Phleum

This study was carried out in the Jena Experiment, which is a long‐term grassland biodiversity experiment (Roscher et al., 2004 (link)). The study site is located in the floodplain of the Saale river near the city of Jena (Thuringia, Germany, 50°55′ N, 11°35′ E, 130 m.a.s.l.) and had been used as a high‐fertilized arable field for growing wheat and vegetables until the biodiversity experiment was established in 2002. The soil is a Eutric Fluvisol, while soil texture changes from sandy loam to silty clay with increasing distance from the river on the experimental site. The study site was divided into four blocks to account for differences in soil texture, while blocks were arranged parallel to the riverside (Roscher et al., 2004 (link)). The mean annual air temperature was 9.7°C and the mean annual precipitation was 574 mm from 2003 to 2016, which was recorded with a meteorological station at the study site (Weather Station Jena‐Saaleaue, Max Planck Institute for Biogeochemistry Jena, https://www.bgc‐jena.mpg.de/wetter/).
For the present study, the Dominance Experiment was used, which was a sub‐experiment of the Jena Experiment (Roscher et al., 2004 (link)). The species pool of this experiment consisted of nine plant species, which often reach dominance in Central European mesophilic grasslands of the Arrhenatherion type (Ellenberg, 1988 ): five grass species (Alopecurus pratensis L., Arrhenatherum elatius (L.) P. Beauv. ex J. Presl et C. Presl, Dactylis glomerata L., Phleum pratense L., Poa trivialis L.), two legume species (Trifolium pratense L., T. repens L.), and two forb species (Anthriscus sylvestris (L.) Hoffm., Geranium pratense L.). Species richness levels ranged from one to nine species (1, 2, 3, 4, 6, and 9 plant species plots), while each species and every species pair occurred the same number of times at each species richness level. All species compositions were replicated twice (i.e., same mixture identity), with the exception of the nine‐species mixture, which was replicated eight times. There was the same number of plots per species richness level in each block, ensuring that replicates with identical species composition were distributed in different blocks. From the year of establishment (2002) until 2009, plants were grown in plots of 3.5 × 3.5 m, while plot size was reduced to 1 × 1 m in 2010. Seeds for the establishment of the experiment were purchased from a commercial supplier (Rieger‐Hoffman GmbH) and were sown in May 2002 with a density of 1000 viable seeds per m². One species, A. sylvestris, which failed to establish in the first growing season, was re‐sown with half density in late autumn 2002 (Roscher et al., 2004 (link)), while no further re‐sowing was done later. All plots were mown every year in June and September (mown plant material was removed), were regularly weeded to maintain the sown species compositions, and have never received any fertilizer.
To keep the number of samples and measurements manageable, we used the 1‐, 2‐, 6‐, and 9‐species plots of the Dominance Experiment (85 plots out of 206 plots). Due to very low amounts of standing biomass in some monocultures, we decided to carry out destructive measurements in only one of the two monocultures per species (with the exception of aboveground biomass, which was measured in both monocultures), so that the other can still be sampled in the future. Furthermore, the monocultures of the grass species P. pratense and the forb species A. sylvestris showed no biomass production in 2016/2017, and both species were extinct or had a very low biomass in all other plots, so that we did not sample these monoculture plots and did not measure any functional traits of these two species in mixtures (i.e., these two species were excluded from the analyses). For all remaining plant species, we conducted measurements in each of the seven monoculture plots, a subset of the two‐species mixtures (= 46 plots), and all 6‐ and 9‐species plots (= 24 and 8 plots, respectively; Table 1). In case of the two‐species mixtures, we used all existing two‐species combinations of the seven species (both replicates), and one replicate with A. sylvestris and P. pratense, respectively, although there are some exceptions due to local extinctions (Table S1). Overall, each of the seven species was present nine to 12 times in the two‐species plots (for detailed information see Table S1), 16 times in the six‐species plots, and eight times in the nine‐species plots. Because of the extinction of several plant species, we counted how many of the originally sown plant species were actually growing in the plots in May 2017 and used this “realized plant species richness” as another explanatory variable, in addition to sown plant species richness.

Recombinant pollen allergens from birch (Betula pendula) and timothy grass (Phleum pratense) pollen allergens Bet v 1.0101 and Phl p 5.0101, referred to as Bet v 1 and Phl p 5 hereafter, were obtained from Biomay AG (Vienna, Austria). Ovalbumin (OVA) was purchased from InvivoGen (Toulouse, France) and was treated the same way as the allergens to serve as a negative control in the cell culture experiments described below. Protein stock solutions (1 mg mL⁻¹) for chemical modification were prepared with pure water as described in Backes et al. (24 (link)).

PNL were isolated from commercially available roasted peanuts as previously described (Palladino et al., 2018 (link)). The same protocol was used to extract lipids from raw peanuts (PNL-Raw), and Timothy grass (Phleum pratense) pollen (GPL). Lipids were also extracted from peanuts boiled in water for 30 min (PNL-B). PNL and GPL were fractionated as follows: the lipid extracts diluted in chloroform were loaded onto thin-layer chromatography (TLC) plates (TLC AI foil, Sigma–Aldrich, MO, United States). TLC was conducted as previously described (Palladino et al., 2018 (link)), and visualized by iodine (Sigma–Aldrich, MO, United States). Once separated, fractions were scratched from the plate, and lipids were re-extracted with 10 mL of MeOH:CHCl₃ 1:2. Every fraction was evaporated by a nitrogen stream and weighed. Lipids resulted negative when tested for the presence of innate immune system contaminants by means of a cellular assay with the reporter cell lines THP1-XBlue and THP1-XBlue-MD2-CD14 (InvivoGen, San Diego, CA, United States). Glyceryl trioleate and oleic acid were purchased from Sigma–Aldrich, MO, United States and used as control for lipids. PNL and GPL total extracts, as well as individual fractions and synthetic lipids were applied to 16HBE14o- cells at final concentrations of 100 μg/mL, or at a varying concentrations if specified, after 5 min of sonication in the medium.