Biocontrol of invasive pheretimoid earthworms using Beauveria bassiana

Specimen collection

Juvenile pheretimoids were collected at the University of Vermont’s Centennial Woods (CW) Natural Area (44° 28′ 36″ N, 73° 11′ 12″ W) in June of 2018 and May and July 2019. At these times pheretimoids are still in their juvenile stage and cannot be identified to species by morphology. In July 2019, the earthworms were fully extended but still had not developed a clitellum. At this site the earthworm community is exclusively pheretimoids (Keller, Görres & Schall, 2017), they were also identified by their trashing behavior when picked up which is unique to these taxa in Vermont.

The collection site was on alluvial, poorly drained Limerick silt loam (coarse-silty, mixed, super active, nonacid, mesic Fluvaquentic Endoaquepts) (Soil Survey Staff, 2019). The canopy comprised white pine (Pinus strobus Linnaeus), hemlock (Tsuga canadensis (L.) Carriere), and sugar maple (Acer saccharum Marsh). The understory at the site was dominated by jewel weed (Impatiens capensis Meerb.). The likely local origin of the earthworms is a housing development with ornamental and vegetable gardens some 250 m up-slope from the study site where the presence of A. agrestis, A. tokioensis and M. hilgendorfi was confirmed. Collections from CW for another project indicated that A. agrestis was dominant there accounting for 95% of specimens (Nouri-Aiin & Görres, 2019; Keller, Görres & Schall, 2017). In 2018, haphazard collections obtained exclusively A. agrestis. However, in the second year of this study, collections for bioassays also yielded large numbers of A. tokioensis.

Fungal isolate and preparation of liquid culture

We cultured pure B. bassiana from the commercial product (BotaniGard^® ES, GHM; BioWorks, Victor, NY, USA) on Potato Dextrose Agar (PDA, OXOID, Basingstoke, UK) in the dark at 23 ± 2 °C. Liquid culture inoculate was prepared, according to a protocol developed by Kim et al. (2014), by adding a conidial suspension (adjusted to 1 × 10⁸ conidia mL⁻¹ in a 250 mL flask) of B. bassiana from the PDA to one-quarter-strength Sabouraud dextrose broth (1/4 SDB; Difco™, Bordeaux, France). Flasks were held on a rotary shaker (150 rpm) at 24 ± 2 °C for 3 days. The treatment strength of 10⁸ conidia per mL was obtained by serial dilution.

Preparation of mycotized millet grains

We mycotized organic whole millet (Panicum miliaceum L.) grains (Healthy Living, South Burlington, Vermont) using a method modified from Hua & Feng (2005) and Bartlett & Jaronski (1988). Millet grains were selected as substrate for granular fungal formulation for two major reasons: it is a standard substrate for B. bassiana (Hua & Feng, 2005; Saranraj & Javaprakash, 2017) and it attracts earthworms as a food source (M. Nouri-Aiin and J. Görres, 2019, personal observation). Dry millet grains (250 g) were rinsed with distilled water and placed in autoclavable polyvinyl bags and soaked in 125 mL of water containing citric acid (0.4 mL⁻¹). All bags were flattened in a tray and autoclaved at 121 °C for 40 min and cooled to ambient temperature. Each bag of autoclaved grains was inoculated with a homogenized mixture of 5 mL of liquid culture broth and held for 14 days at 23 ± 2 °C and a 16:8 h L:D photoperiod. After 10 days when millet grains were entirely covered by B. bassiana, the bags were opened at ambient temperature until a moisture content of <5% was reached.

Mycotized millet grains were refrigerated at 4 °C in the dark until use. All batches of mycotized millet grains were assessed for conidial concentration with a Neubauer hemocytometer. To determine germination rate for each bag, four plates of Sabouraud dextrose agar (30 g in 1 L water) were inoculated with 0.1 mL of the conidial suspension and held in the dark at 25 °C for 24 h. Conidia with germination tubes longer than their width were considered germinated by inspecting 400 conidia per Petri dish (Hywel-Jones & Gillespie, 1990). Conidial yield was determined with a Neubauer hemocytometer to quantify the number of conidia per gram. Three batches with germination rates >98% and ≈ 1 × 10⁸ conidia per gram were used in the bioassays. It was not possible to get accurate estimates of the conidia count and germination rate because of the inert ingredients.

