Experimental procedure
Twenty queenright colonies were ordered from Biobest (Westerlo, Belgium) in January 2022. Colonies were kept in a climate chamber at 26° C and 70% humidity under a constant, inverted 12:12 h light:dark cycle. At the start of the experiment, we removed the queen and ten randomly selected workers from each colony. To observe colony initiation, we placed them in fresh boxes (21 x 13 x 27 cm) provided with sugar water (50% Apiinvert, Südzucker AG, Mannheim, Germany) ad libitum. Half of the queen-worker groups were randomly assigned to the control group receiving pollen cakes made from ground pollen (Imkerpur, Osnabrück, Germany) and Apiinvert (ratio 5:1 (g:ml)) (Figure A1A). The other half we assigned to the DEP treatment receiving pollen cakes made from ground pollen spiked with 0.2% DEPs (w/w) and Apiinvert (ratio 5:1 (g:ml) (Figure A1B). At the beginning of the experiment each colony received 4 pellets (ca. 1.5 g each) of pollen cake. Afterwards, we added two pollen pellets to the colonies twice per week. We replaced the sugar water once per week to prevent molding. Once per week, we removed dead workers. On day 35, 43 and 63 we randomly removed 20% of the workers from each colony to simulate a natural mortality rate and prevent overcrowding (Otti and Schmid-Hempel 2007). After 66 days all colonies were frozen at -20° C.
After defrosting, we counted the number of workers, larvae, and pupae. For each colony, we pooled all larvae and pupae in 10 ml glass vials, respectively, and placed them at 70° C for 96 hours in a drying oven (UFE 600, Memmert GmbH + Co. KG, Schwabach, Germany) to measure larvae and pupae dry weight using a fine scale to the nearest 0.01 mg (SM1265Di, VWR, Darmstadt, Germany). We calculated brood number and dry weight by adding numbers and dry weights of larvae and pupae, respectively.
Fat body assay
We randomly selected ten workers per colony and measured the radial cell length of the right forewing under a light microscope (Leica DFC290, Leica M165, Leica, Wetzlar, Germany) to determine worker size (Medler 1962). Additionally, we measured the fat body content of the selected workers according to Hüftlein et al. (2023). For that purpose, we separated the abdomen from head and thorax and slightly opened the sternites with tiny scissors. Then we placed each abdomen in a 5 ml glass vial and dried it in a drying oven (UFE 600, Memmert GmbH + Co. KG, Schwabach, Germany) at 70° C for 72 hours. Then, the dry weight including the fat body of each abdomen was weighed on a fine scale to the nearest 0.01 mg (PLE 420-3N, Kern & Sohn GmbH, Balingen-Frommern, Germany). Then we removed the fat content added 2 ml chloroform to each vial. Every 24 hours we replaced the chloroform with 2ml fresh one. After 72 hours we removed the chloroform and placed the vials in the drying oven again at 70° C for 72 hours. Then we weighed the abdomens again without the fat body. We calculated the fat body content by subtracting the dry weight after fat body removal from the initial dry weight of the abdomen. To correct for body size, we calculated relative fat body weight by dividing the fat body content by the radical cell length.
Generation of diesel exhaust particles (DEPs)
We produced DEP-samples on a test bench with a four-cylinder diesel engine (OM 651, Daimler AG, Stuttgart, Germany) as described in Zöllner (2019). When analyzed with a fast response differential mobility particulate spectrometer DMS500 (Combustion, Cambridge, England), raw exhaust samples showed a median diameter between 52.1 nm and 101.9 nm within an inner-city driving cycle. Elemental carbon accounted for 72.2 ± 1.1% of the DEP mass, organic matter for 23.2 ± 0.9% w/w, and inorganic matter for 4.6 ± 0.7% w/w, according to thermogravimetric analysis (TGA, STA 449 F5 Jupiter, Netzsch-Gerätebau GmbH, Selb, Germany). We analyzed the DEP-samples for polycyclic aromatic hydrocarbons (PAH) and detected a concentration of 444 ppm for pyrene, 220 ppm for phenanthrene, and 107 ppm for fluoranthene. The elemental composition showed fractions of calcium (1.63% w/w), copper (1.03% w/w), zinc (0.53% w/w), phosphorus (0.50% w/w), sulphur (0.17% w/w), boron (0.13% w/w) and magnesium (0.10% w/w). More details on the characterization and sampling methods of DEP-samples can be found in Hüftlein et al. (2023) and Seidenath et al. (2023).
Statistical analysis
All statistical analyses were performed using R 4.2.1 (R Core Team 2022). Worker number, larvae number, pupae number, brood number (sum of larvae and pupae), larvae dry weight, pupae dry weight and brood dry weight (total weight larvae and pupae) were analyzed fitting generalized linear models (GLMs) with treatment as a predictor. Relative fat body weight and worker size were analyzed fitting generalized linear mixed models (GLMMS) with treatment as a predictor and colony as a random factor using the package glmmTMB (Brooks et al. 2017). We checked model assumptions using model diagnostic test plots, i.e., qqplot and residual vs. predicted plot from the package DHARMa (Hartig 2022). For the GLMs we then produced F-statistics with the function Anova() from the package car (Fox & Weisberg 2019) to calculate p-values for differences between the two groups. For the GLMMs we performed chi-squared tests with the function Anova() from the package car (Fox & Weisberg 2019) to calculate p-values for differences between the two groups.