Effects of a heat wave event on the chemical ecology of species interactions in the potato agroecosystem

Heat waves, brief periods of unusually high temperatures, are increasing in frequency and intensity globally. Such extreme weather events can alter plant chemistry, disrupting species interactions that contribute to pest suppression or increase their performance. Yet, most heat wave studies focus on pairwise interactions, leaving us with a poor understanding of how complex agroecosystems respond to temperature extremes. We addressed this knowledge gap by simulating an experimental heat wave in the field on potato plants (Solanum tuberosum) and the Colorado potato beetle (CPB), Leptinotarsa decemlineata (Coleoptera: Chrysomelidae), in the presence or absence of their mutualistic microbial symbionts and another pest, the potato aphid (Macrosiphum euphorbiae). Here we report changes in CPB performance and leaf chemistry, specifically, volatile organic compounds (VOCs) and glycoalkaloids content in host plants.

Methods
Experimental design

To assess the effect of a heat wave event on pest performance and plant chemistry in different ecological contexts, we conducted a field experiment during the 2022 growing season at the Michigan State University Kellogg Biological Station (KBS) (Hickory Corners, MI). We generated a 4-day heat wave event on potato plants exposed to damage by CPB larvae with or without microbial symbionts and/or simultaneous damage by potato aphids.

CPB colony

All larvae were obtained from a colony maintained at the KBS greenhouse facilities with adults sourced from the Montcalm Research Station, MI. We obtained CPB without symbiotic bacteria by splitting the colony into separate cages with antibiotic-free and antibiotic-treated potato plants. Antibiotic-treated plants were sprayed thoroughly with an antibiotic solution consisting of 5 % Tetracycline, 5 % Neomicine, and 2.5 % Streptomycine in DI water.

Field experiment

We planted seed potatoes individually within 1 m2 plots in mid-May. When plants began sprouting and before naturally occurring CPB populations arrived, we allowed a subset of plants to be colonized by natural populations of potato aphids and covered the rest with 1 m2 by 80 cm tall mesh cages (Lumite, Inc., GA). When plants were about six weeks old, they were assigned to one of the following herbivory treatment combinations: 1) “control” plants with no herbivores; 2) “CPB” alone; 3) CPB treated with antibiotics “CPB(ab)”; 4) CPB with aphids “CPB+aphid”; and 5) antibiotic-treated CPB with aphids “CPB(ab)+aphid”. Plants in the “CPB” and “CPB(ab)” groups received 10 CPB neonate larvae from the antibiotic-free and the antibiotic-treated colony, respectively. For the “CPB+aphid” and “CPB(ab)+aphid” treatment groups, we added CPB larvae from their corresponding colonies as described above to the aphid colonized plants and covered with the protective cages.

Heat wave treatment

After one week of initiating bioassays, half of these plants were assigned randomly to either an ambient or heat wave group. Each herbivore x heat wave treatment combination had 15 replicates, totaling 120 plants. We simulated a single 4-day heat wave event in mid-July using an electric 300-Watt ceramic heater (Tempco, Inc., IL) hung in the opening of a pyramidal open-top chamber with a wood frame and anti-condensate greenhouse plastic sides (6 mil, 91 % light transmittance, Poly-Ag Corp., CA). This system increased daily temperatures by an average of 10 °C and 4 °C at night, achieving average temperatures of 41.84 °C ± 9.86 (diurnal) and 25.16 °C ± 4.2 (nocturnal) in the heat wave plots compared to 31.8 °C ± 9.26 (diurnal) and 21.91 °C ± 3.43 in ambient plots.

Data collection

Insect performance

We censused CPB larvae twice, once before the heat wave and once after to quantify performance (searching for 5 minutes per plant). Adult beetles were collected in mid-August prior to removing experimental cages, frozen immediately after collection, oven-dried and sexed. We then weighted them and measured pronotum length to assess adult performance.

