High susceptibility of Tetranychus merganser (Acari: Tetranychidae), an emergent pest of the tropical crop Carica papaya, towards Metarhizium anisopliae s.l. and Beauveria bassiana strains

Sampling and identification of mites in papaya orchards of Campeche, México

During 2017 and 2018, five papaya orchards located in the state of Campeche, México were monitored for the presence of mites (Fig. 1). We sampled one hectare in each orchard that was previously identified by the producers as a site with a possible mite presence. Fifteen plants were randomly selected within that area and two mature leaves were sampled from the lower stratum of each plant. The leaves were transported to the laboratory in a plastic bag inside a container. All mites present on the abaxial surface of the sampled leaves were observed with a stereoscope and collected to be identified. The specimens were incubated in lactic acid as a clearing agent, for seven days. Subsequently, the males were mounted on microscope slides with Hoyer’s mounting medium (Henderson, 2001). The taxonomic keys of Gerson, Smiley & Ochoa (2003) were used to identify the families of the collected mites. The Tuttle, Baker & Abbatiello (1976) keys were used for identification at the genus and species level (Fig. 2A–D).

Establishing a T. merganser colony

T. merganser were collected in papaya fields and were raised on thirty 15-day-old plants of Phaseolus vulgaris (Fabaceae) var. Jamapa. The plants were watered every third day. The mite colony was protected from external depredators by containing them inside cages built with a 200 µm mesh. The colony was raised at 27−28 °C, 67–77% relative humidity (RH), and a 12:12 h (light:dark) photoperiod. The density of the mite colony was maintained at approximately 200 mites per leaf. Plants severely damaged by mite infestation were replaced with new plants.

Entomopathogenic fungal strains

Thirty strains belonging to the entomopathogenic fungi collection of the Microbial Pest Control Laboratory of the Postgraduates College (Colegio de Postgraduados, Campus Campeche) were tested against T. merganser. A total of 11 strains were identified as B. bassiana (Bb005, and Bb014 to Bb023) and ten as M. anisopliae sensu lato (Ma002 to Ma009, Ma014, and Ma034) using conidiophores morphological keys. The M. anisopliae s.l. and B. bassiana strains were isolated from the soil or infected insects that were collected from several agricultural fields or semi-conserved rain forest areas from Campeche. Nine strains of Lecanicillium sp. isolated from Onychiurus folsomi (Collembola: Onychiuridae) collected in México City (19°18′58.59″N; 99°11′30.12″W) were also tested, as well as the CHE-CNRCB 377 strain of Hirsutella thompsonii var. sinematosa from the collection of the National Reference Center for Biological Control (Centro Nacional de Referencia de Control Biológico) (SENASICA, 2016) isolated from Phyllocoptruta oleivora (Acari: Eriophyidae).

The internal transcribed spacer (ITS) region of seven strains identified in this work as the more lethal for T. merganser (Ma002, Ma003, Ma004, Ma005, Ma014, Ma034, Bb016) were amplified by PCR using 100 ng of gDNA (Raeder & Broda, 1985), 40 pmol of ITS1 and ITS4 primers (White et al., 1990), and 30 µL of high fidelity supermix (cat 10790-020; Invitrogen, Waltham, MA). The PCR cycles consisted of 1 min at 95  °C, 55 °C and 72 °C with 2 min at 95 °C for initial denaturation and 7 min at 72 °C for the final extension. The PCR products were purified with the QIAquick PCR Purification Kit (cat. 28104; Qiagen, Hilden, Germany) and sequenced by an external service provider. The ITS sequences were compared by BLASTn analysis with NCBI (nr/nt) and UNITE 8.3 (Nilsson et al., 2019) databases. The M. anisopliae s.l. ITS sequences were aligned with ITS sequences of reference strains from UNITE/NCBI databases with Clustal Omega (1.2.4) and a guide tree was constructed. A phylogenetic tree was constructed with MEGA X using the UPGMA and Neighbor-Joining method with 1,000 replicates.

