Selection of Beauveria bassiana (Hypocreales: Cordycipitaceae) strains to control Xyleborus affinis (Curculionidae: Scolytinae) females

Beauveria bassiana source

Nineteen coleoptera-isolated B. bassiana strains were selected from the catalogue from the EPF collection (http://www.gob.mx/senasica/documentos/coleccion-de-hongos-entomopatogenos, accessed 28 January 2019) from the Centro Nacional de Referencia de Control Biológico (CNRCB), located in Colima state, México. Fifteen of the 19 strains were isolated from beetles, from which 13 were the coffee berry borer Hypothenemus hampei Wood & Bright (Curculionidae: Scolytinae).

All strains were stored on silica gel at 5 °C to be further grown on Petri dishes containing Sabouraud Dextrose Agar (SDA), supplemented with 1% yeast extract (SDAY), which is commonly used for entomopathogenic fungi culture and assays (Talaei-Hassanloui et al., 2007), and incubated at 27 ± 2 °C in darkness for 20 d.

Xyleborus affinis colony

To establish X. affinis colony under laboratory conditions, adult beetles were collected in Comala, Colima, México (19.45965° N, −103.65603° W), from avocado trees showing such signs of ambrosia beetle infestation as decay and wilt. Permit to collect beetles was granted by the Colima State Plant Health Committee (Comité Estatal de Sanidad Vegetal de Colima) and the landowner, Julio Aguirre González. Beetles were identified by using Wood’s (1982) taxonomic keys (Castrejón-Antonio et al., 2017a). Adults were kept in plastic conical 50 mL tubes and fed on artificial avocado sawdust diet (Cooperband et al., 2016), placing five insects per tube and incubating at 25 ± 2 °C for 4 weeks.

Selection stage 1: B. bassiana strains growth profile

For B. bassiana strains selection, before performing bioassays against ambrosia beetles, and since germination and infection rates, high conidium size, and sporulation are factors that may correlate with their virulence, mycelium growth rate (mm day⁻¹), conidial production (conidia mm⁻²), conidia germination (%), and germ tube length (mm) were analysed on SDAY culture medium. Each experiment was done in triplicate.

Growth rate

A 6-mm diameter sterile filter paper disc was placed at the centre of a 90-mm Petri dish containing SDAY medium. Based on previous tests, 5 μL of an 8 $\times$ 10⁶ conidia mL⁻¹ suspension was added on each disc. All strains were cultured at the same time and incubated at 27 ± 2 °C. Colony diameter size was daily measured, calculating the average (mm d⁻¹) of the two perpendicular diameters. Diameter size data of fungal growth were completed when one of the 19 tested strains grown within 0.5 cm of the Petri dish edge. Data were analysed using linear regression model (SPSS, V21), where the slope value was considered as the growth rate.

Conidia production

Beauveria bassiana strains were cultured as described above and incubated at 27 ± 2 °C for 20 d. A 20-mm-diameter sample mycelial plug was taken from the growing plate, halfway between the colony centre and edge (Montesinos-Matías et al., 2011b). Conidia were released from the resulting culture plug by transferring it into a 50 mL conical tube containing 10 mL Tween 80 (0.05%) and vortexed for 1 min (maximum speed) (Vortex-Genie 2, Scientific Industries, Bohemia, NY, USA). Conidia were counted in a haemocytometer.

Conidial germination

Ten microliters of an 8 $\times$ 10⁶ conidia mL⁻¹ suspension were transferred into a conical microtube containing 990 μL of Tween 80 (0.05%), after which it was vortexed at maximum speed (7) for 1 min, and 10 μL of the resulting suspension were cultured in a 17-mm diameter disk of SDAY medium at 27 ± 2 °C for 18 h. After incubation, the disk was transferred onto a coverslip and a drop of blue lactophenol (20 g of phenol, 20 g of lactic acid and 40 g of glycerol mixed in 20 mL distilled water) was added along with a coverslip. Viable propagules percentage was determined by counting 100 conidia (germinated or not), using an optical microscope with a 40$\times$ objective (Safavi et al., 2007). Conidia were counted as viable if germ tube length was twice the conidia diameter size (Wraight, Inglis & Goettel, 2007).

Germ tube length

Conidial germination tubes were detected from the slides used in the germination tests (18 h). Then, germ tube length from 30 conidia was measured for each strain. Photographs were taken using an Axio Scope A1 microscope with a 40$\times$ objective and an Axiocam ICc 1 camera (Carl Zeiss de México, S.A. de C.V., Ciudad de México, México) and analysed by the Axiovision Release 4.8.2 software.