Plant chemistry

We analyzed potato’s four main glycoalkaloids (solanine, chaconine, dehydrosolanine, and dehydrochaconine), and leaf volatiles to assess whether heat stress changes plant secondary chemistry under different types of biotic stresses. Glycoalkaloids were extracted from plant tissue and analyzed using LC/MS/MS at the Michigan State University’s Mass Spectrometry and Metabolomics Core.

Approximately 100 mg of plant tissue from each plant were collected in pre-weighted 2 mL tubes. Frozen tissue was ground in a pre-frozen bead beater (Retsch, MM400) at 30/s until fully ground. Samples were extracted with 1 mL extraction buffer (80:20 v/v methanol:water, 0.1% formic acid, and 100 nM of digitoxin as internal standards). After incubating at 4°C on a rocking platform for 16 hours, samples were centrifuged at 4 °C for 10 minutes at 14,000 rpm (Eppendorf, 5810 R). Supernatants (10 µL) were diluted into a 990 µL cold extraction buffer and stored at -20 °C. Samples were analyzed using a Waters Xevo G2-XS Quadrupole-Time-of-flight LC/MS/MS system with a Waters Acquity BEH-C18 UPLC column (2.1 x 100 mm) in positive ion mode. Compounds were eluted using a binary gradient of solvent A (0.1% formic acid in water) and solvent B (acetonitrile) at a flow rate of 0.3 mL/ minute at 40°C following a stepwise gradient: 98.0 % A, 2.0 % B; 0.50 min, 85.0 % A, 15.0 % B; 5.00 min, 40.0 % A, 60.0 % B; 7.00 min, 1.0 % A, 99.0 % B; 8.00 min, 1.0 % A, 99.0 % B; 8.01 min, 98.0 % A, 2.0 % B; 10.00 min, 98.0 % A, 2.0 % B. We identified compounds in the Waters MassLynx software based on mass spectrometry, confirmed with digitoxin as internal standard, and quantified using the Waters Quanlynx MS software. Prior to statistical analysis data was normalized to internal standards and tissue sample mass.

Leaf volatiles were collected immediately after the heat wave event over a period of three days from 118 plants including at least 10 replicates per each treatment combination. Sampling occurred between 11 AM and 2 PM during which time temperatures ranged from 23-43 °C following dynamic headspace sampling procedure. Briefly, we selected one vegetative stem per plant of similar biomass and enclosed them in a 35 x 35 cm nylon oven bags (Reynolds, USA) unsealed on one side. The bags were tied with twisters and volatiles were allowed to equilibrate for 30 minutes, after which, samples were pumped for 30 minutes through a scent trap using an air sampling pump (Ointik; Push-Pull Active Air Sampling Vacuum Pump) set to a pre-trap flow rate of 350 mL/min. Stems were saved, dried, and weighted. Ambient controls (n = 10) were taken from the area adjacent to the experimental plots where air was collected using an empty oven bag and sampled following the same methodology as the experimental samples. These ambient samples were used to identify contaminants or background compounds from surrounding vegetation.

Leaf and ambient samples were analyzed using thermal desorption Gas Chromatography-Mass Spectrometry (GC-MS), together with one blank/unused scent traps to detect potential contaminants in the trap system. Peak deconvolution, integration, and tentative compound identification were performed in the Automated Mass Spectral Deconvolution and Identification System using the 2020 NIST mass spectral library. Data filtering was performed in the bouquet package. Peaks were included if they had mass spectral match scores greater than 80 %, a maximum peak area of at least 20,000 counts, and if they were present in more than 20 % of the samples. Additionally, we only included compounds with peak areas four times higher than the mean area of the ambient and blank controls. Caprolactam (compounds present in oven bags) and compounds with high retention times (above 15 minutes) were excluded as contaminants. Stems used for volatile collection were saved, dried, and weighted. Volatile emissions were quantified based on peak values and were standardized by dry weight of the sampled stem.