The strains of Beauveria, Metarhizium, and Lecanicillium were grown in potato dextrose agar and incubated at 28 °C until conidiogenesis occurred. H. thompsonii was grown in Sabouraud dextrose agar with yeast extract and transferred to medium H (Cabrera, Vega & Ayra, 2006) for sporulation. The conidia were recovered from sporulated cultures with 0.01% triton in sterile distilled water by scraping the surface. The suspension was homogenized and the conidia were counted in a Neubauer chamber to adjust the concentration. The percentage of viable conidia was also determined. Isolates with more than 90% germination were used in bioassays.

Screening of fungal pathogenic strains on adult T. merganser females

To evaluate whether any B. bassiana, M. anisopliae s.l., or Lecanicillium sp. strains from our collection or the H. thompsonii strain from SENASICA were pathogenic to T. merganser, we performed a mass screening bioassay evaluating mortality on adult females (Fig. 2A). For this infectivity bioassay, papaya leaf disks of three cm in diameter were disinfected under aseptic conditions in a solution of sterile distilled water with sodium hypochlorite (0.06%) and 10 µL of liquid soap, and continually stirred for 1 min. Leaf disks were rinsed three times with sterile distilled water and subsequently allowed to dry on a sterile paper towel. Each leaf disk was immersed in 2 mL of 1 × 10⁸ conidia/mL of each strain for 30 s, before allowing it to dry on a sterile paper towel. This conidia concentration was previously selected by testing three strains at 1 ×10⁴, 10⁶, and 10⁸ conidia/mL. We observed a mortality of more than 50% only at the highest concentration. All strains of the Lecanicillium sp. showed very low sporulation levels, thus we used 1 ×10⁷ conidia/mL for bioassays. The viability of the conidia of all tested strains was evaluated before each experiment. The inoculated disks were placed upside down on an Oasis^® floral sponge base of 3.6 × 3.3 × 0.5 cm inside a five cm diameter Petri dish containing 3 mL of water. Twenty adult T. merganser females were placed onto each disk using a fine brush. In the control group the disks were immersed in sterile distilled water with 0.01% triton. The Petri dishes were kept closed and maintained at 28 °C with 50–55% RH and a 12:12 h (light:dark) photoperiod. On the third day, the mites were transferred onto a new leaf disk without inoculum. The mortality of the mites was quantified daily from the third to the sixth day post-infection. The mortality caused by mycosis was verified by the microscopic observation of specific fungal sporulation on the mummified mites (Fig. 2E). Three independent assays comparing all strains were performed. The average percentage and the standard deviation of adult mortality caused by each strain were calculated and then plotted with Sigma Plot 14.0. Afterwards, the percentage of mortality was arcsine-transformed to normalize the data before one-way analysis of variance (ANOVA) was performed. Differences between the means were compared by Tukey tests, and p = 0.05 was considered as significant. To estimate the lethal time to 50% mortality (LT₅₀) with the screening assay data, the accumulated mortality recorded at the concentration of 1 ×10⁸ conidia/mL for each strain were corrected for natural mortality (Abbott, 1925) and used to estimate LT₅₀ using the Probit function (Finney, 1971) with a significance level of 0.05%. The statistical analyses of all bioassays in methods section were carried out using Minitab 14 Statistical Software (2003).

LC₅₀ & LT₅₀ of the most virulent strains

The seven more virulent strains (causing 90–100% mortality of T. merganser in less time) were selected to perform a bioassay similar to the one described above. Each strain was evaluated in serial concentrations from 1 ×10⁴ to 1 × 10⁸ conidia/mL, from the third to the seventh day. In this range, the recorded mortality was between 0 and 100%. For each strain, at least six independent experiments were performed and the lethal concentration to 50% mortality (LC₅₀) was estimated and the LT₅₀ was recalculated using the Probit analysis with a significance level of 0.05%.