Experimental design

Three mesocosm bioassays were conducted to test the effect of the entomopathogenic fungus Beauvaria bassiana (BotaniGard^®) on juveniles of Amynthas spp. The first was conducted in the spring of 2018, the second and third in the spring and summer of 2019, respectively. The assays in the spring of 2018 and 2019 were conducted with the same batch of a BotaniGard^® obtained in April 2018. These two assays differed only in the age of the BotaniGard^®. For both of these assays fresh cultures were prepared as described above and the age of the juveniles was comparable. For assay 3 in the summer 2019, fresh product was purchased 4 weeks before the trial and, in this aspect, it resembled the first bioassay. However, the age of the juveniles was more mature in the summer 2019 assay. The bioassays were distributed over 2 years because these earthworms are annual species and develop fast so that similar life stages had to be collected over 2 years.

The treatments used in all trials included two controls (water, 15 g millet) and the following three treatments: suspension of pure B. bassiana cultured from the commercial product, 15 g and 25 g of mycotized millet (Table 1). To have a paired control for each treatment, an additional control of 25 g millet was added in the 2019 trials. In the June 2019 trial and the August 2019 trial, we added BotaniGard^® suspension prepared according to the label (1 × 10¹² spores/mL). Treatments and controls were each replicated 15 times. The suspensions were applied directly to the earthworms before adding them to the pots. Millet was added to the growing medium prior to the earthworms.

Experimental units were prepared using 1-gallon pots lined with zippered, fine-meshed laundry bags 11 × 15 inch (Aspire, LADB-EA70339 ordered through Opentip.com) to prevent worms from escaping. To each bag were added 1,000 g of organic potting mix (All Purpose Mix, Promix, Quakertown, PA, USA) and seven non-clitellated (juveniles) pheretimoid earthworms. The pots were then arranged on a growing bench at the center of a greenhouse. They were watered twice a week until the pots leaked leachate. In 2018, the earthworms were enumerated at the end of a 5-week period. In the summer of 2019, earthworms were counted every 3 to 4 days during the experimental period in order to calculate LT₅₀.

Analysis of mortality

Prior to analysis we used the arcsine-square-root transformation [x′ = asin(sqrt(x))] to normalize mortality data, followed by the Kolmogorov-Smirnov test (KS test) to confirm normality. The arcsin square root transformed mortality data sets for each treatment were normal at the 0.05 level. In all trials, ANOVA was used to test whether there were any differences in mortality data among treatments at the end of the experiments. Tukey’s Honestly Significant Difference (HSD) test was used to test whether the treatments were significantly different from the appropriate control. While statistics was done on arcsine-square-root transformed data, in Fig. 2 we report untransformed mortality data.

In addition, Abbott’s formula was applied to calculate mortality rates corrected for the control response (Abbott, 1925) to calculate LT₅₀ and LT₉₉:

(1) $$M_{\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{r}}={\displaystyle \frac{⟨M_T⟩-⟨M_c⟩}{1-<M_c>}}$$

where M_corr is the corrected mortality, <M_T> is the average observed treatment mortality, and <M_c> is the average observed mortality in the control.

The variance of M_corr, Var (M_corr) was calculated following Rosenheim & Hoy (1989) as

(2) $$Var\left(M_{corr}\right)={\displaystyle \frac{1}{{\left(1-⟨M_c⟩\right)}^2}\left[{\displaystyle \frac{Var\left(M_T\right)}{n}+{\left({\displaystyle \frac{1-⟨M_T⟩}{1-⟨M_c⟩}}\right)}^2{\displaystyle \frac{Var\left(M_c\right)}{n}}}\right]\left[2\right]}$$

where n is the number of replicates, Var (M_T) and Var (M_c) are the variances of the treatment and control mortalities.

In the third assay mortality was recorded every 3 to 4 days starting on the second day of the experiment to calculate lethal time LT₅₀ and LT₉₉ assuming a logit function. These were estimated using R package Ecotox (Hlina et al., 2019).

