Plants and insects
Commercial Bombus terrestris queens
Unmated Bombus terrestris queens were obtained from commercially available queenright colonies provided by the Biobest group NV (Westerlo, Belgium). Nests were provided with BIOGLUC® sugar solution (2.1 kg) and fed with lyophilized pollen (Bio-Blütenpollen, naturwaren-niederrhein GmbH, Germany) provided either as pure ground powder or mixed with nectar. Colonies were maintained in climate chambers in complete darkness (Kälte 3000: day/night cycle: 16:8, 24 h darkness; RH70%/80%; 27/26 °C). Founding queens were removed after we received the hives, in order to accelerate the production of new queens from developing brood as described by Lopez-Vaamonde et al.,26 (colonies arrived 27.04.2022; new queens eclosed 18.05.2022). Queens were readily distinguishable from workers based on size. Newly eclosed gynes were isolated and kept in small cages (W60 × D40 × H40 cm) in a climate chamber (Kälte 3000: light/dark cycle: 12:12, 150 μE; RH60%/70%; 22/18 °C;). Queens were deprived of pollen 3 days prior to experiments. A separate batch of unmated queens from commercially obtained colonies were mated with males (from a different commercially obtained colony) and used to confirm the persistence of damaging behavior following experimentally induced hibernation, using methods described by Röseler27.
Wild bumblebee queens
Wild bumblebee queens were caught using insect nets in Spring 2023 (end of February 2023 until end of July 2023). We collected queens from different regions of Switzerland (Fig.S2) and from different elevations: Haldenstein (Calanda 1400 m a.s.l., 2000 m a.s.l.), greater Zurich area (City Center 400 m a.s.l., Uetliberg 870 m a.s.l.), Baden 500 m a.s.l., Landquart 550 m a.s.l., Niederwil 405 m a.s.l. Delayed phenology at higher elevation shifted the timing of queen emergence to mid-July (Snow melt dates for 2000 m asl: mid-May). To avoid collecting queens that had already initiated a colony, we only kept queens without pollen bags and stopped collecting queens as soon as we observed flying workers at each collection site. Queens were anesthetized with CO2 for 1 min and identified under a microscope using the Fauna Helvetica Apidae 1 identification guide for Swiss bumblebees22. Individual queens were kept in small enclosures (W30 × D30 × H30 cm) in a climate chamber (Kälte 3000: light/dark cycle: 12:12, 150 μE; RH60%/70%; 22/18 °C) for at least 2 weeks and given access to nectar (Biogluc sugar solution, Biobest group) ad libitum. Pollen (Bio-Blütenpollen, naturwaren-niederrhein GmbH, Germany) was provided once a week, but during exposure to plants no pollen was available to ensure conditions of pollen-deprivation.
Plants
Brassica nigra plants were grown from seed provided by the Center of Genetic Resources in Wageningen, the Netherlands (accession number: CGN06619) and further propagated in the lab. Seeds were sown in plastic trays with “substrate 2 + greenfibre (120)” as soil (Klasmann-Deilmann GmbH, Geeste, Germany) and irrigated with water mixed with ~0.25% (v/v) Solbac (Andermatt Biocontrol Suisse AG, Grossdietwil, Switzerland). Trays were stored in darkness for 3 days at 4 °C to synchronize germination and subsequently kept in a climate chamber (Kälte 3000, light/dark cycle: 12:12, 150 μE; RH60%/70%; 22/18 °C). One week after germination, seedlings were transplanted into single pots (Desch PlantPak, 9 × 9 × 10 cm) and treated with ~0.25% (v/v) Solbac (Andermatt Biocontrol Suisse AG, Grossdietwil, Switzerland). Solanum lycopersicum var. Sibirian Early 34500 (tomato) plants were grown from seed obtained from Zollinger bio (Port-Valais, Switzerland). Seeds were sown in plastic trays filled with “substrate 2 + greenfibre (120)” (Klasmann-Deilmann GmbH, Geeste, Germany) and watered with water mixed with ~0.25% (v/v) Solbac (Andermatt Biocontrol Suisse AG, Grossdietwil, Switzerland). Trays were kept in a climate chamber (Kälte 3000: light/dark cycle: 16:8, 150 μE; RH60%/70%; 22/18 °C). One week after germination seedlings were transplanted into single pots (Desch PlantPak, 9 × 9 × 10 cm) and treated with ~0.25% (v/v) Solbac (Andermatt Biocontrol Suisse AG, Grossdietwil, Switzerland).