Stage 1 data analysis

Mean values of each variable for the 19 strains were standardised to a Z-score. Prior to standardisation, data expressed as percentages were transformed using the arcsine function. Data were analysed using one-way analysis of variance (ANOVA) and Tukey’s test. As well as Principal Component Analysis (PCA), using the Spearman method setting of the XLSTAT statistical package (V.2018 for Windows, Addinsoft, NY, USA). PCA correlations were considered significant when the Bartlett’s sphericity test p value was ≤0.05. Variables with a correlation coefficient ≥0.6 were considered relevant.

Following growth variables correlation from PCA analysis, 10 strains were selected for further analysis, because they grouped in at least two of the clusters. Strains that met this criterion were CHE-CNRCB 21, 26, 37, 38, 171, 174 and 485. Strain CHE-CNRCB 431 was selected for possessing the highest germination rate, whereas strains CHE-CNRCB 44 and CHE-CNRCB 117 were chosen based on conidia production.

Selection stage 2: evaluation of fungal virulence factors

Two cell wall-bound cuticle-degrading enzymes–Pr1-like proteases and β-N-acetyl glucosaminidases (NAGase) chitinase (St. Leger et al., 1991; Safavi et al., 2007) were evaluated for the 10 selected B. bassiana strains. In addition, conidial hydrophobicity and monopolar conidial germination percentages were assessed (Talaei-Hassanloui et al., 2007). To stimulate enzyme production, strains were grown on plates in a culture medium containing a 10:1 carbon: nitrogen ratio at 27 ± 2 °C (Safavi et al., 2007). Each experiment was performed in triplicate.

Pr1 protease activity

To determine Pr1 protease activity, 500 μL of Tris-HCl 0.1 M buffer (pH 7.9) (Sigma–Aldrich, St. Louis, MO, USA), containing one mM of N-succinyl-alanine-alanine-2-proline-phenylalanine-p-nitroanilide substrate (Sigma–Aldrich, St. Louis, MO, USA), was used following protocols reported by Ayala-Zermeño et al. (2015) and Castrejón-Antonio et al. (2017a). For this, 500 μL of a 3 $\times$ 10⁸ conidia mL⁻¹ suspension were mixed in a 1,500 µL conical microtube containing Tween 80 (0.05%). Mixture was incubated at 27 ± 2 °C for 10 min and the enzymatic reaction was stopped with 100 μL of 0.1 M HCl. Tubes were then placed on ice for 5 min to stabilise the resulting chromogen and then centrifuged (Centrifuge 5430 R, Eppendorf, Hamburg, Germany) at 15,295×g at 4 °C for 5 min. Supernatant optical densities were measured using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) at 405 nm. A mixture of Tris-HCl buffer (0.1M), substrate and Tween 80 (0.05%) solution was used as a blank. Protease activity was reported in nmol of nitroaniline mL⁻¹ min⁻¹ (Castrejón-Antonio et al., 2017a).

NAGase chitinase activity

To determine NAGase chitinase activity, 200 μL of citrate-K₂HPO₄ 0.2 M buffer (pH 5.6) and 200 μL of 1 mg mL⁻¹ of p-nitrophenyl N-acetyl-β-D-glucosaminide (Sigma–Aldrich, St. Louis, MO, USA), were used according to previous reports by St. Leger et al. (1991) and Castrejón-Antonio et al. (2017a). For this, 200 μL of a 3 $\times$ 10⁸ conidia mL⁻¹ suspension were mixed in Tween 80 (0.05%) in a 1.5 mL conical microtube. Mixture was incubated in a shaker (Lab-Companion, BS-11, Korea) at 120 rpm and 37 ± 2 °C for 60 min, and enzymatic reaction was stopped by adding 1 mL of 0.02 M NaOH. Microtubes were then centrifuged for 5 min at 15,295 × g at 4 °C (Thermo Scientific, Hamburg, Germany) and optical densities were determined at 400 nm. A mixture containing citrate-K₂HPO₄ buffer, substrate, and Tween 80 (0.05%), was used as a blank. NAGase chitinase activity was reported in nmoles of p-nitrophenol mL⁻¹ min⁻¹.

