Beetle Herding: Development of Strategies to Optimize Biological Control of Air Potato Using Attractants

Objective 1: Determine response of Lilioceris cheni to Dioscorea bulbifera (undamaged, mechanical, larval, and adult)

            To improve the air potato biological control program, attractive compounds to L. cheni were identified. Field observations of the beetles have indicated that they aggregate, suggesting they rely on odor mediated cues to form these aggregations. The first step of identifying attractive host plant volatiles of L. cheni was to conduct insect behavior studies. Specifically, the behavior of L. cheni was characterized in a wind tunnel. Wind tunnels are widely used in insect chemical ecology research and are a reliable, and predictable way to assess insect behavior. Treatments consisted of an undamaged plant, which acted as a control, mechanically damaged plants, and plants that have been fed on by larval or adult L. cheni. A fully crossed experimental design will be utilized to determine which treatment is most attractive to the adults. There were 20 replications per treatment, with 10 adult beetles per replicate and difference were assessed with a binomial test. Plants were changed between each replication and wind tunnel was cleaned with 70% ethanol to remove any residue from the test beetles.

Objective 2: Identification of chemicals governing L. cheni attraction to D. bulbifera

            Once the underlying insect behavior were characterized in objective 1, the chemicals which govern this behavior were identified. Volatiles from undamaged, mechanically damaged, larvae-damaged and adult-damaged plants were collected and analyzed. Chemical analysis and identification were primarily utilizing techniques for volatile collection and analysis with gas chromatography coupled to mass spectrometry (GC-MS). Mesh cages (0.3 m x 0.3 m) were used to enclose a single plant to receive the experimental treatment. Four treatments were evaluated: 1) adult damage; 2) larval damage; 3) mechanical damage; 4) undamaged, which acted as the control. There were 16 samples of each treatment collected and evaluated. Plants were placed inside the mesh cage and exposed to the different treatments and placed in a Percival incubation chamber at 22-25°C with a 14:10 L:D cycle for 24 h. Plants exposed to insect damage had either 10 unsexed adults or 10 larvae added to the cage. Mechanically damaged plants were damaged with a 1 cm diameter paper hole punch 10 times. Plants were removed from the cages and conspecifics were removed prior to placement in glass collection jars. Collection jars were created from 7.5 L, glass jars with 2 inflow ports drilled through the lid. Volatiles were collected for 6 h using a volatile collection system. Due to the short collection period, we were not concerned with breakthrough of the volatiles. External air was purified by charcoal-filter then pushed into the jar at 1 L/min and pulled at 1 L/min through a HayeSepQ filter connected to PTFE tubing. Volatile filters were cleaned prior to use with 500 µL of dichloromethane. After sampling, filters were eluted with 200 µL of dichloromethane into 2 mL glass vials with a 250 µL glass insert, with 32.4 ng of nonyl acetate as an internal standard. Vials were stored in a -20 °C freezer until they were analyzed by gas chromatography coupled with mass spectrometry (GC-MS). One µL of each sample was injected into the GC-MS using an autosampler. Helium was used as the carrier gas with a pressure of 8.5 psi and velocity of 1.1 mL/min. All samples were analyzed by a HP-5MS ultra-inert capillary column, 30 m long with 0.25 mm diameter. The column temperature was maintained at 40 °C for 4 min and increased at a rate of 10 °C/min to a final temperature of 300 °C for 5 min. The inlet temperature was set to 250 °C and the source temperature was set to 230 °C. Mass spectrometry was performed using electron impact at 70 eV. Chromatograms were integrated using the auto integration function from MassHunter Qualitative Analysis. Compound identification was based on authentic standards when available and tentative identification through the NIST 2020 spectral library.

Objective 3: Validation of lures and field trial of D. bulbifera semiochemical attractants

            The  volatiles identified in objective 2 were tested for behavioral activity in L. cheni with electroantennography (EAG). This was conducted to ensure the adults were able to detect the synthetic ocimene and farnesene racemic mixtures prior to field deployment. Beetles reared in laboratory settings were used to assess antennal responses of adult male and female L. cheni stimulation of ocimene or farnesene. Negative control treatment consisted of 10 µL of dichloromethane and semiochemical treatments consisted of 10 µL of 100 ng µL^-1 ocimene in dichloromethane, or 10 µL of 100 ng µL^-1 farnesene in dichloromethane, were tested. Treatments were applied in the following sequence: solvent control, ocimene or farnesene, solvent control, ocimene or farnesene, and solvent control, with approximately 30 s between each treatment and replicated 5 times per beetle sex.

