae and adults on tomato plants in the greenhouse ( Figure 1D ), indicating that it can control the insect pests and powdery mildews simultaneously.

We next investigated whether entomopathogenic fungus B. bassiana strain GHA, the active ingredient of Botanigard, is involved in the suppression of cucumber powdery mildew P. xanthii. First, to verify whether the biocontrol effect of strain GHA is systemic or not, one-half of the cucumber leaves grown in pots were treated with GHA and the another with water, and, then, the whole leaves were inoculated with P. xanthii spores. Typical symptoms appeared only on the mock-treated leaves, indicating that the biocontrol effect was limited to the area of GHA application ( Figure 2A ).

In the test to determine whether pretreatments of leaves with different concentrations of B. bassiana GHA induce resistance in cucumber, GHA had a concentration-dependent suppressive effect on powdery mildew ( Figure 2B ). In particular, pretreatment of a 1,000-fold dilution of Botanigard and high concentration of GHA (1 × 10⁹ spores/mL) showed a similar biocontrol effect against P. xanthii. When the leaf surfaces were observed with fluorescence microscopy at 24 h after inoculation, fluorescent epidermal cells were observed under germ tubes of P. xanthii, indicating a HR-like cell death ( Figure 2C ; Supplemental Figure 3 ). The pretreatments with either the bioinsecticide or GHA strain increased the number of HR-like cell death with increasing concentration of the active ingredient ( Figure 2C ). Spore germination of P. xanthii was inhibited about 60% by Botanigard, but not by GHA ( Figure 2C ). B. bassiana hyphae were rarely present around epidermal cells with HR-like cell death.

Cucumber roots were treated with spore suspension of GHA or distilled water and grown in pots for 2 weeks. The mass of aerial plant parts and roots did not differ significantly from the control, indicating that GHA is not a PGPF per se under this condition ( Supplemental Table 3 ).

For the comparison of the transcriptomes from the cucumber leaves to investigate the influence of B. bassiana colonization on the plant, 26.7 million reads were mapped to the cucumber genome, and 150 DEGs were identified; 137 DEGs were upregulated and 13 were downregulated ( Figure 3A ; Supplemental Table 4 ; Figure 4 ). We did not have enough transcripts for P. xanthii and GHA to map to the fungal genomes.

In the GO annotations with a cutoff of p < 0.05, the 137 upregulated DEGs were classified into 39 functional subgroups, including 10 in molecular function and 29 in cellular component ( Figure 3B ). For molecular function, the GO terms heme binding (GO:0020037), tetrapyrrole binding (GO:0046906), and iron ion binding (GO:0005506), which all may be related to iron uptake, were detected. For cellular component, the DEGs were significantly enriched for the primary metabolic pathway and biosynthetic processes for carboxylic acid, oxoacid, organic acids, lipids, and fatty acids ( Figure 3B ; Supplemental Table 5 ). DEGs that were upregulated during the response of cucumber plants to B. bassiana GHA and powdery mildew were enriched for response to biotic stimulus (GO:0009607), response to external biotic stimulus (GO:0043207), response to other organism (GO:0051707), and biological process involved in interspecies interaction between organisms (GO:0044419) ( Figure 3B ). Two GO terms (GO:0071554, cell wall organization or biogenesis; and GO:0071669, plant-type cell wall organization or biogenesis) indicated reconstitution of plant cell walls in response to powdery mildew ( Figure 3B ). In addition, DEGs were significantly enriched in two GOs related to phytohormones including SA, which is involved in induction of plant resistance (GO:0009725 and GO:0009751).

To determine whether the phytohormone is involved in B. bassiana GHA-mediated resistance to cucumber powdery mildew, a quantitative analysis of endogenous phytohormones was performed. SA accumulated significantly in leaves treated with Botanigard or with a high concentration of GHA (1 × 10⁸ and 1 × 10⁹ spores/mL), but the levels of JA and ET did not change significantly ( Figure 4 ).

To further elucidate the influence of SA in B. bassiana GHA-mediated resistance, we established a tomato system with MM-NahG, which is unable to accumulate SA, and powdery mildew P. neolycopersici. As NahG plant responds excessively to inoculation with tomato powdery mildew, leaf discs were prepared and sprayed with the spore suspension of P. neolycopersici. Powdery mildew symptoms appeared on all leaves including the mock control and did not visibly differ in severity ( Supplemental Figure 5 ), so fungal biomass in the tomato leaves was quantified by qRT-PCR. The GHA treatment of the MM tomato did inhibit powdery mildew growth compared with the untreated control but was not as effective in suppressing growth in the MM-NahG ( Figure 5 ). A significant difference in fungal biomass was also found for MM and MM-NahG treated with GHA. These results suggest that SA is partially involved in the resistance induced after treatment with B. bassiana GHA.