Infectivity on eggs

The strains that showed the ability to infect eggs laid by the females during the screening assay (Fig. 2F) were evaluated with a novel infectivity bioassay. Sixty adult females were placed on a disinfected papaya leaf fragment (8 × 5 cm) for 24 h to lay eggs. The females were removed by washing the leaves with 0.05% triton with strong agitation for 1 min. The leaves were allowed to dry in a laminar flow cabinet. Twenty eggs were carefully recovered with a fine brush and placed onto 1 cm² of sterilized screen-printing mesh (130 µm). The mesh was previously placed onto 1 cm² of a sterilized cotton pad and both were placed onto a sterilized glass slide. The eggs were separated with the brush, while avoiding contact between the eggs (Fig. 2G). A volume of 1.5 µL of a 1 × 10⁸ conidia/mL suspension was added onto the glass slide, which was sufficient to cover all eggs, allowing the excess liquid to be absorbed by the cotton pad. Sterile distilled water with 0.01% triton was added to eggs in the control group. Each slide was placed onto a cotton pad moistened with sterile distilled water to promote fungal growth, inside a 15 mm Petri dish. The Petri dishes were kept at 28 °C, 50–55% RH, and a 12:12 h (light:dark) photoperiod. Infectivity was recorded on the fifth day based on the presence of mycelia on the eggs (Fig. 2H) and by verifying the characteristics of the spores of each fungus. The fungi were stained with lactophenol cotton blue at 0.5%. Three independent experiments were performed. The average percentage and the standard deviation of infectivity on eggs were calculated, and plotted (Sigma Plot 14.0). The percentage of infectivity was arcsine-transformed and a one-way analysis of variance (ANOVA) was performed. Differences between the means were compared by Tukey tests; p = 0.05 was considered as significant.

Identification of T. merganser and T. urticae in papaya orchards of Campeche, México

Five papaya orchards located in two distant sites in the Campeche State were sampled repeatedly over two years, reaching a total of twelve samples (Fig. 1). The presence of two Tetranychus species was recorded: T. merganser and T. urticae. On some sampling dates, we did not find any mites, possibly because the producers use acaricides to control the pest. T merganser was found in high density in three of the five sampled orchards growing on two different papaya varieties: Maradol and a Tainung hybrid. T. urticae was observed in less density than T. merganser, when it was present. The characteristics confirming the identification of species from the Tetranychus genus are the distance between the pair of duplex setae, the presence or absence of tenant hair, and the morphology of the aedeagus knob. The aedeagus knob of T. urticae has an anterior projection acute and a rounded shape in posterior projection, their females are greenish (Figs. 2B, 2D). The aedeagus of T. merganser resembles the head and neck of a goose (Fig. 2C); the pretarsus do not have tenant hair, and each empodium ends in a tuft of three pairs of hair. The specimens that were identified as T. merganser had the unique coloration of its species, namely bright red females and pale yellowish males (Fig. 2A).

Given that T. merganser has been previously reported in C. papaya orchards of Yucatán and Veracruz (Fig. 1), other states located along the Gulf of México (Valencia-Domínguez et al., 2011; Abato-Zárate et al., 2014; Monjarás-Barrera et al., 2017), we looked for biological control for this emerging pest to avoid the use of chemical acaricides and resistance selection.

Screening of fungal pathogenic strains on adult T. merganser females

The H. thompsonii strain that was tested showed no infectivity in T. merganser adult females at 1 × 10⁸ conidia/mL, even on the sixth day post-inoculation. However, many strains in our collection were infectious towards adult T. merganser females, indicating a considerably different mortality effect between species (Fig. 3). The M. anisopliae s.l. strains caused the highest T. merganser mortality levels in comparison to the other fungal species (Fig. 3A). All these strains resulted in significant differences in mortality, compared to the control (F = 74.85, p < 0.0001) (Data S1). Five M. anisopliae s.l. strains (Ma002, Ma004, Ma005, Ma014 and Ma034) caused mortality to 100% of mites, Ma003 to 91%, and the other strains (Ma006, Ma007, Ma008 and Ma009) caused mortality from 75% to 85%.

In the case of B. bassiana (Fig. 3B), all strains were infectious to adult females, but only seven of the eleven strains showed mortality with statistically significant differences compared to the control (F = 17.55, p < 0.0001) (Data S1). Three strains (Bb005, Bb014, Bb016) showed the highest mortality levels, reaching 95% in the case of Bb016. The Bb015, Bb019 and Bb021 strains showed a mortality of only 40–43%.

Very low sporulation levels were observed in the strains of the Lecanicillium sp., thus we were not able to obtain a highly concentrated suspension, but performed the test at 1 × 10⁷ conidia/mL. At this concentration, no strain of Lecanicillium sp. showed mortality with a statistically significant difference from the control (F = 2.9, p < 0.023) (Fig. 2C and Data S1), although all strains showed some degree of infectivity (mycosis on the mummified mite).