Experimental methods
Flowering-time assays with B. terrestris queens
In both flowering-time experiments, plants were randomly assigned to two treatment groups: queen-damage and mechanical damage (control) (B. nigra, n = 8 per treatment; S. lycopersicum, n = 7 per treatment). Each queen-damaged plant was paired with a mechanically damaged control plant, on which we tried to replicate the damage pattern as closely as possible using needles (Fig. 1a, b). Findings from our previous study8 indicate that mechanical damage can cause slight acceleration of flowering in S. lycopersicum, but that this is effect is consistently weaker than that observed for damage by B. terrestris workers. All plants were of uniform age at the start of the experiment (B. nigra, 24 days post germination (dpg); S. lycopersicum, 25 dpg) and had the same number of leaves (B. nigra, 8 true leaves; S. lycopersicum, 9 true leaves). Plants in the queen-damage treatment were placed into cages with individual, pollen-deprived queens (B. terrestris queens, n = 4, 2 or 3 plants per queen) in a climate chamber (Kälte 3000: light/dark cycle: 12:12, 150 μE; RH60%/70%; 22/18 °C). It took between several minutes and 24 h for queens to begin damaging individual plants, and plants were removed once queens were observed having damaged a plant. On average, Brassica plants received 9 ± 7 holes, Solanum plants received 2 ± 1. After treatment, plants were placed at random positions in their respective climate chamber (B. nigra, LD 12:12; S. lycopersicum, LD 16:8). Trays were moved to new random positions every 3 days. Plants were monitored daily for the development of macroscopic signs of flowering (B. nigra, first flowering primordium; S. lycopersicum, first open flower), and flowering time was assessed as “time (days) elapsed since damage treatment”.
Leaf-damaging assays with wild bumblebee queens
Individual flowerless plants of Brassica nigra (20–30 days post germination) or Solanum lycopersicum (30–40 days post germination) were placed in enclosures with individual wild bumblebee queens in a climate chamber (Kälte 3000, light/dark cycle: 16:8, 150 μE; RH60%/70%; 22/18 °C). The plant species a queen received was based on availability. Plants were inspected daily for new leaf-damage. Damage was documented by cutting damaged leaves and scanning them using the Epson Perfection V850 Pro Scanner. The following settings were applied in the Epson Scan Software: document type: film; film type: color positive film; image type: 24-bit color; resolution: 1200 dpi; adjustments: none; file type: TIFF. Leaves were scanned in batches every 2–3 days.
Leaf-damaging assays with wild bumblebee workers
In May 2023, we started collecting wild foraging workers from different species to form microcolonies, which could be used to test leaf-damaging behavior of the worker caste. 10–20 worker per species, originating from different colonies, were collected in the center of Zurich, kept together in a plastic box with sufficient ventilation in a climate chamber (Kälte 3000: day/night cycle: 16:8, 24 h darkness; RH60%/70%; 22/18 °C), and given around 10 days to establish a new hierarchy. Only B. pratorum workers formed a microcolony that could be used for damaging experiments. Microcolonies are usually formed by separating workers from their parental colony. Our method provides a way to circumvent the time-consuming process of rearing a colony first from wild queens. The colony was pollen-deprived for 3 days before being moved into a cage (W60 × D60 × H60 cm) in a climate chamber (Kälte 3000: light/dark cycle: 16:8, 150 μE; RH60%/70%; 22/18 °C) and presented with flowerless S. lycopersicum plants. Damaged plants were removed from the cage and the leaves were scanned as described above. A selection of leaf scans is found in the supplementary material.
Statistical analyses
All statistical analyses were carried out using R version 4.3.1 (R Core Team 2023)28. For flowering-time experiments, displayed in Fig. 2, we used an exact, two-sided Wilcoxon-Mann–Whitney Test to test for the equality of means between two independent groups. We used the function “wilcox_test ()” from the R package coin. Boxplots were plotted using the “ggboxplot ()” function from the ggplot2 package and p-values were added to the plot using the function “stat_pvalue_manual ()” function form the package rstatix and the p-values were calculated using the function “dunn_test ()” also from the package rstatix. A summary of the statistics can be found in the supplementary table 1 (ST1).
Ethical note
In Switzerland, there is no legislation regulating research with bumblebees; however, our experimental design and procedures were guided by the 3 R principles29. Bees received daily care by trained staff and were provisioned with adequate food. Behavioral tests were non-invasive, and we tried to minimize stress wherever possible.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.