EAG was performed on isolated antennae attached with electrode gel onto two metal electrodes. Treatments used were prepared by adding 10 µL of solution to a 14.5 cm Pasteur pipet . The solvent was given 60 s to evaporate from the Pasteur pipet before the tip of the Pasteur pipet was placed about 3 mm inside a small hole through the wall of a metal tube (13 cm x 1 cm). The metal tube was positioned in the direction of the antennal preparation with a continuous flow rate (1 L/min) of humidified, charcoal-filtered air from the stimulus controller with a stimulus duration of 1 s. A 90 s interval between stimulations was given to allow for the antennae to recover. After each EAG assay, the beetle was sexed based on dissected genitalia.

 Behavioral activity of the lures were evaluated in the field utilizing a field trial conducted in summer 2023 at the USDA: ARS facilities in Tallahassee, FL, USA. A field plot of 120 m x 50 m was created to support 50 potted 1 m tall D. bulbifera plants.  Each plant was spaced 10 m from another plant in an alternating pattern of a buffer (no treatment) and experimental plant (control or treatment). Based on previous volatile analyses, ocimene and farnesene mixtures were used singularly and in combination as experimental lures. Field lures were created by inserting a 2 cm cotton tube inside a microcentrifuge tube (1.5 mL) with 1 mL of mineral oil and 4 µL of one of the different treatments. Treatments evaluated were 1) ocimene; 2) farnesene; 3) combination of ocimene (2 µL) and farnesene (2 µL); 4) control (mineral oil only). The lures were covered in aluminum foil to prevent solar degradation of the semiochemicals, then attached to the metal posts using a 15 cm long wire. Lures were replaced weekly after data was collected and experimental trials lasted 4 weeks.

Data were collected on seven metrics: 1) visual count of adult L. cheni; 2) observed number of cupped leaves on terminal vines indicating oviposition occurrence (20,27,28); 3) number of un-cupped leaves on terminal vines; 4) larval damage to terminal vines; 5) count of undamaged leaves; 6) count of adult damaged leaves; and 7) count of larval damaged leaves. A cupped leaf is defined as a leaf that has been deformed by adult L. cheni to form a “cupping” shape around oviposition sites (Fig. 1A). The visual count of adults was conducted during a 1 min duration where the observer moved around the plant to count the adult L. cheni on the plant. Differentiation of adult and larval L. cheni damage was completed based on the different characteristics presented on the herbivore damaged leaves (Fig. 1B). Adult L. cheni herbivore damage had a pocked appearance on the leaves while larval L. cheni herbivore damage skeletonized the leaf causing it to turn brown and desiccate (Fig. 1B). To randomize the observation on each plant, a ribbon was tossed onto the plant with the end of it being used as the center of a metal hoop (0.3 m diameter) used to subsample leaves for L. cheni herbivore damage on the observed plant. Every leaf within the hoop (even partially) were categorized either as undamaged, adult damaged, or larvae damaged. If a leave was both adult and larvae damaged, the dominant type of damage was considered.  Two subsamples were taken with the metal hoop for each individual plant. Terminal vines are defined as the newly grown vines that produce tender new leaves preferred by L. cheni females to “cup” for use as ovipositional sites. A terminal vine was measured to 3 m and the leaves on the measured portion of the vine were assessed for cupped or un-cupped leaves as well as larval damage to the terminal vine.

For each field trial, each treatment was replicated six times with a total of six replicates for each sampling date. Data for the first field trial was collected weekly between 13 July 2023 to 3 August 2023 and collected weekly between 5 September 2023 to 26 September 2023 for the second field trial. ne week prior to the first weekly data collection, 200 adult beetles were divided into 3 groups and released in the field at 40 m intervals within the plot. Initial lures were also deployed at this time.