To compare the fungal strains’ speed to kill T. merganser adult females, we calculated the time it took to reach 50% of mortality (LT₅₀) at 1 × 10⁸ conidia/mL with strains that showed significant differences in mortality, compared to the control (Table 1). In this preliminary assay, the five M. anisopliae s.l. strains that caused 100% mortality to T. merganser also showed the fastest mortality times (Ma002, Ma005, Ma034, Ma004, and Ma014). Of the B. bassiana strains, Bb016 exhibited the shortest LT₅₀, followed by Bb005 and Bb014. These are also the three B. bassiana strains that caused the highest mortality to T. merganser. The chi-square, fiducial limits (FL) and slope (m) from Probit analysis are shown in Table 1.

LC₅₀ & LT₅₀ of the most virulent strains

We selected the six strains of M. anisopliae s.l. (Ma002, Ma003, Ma004, Ma005, Ma014 and Ma034) and the one of B. bassiana (Bb016) that caused 90%–100% mortality to T. merganser and showed the lower LT ₅₀ in the mass screening assay to evaluate their virulence with serial doses for seven days and with more replicates (Data S1). The doses required to cause mortality to 50% of a population (LC₅₀) were determined and the LT₅₀ was recalculated. The most virulent strains were Ma034, Ma005, Ma003, Bb016 and Ma002, with LC₅₀ values around 10⁶ conidia/mL (1.7−3.2 ×10⁶, respectively) (Table 2); which also have the lowest LT₅₀. The Ma004 and Ma014 strains had LC₅₀ values above 10⁷ conidia/mL and the highest LT₅₀. The chi-square, fiducial limits (FL) and slope (m) from Probit analysis for LC₅₀ are shown in Table 2. Taken together, the LT₅₀, and LC₅₀ data allowed us to identify Ma034 and Ma005 as the most aggressive strain towards T. merganser adult females followed by Ma003, Bb016 and Ma002.

The ITS of these seven strains were sequenced and we confirmed the identification of Bb016 as B. bassiana (Table 3). According with BLASTn hits, ITS from three strains of M. anisopliae s.l. (Ma002, Ma003 and Ma014) showed the highest similarity with M. anisopliae and three with M. robertsii (Ma004, Ma005 and Ma034) (Table 3). The phylogenetic analysis confirmed that these six strains belong to the so-called PARB clade of M. anisopliae (Table 3, Fig. S1). Altough the ITS information is not enough to distinguish between the species of PARB clade (Bischoff, Rehner & Humber, 2009; Rehner & Kepler, 2017); the ITS alignment showed that Ma004, Ma005 and Ma034 have a different number of adenines in a short stretch of this nucleotide, comparing with Ma002, Ma003 and Ma014 (Fig. S2). That difference is also present in other M. robertsii ITS sequences from the Genbank or UNITE databases, including sequences from the reference strains ARSEF 2575 and ARSEF 23, compared to M. anisopliae sequences from ARSEF-7487, JEF-290, BRIP 53293, BRIP 53284.

Infectivity on eggs

We recorded that seven Metarhizium anisopliae s.l. strains (Ma002, Ma003, Ma004, Ma005, Ma008, Ma014 and Ma034) and two B. bassiana strains (Bb016 and Bb017) were infectious to T. merganser eggs laid as an additional observation during the screening assay (Figs. 2F and 3). Most of these strains also showed a high mortality to the adult females. Thus, we developed an in vitro egg infectivity assay (Fig. 2G).

All of the tested strains showed mycosis on the eggs, although four strains (Ma002, Ma003, Bb016 and Bb017) did not show statistically significant differences compared to the natural mortality in the control group (F = 14.83, p < 0.0001) (Fig. 4 and Data S3 ). However, the Ma014 and Ma034 strains showed a high infection level, reaching 71.7 and 81.7%, respectively (Data S3 ). The data of the Ma004, Ma005 and Ma008 strains were excluded from the analysis due to their high variability. This analysis points to Ma034 as the most promising strain since it is the most virulent to adult females with 100% mortality, has the lowest LC₅₀ and low LT₅₀ value, and is highly infectious to eggs.