Hydrophobicity

Conidial hydrophobicity was performed according to Kim et al. (2010). For this, strains were grown on SDAY slant culture tubes (18 $\times$ 150 mm) at 27 ± 2 °C, and after 15 d, culture surface was scraped using 0.1 M KNO₃ (pre-filtered through a 0.2 μm filtration membrane (Thermo Scientific, Hamburg, Germany)) to release the conidia. The resulting 15 mL suspension was transferred into a 50 mL conical tube and vortexed for 5 min at highest speed (7, Vortex-Genie 2, Scientific Industries, Bohemia, NY, USA). Suspension was then filtered into a new 50 mL corning tube using sterile double-layer gauze and labelled as a stock solution. The conidial stock solution concentration was determined using a haemocytometer (Marienfeld, Lauda-Königshofen, Germany) and 5 mL of a suspension of 1 $\times$ 10⁷ conidia mL⁻¹ was prepared. Then, one mL of n-hexadecane (Sigma–Aldrich, St. Louis, MO, USA) and three mL of the conidial suspension were added to 10 $\times$ 100 mm glass test tubes. This mixture was vortexed at high speed (6) for 20 s and allowed to stand for 30 min to allow separation of the aqueous and organic phases. Then, one mL of the aqueous phase was carefully extracted with a long neck glass Pasteur pipette and poured into a new glass tube. Conidia were counted using a haemocytometer. Relative hydrophobicity was calculated using the following formula: $$H\left(\mathrm{\%}\right)=\left(1-{\displaystyle \frac{C}{C_0}}\right)100\mathrm{\%}$$where H represents hydrophobicity, C₀ represents the conidial stock suspension, and C represents the conidial concentration in the aqueous phase.

Percentage of monopolar conidial germination

A 1:1,000 dilution was prepared from each 3 $\times$ 10⁸ conidia mL⁻¹ suspension used for the enzyme activity evaluation. From these dilutions, 10 μL were used to inoculate a 2-cm diameter disk of culture medium. After inoculation, each disk was incubated at 27 ± 2 °C for 18 h, then the disk was placed onto a clean slide. This sample was stained by adding 50 µL of blue lactophenol and covered with a coverslip. Sample was observed under a microscope at 40×. The conidial percentage showing only a single germ tube was determined by observing 100 conidia (Talaei-Hassanloui et al., 2007).

Stage 2 data analysis

Mean values of each parameter were clustered using PCA analysis, as described in “Stage 1 data analysis”. Variables showing the highest correlation such as enzymatic activity and hydrophobicity were clustered. Upon clusters, four strains were selected to test their virulence against X. affinis females.

Biocontrol of X. affinis by B. bassiana

To determine B. bassiana pathogenicity, we selected X. affinis females from reared colonies of approximately 10–15 d old, assuring females were not too teneral nor too sclerotised (with a similar sclerotisation degree). Females were externally hygienised in order to remove as much as possible those contaminants attached to the insect, either by remaining diet and by microbiota present in the galleries. A cleaning process reported for ambrosials (Castrillo, Griggs & Vandenberg, 2013; Cruz et al., 2018) was used as follows: (1) Tween 80 (0.05%), (2) 70% ethanol, (3) again Tween 80 (0.05%) and (4) 0.1% sodium hypochlorite, with each cleaning lasting 10 s. Insects were allowed to dry for 20 min on sterilised absorbent paper. Once dried, they were separated in 10 individual groups in 150 $\times$ 600 mm sterile glass Petri dishes.

Xyleborus affinis females were treated with each B. bassiana strain using a Potter Tower (BS00281, Burkard Scientific, Uxbridge, UK) calibrated at 103.4 kPa pressure, which applied a dose of 2 mL of 1 $\times$ 10⁸ viable conidia mL⁻¹. Strains conidial viability was determined as previously described. Experimental set-up was a completely randomised design with two replicates (n = 10 for each replicate), repeated five times.

After females were sprayed with the conidial suspension, beetles were individually placed into plastic vials (50 mm high and 25.7 mm in diameter), containing a piece of avocado sawdust artificial diet, same as used for rearing purposes. Vials were incubated in a humidified chamber at 27 ± 2 °C and 70 ± 5% humidity for 8 d, replacing the feeding source every 2 d. Vials were then daily observed to identify and collect dead individuals, and mortality rates were registered. Dead beetles were disinfected using the described external hygienization, by vortexing at low speed (2) in each solution for 10 s. Clean insects were then placed in a humidified chamber and incubated to allow B. bassiana sporulation for up to 10 d.

Since mortality results were lower than 60%, the time required to reach 40% mortality (LT₄₀) was estimated from a plot of accumulated mortality versus time, for all strains. Data were fitted to an exponential decaying function (Rodríguez-Gómez et al., 2009) using the following formula: $$Y=\left(100-S\right)e^{-k\left(t-t_0\right)}+S;\mathrm{I}\mathrm{f}t>t_0$$where Y = 100 if 0 ≤ t ≤ t₀, where Y is percent survival at time t, k is specific death rate (d⁻¹), t_o is delay time (d) and S is estimated asymptotic survival level (%). This model corresponds to a first order differential equation with the indicated time delay, the aforementioned initial condition and the asymptotic value, where $Y\to S$ as $t\to \mathrm{\infty }$.