Sporulation and germination rates of mycotized millet grains

Fungal mycelium started to grow gradually on the inoculated millet grains after 3–4 days. After 10 days all grains were fully covered by the fungus (Fig. 1). For each assay, the three bags with highest conidial production and germination rate > 98% were selected. Conidial densities per gram of grains for the selected bags for assay 1 were 3.1 × 10⁸, 1.9 × 10⁸ and 2 × 10⁸, assay 2, 2 × 10⁸, 2.1 × 10⁸ and 1.9 × 10⁸, and assay 3, 1.9 × 10⁸, 2 × 10⁸ and 2.3 × 10⁸.

Mortality of earthworms

The Kolmogorov–Smirnov test showed the arcsine-square-root transformed distributions for each treatment were not significantly different from normal distribution. ANOVA showed that there were differences in mortality among treatments in assay 1 and assay 3, but not in assay 2 (Table 2). In assay 1, mycotized millet grain treatments were significantly different from the control millet (HSD: (P < 0.0001) for the 15 g mycotized millet and (P < 0.0001) for the 25 g mycotized milled). Likewise, the recultured B. bassiana treatment differed significantly from the control (P < 0.0001). The fungal treatments showed mortalities between 72 ± 6% and 78 ± 6%. Compared with the controls that varied between 22 ± 6% 12 ± 4% and (Fig. 2A).

Assay 3 obtained similar results to assay 1 (Table 2). Comparison of millet 15 g and mycotized millet 15 g showed no significant difference (P = 0.079). The comparison between water and BotaniGard^® was also not significant (P = 0.939). The comparison between B. bassiana and water was also significant (P = 0.036). Comparison between millet 25 g and mycotized millet 25 g was significant (P < 0.0001). The recultured B. bassiana suspension added directly to the earthworms prior to placing them in the soil had the highest mortality (96% ± 3%) which was significantly different from all other fungal treatments (P < 0.043) (Fig. 2C). Lowest mortality of the fungal treatments was recorded for 15 g mycotized millet (45% ± 7%) among the treatments. However, when comparing the commercial BotaniGard^® treatment in this experiment, there was no statistical difference between the mortality with commercial BotaniGard^® (59% ± 9%) and the untreated control (71% ± 7%).

There were no significant differences in mortality in assay 2 between controls and treatments in which 1 year old product was used (Table 2). Mortalities varied between 11% and 20% (Fig. 2).

Effect of B. bassiana on earthworm development

While there was no difference in mortality found for treatments in assay 2 (Table 2), there were differences in the development of the earthworms. In all millet treatments some earthworms reached maturity (Table 3), but fewer did so in mycotized millettreatments at the end of the assay (ANOVA, df = 5, F = 34, P > 0.001). There were significant differences between 15 g mycotized millet (4% ± 8%) and the 15 g millet control (40% ± 29%), and between 25 g mycotized millet (50% ± 30%) and the control with 25 g millet (73% ± 24%). The more millet, the greater the advance from juvenile to adult. And, more earthworms advanced to adult in the control treatments than the fungal treatments.

In assay 1 only earthworms in the control with 25 g millet reached maturity (87% ± 35%).

In the third assay, in which juveniles were further along in their development, adults were observed in all treatments but with very significant differences among treatments (ANOVA, df = 5, F = 53, P > 0.001). Controls with 15 g and 25 g millet, not mycotized, had the largest number of adults: 75% ± 16% and 76% ± 13% respectively. In contrast, in treatments with B. bassiana occurrence of adults were lower: 2% ± 6% for recultured B. bassiana, 15 g mycotized millet 33% ± 27% for 15 g of mycotized millet, and 26% ± 28%, for the 25 g mycotized millet treatment. The untreated control scored a similar development rate to the fungal treatments (20% ± 21%).

LT₅₀ and LT₉₉ values calculated with Abbott’s correction are listed in Table 4. The shortest LT₅₀ and LT₉₀ were observed for the recultured B. bassiana suspension added directly on earthworms (16 days and 45 days respectively). In contrast, lethal times were longer for BotaniGard^® (30.2 and 76.1 days). The more mycotized millet is added, the shorter the lethal times.