Infected beetles’ conidial production

Five beetles covered with aerial conidia (not mycelium) removed from each bioassay replicate, were used. In order to have sample representativeness, insects were randomly selected and transferred into two mL Eppendorf tubes with 1.5 mL of Tween 80 (0.05%), and vortexed (maximum speed) for one min. Tubes were maintained at 4 °C for 24 h. After vortexing the tubes for 1 min, six aliquots of 10 μL were removed from the resulting suspension and placed in a haemocytometer to determine the average of conidial number per beetle, per tested strain (n = 4).

Bioassays analysis

To correct mortality collected data with the control treatment, Abbott’s transformation was used (Abbott, 1925), which resulted in <5% adjustment. Data distribution was confirmed using Kolmogorov Smirnov test and were analysed by ANOVA and Duncan’s multiple range tests. Values that were expressed as percentages were converted using an arcsine transformation prior to ANOVA (Studebaker, 1985). For analysis statistical software SPSS was used (V.20, 2012, Chicago, IL, USA).

Stage 1 analysis: growth profile

Growth profiles of B. bassiana isolates evaluated are shown in supplemental Table S1. PCA indicated grouping of data into three main components that accounted for 94.25% of total variance (Eigenvalue 0.94). The best correlation with component 1 (PC1) was observed for germ tube length and germination (correlation coefficients of 0.89 and 0.91 respectively), which represented most of the total variance (40.39%) (Fig. 1; Table S2). Component 2 (PC2) showed the highest correlation with growth rate (correlation coefficient of 0.98), representing 27.5% of total variance (Fig. 1). Component 3 (PC3) correlated with conidial production, accounting for 26.37% of the total variance (Fig. 2). Based on the weighted sum of each strain for component, those that presented values greater than 0.3 were selected (locked in the ovals). The strains were fitted into three clusters: (1) those characterised by longer germ tubes and higher germination levels (CHE-CNRCB 21, 22, 26, 37, 38, 485 and 486); (2) those with higher growth rates (CHE-CNRCB 25, 37, 105, 171, 174, 431 and 485); and (3) those with higher conidia production (CHE-CNRCB 21, 26, 38, 44, 117, 171 and 174). Strains were grouped based on these results (Table S3).

Stage 2 analysis: virulence factors

The average values of the measured parameters Pr1-like proteases, NAGase chitinase, conidia hydrophobicity and monopolar conidial germination percentage for each of the 10 selected strains are shown in Table S4. PCA analysis revealed that 84.26% of the variability could be attributed to two main components (Fig. 3; Eigenvalue 0.96). A high correlation was seen between PC1 and the proportion of monopolar conidia as well as hydrophobicity (correlation coefficient of 0.90). For PC2, a correlation was seen with Pr1 protease activity and NAGase chitinase activity (correlation coefficient of 0.96) (Table S5).

Comparison of the distribution of the weighted sum of each strain for component showed that CHE-CNRCB 44 and 171 had higher hydrophobicity and monopolar conidial germination levels, whereas CHE-CNRCB 431 and 485 exhibited higher NAGase chitinase and Pr1 enzymatic activity, respectively (locked in the ovals of Fig. 3). The described strains distribution by the evaluated variables in this stage allowed their selection for the subsequent in vivo tests.

Bioassays against X. affinis

Xyleborus affinis females were susceptible to B. bassiana infection when applied at 1 $\mathrm{x}$ 10⁸ conidia mL⁻¹, using spray application as the inoculation method. The average mortality rate (n = 100) among the evaluated B. bassiana strains ranged between 40% and 58% (Table 1). Strains 485, 171 and 431 caused mortality rates greater than 40%, with 58 ± 13.7%, 48 ± 9.0% and 48 ± 9.7% mortality rates, respectively, whereas strain 44 induced the lowest mortality (40 ± 13.3%). Mortality in the control was 19.6 ± 2.7%. In order to confirm that beetle mortality was by fungal infection, B. bassiana sporulation was analysed from dead washed insects incubated inside a humidified chamber. Growth of B. bassiana on dead beetles resulted in aerial mycelia development, ranging between 12% (CHE-CNRCB 44) to 30% (CHE-CNRCB 485).

Lethal Time 40 (LT₄₀) averages between 5.32 and 9.62 d were observed by strains CHE-CNRCB 485 and CHE-CNRCB 171, respectively. Strain 44 caused the fastest rate of mortality, where treated insects started dying from the second day (t₀ = 1.95). Beetles treated with CHE-CNRCB 431 begin dying around the fifth day, following treatment. The amount of conidia counted per treated beetle demonstrated that strain 431 caused the highest conidia production per beetle corpse (4.74 ± 0.8 $\times$ 10⁷ conidia insect⁻¹).