Fig. 2: EVs show species-selective binding and impact on microbial community.

a Representative confocal microscopy images of Synechococcus sp. FACHB-1061 (up) and Microcystis aeruginosa FACHB-905 (down) co-incubation with DiO-labeled EVs derived from XL92 (Flavobacterium sp.) (scale bar, 2 μm). b Representative confocal microscopy images of field species from XLR co-incubation with DiO-labeled EVs derived from XL92 (Flavobacterium sp.) (scale bar, 5 μm). c The scheme for tracking DiO-labeled EVs uptake or conglutination by aquatic organisms. d The bar graph showed the results of 16S (left) and 18S (right) rRNA gene sequencing for species with a preference for taking up or attaching EVs (group P, n = 3) compared with the control group (group N, n = 3). The data were representative of three independent experiments. e The scheme of the microcosmic experiment setup. The four experimental groups and one control group lasted for 7 days. The additional amounts of EVs and bacteria were determined according to the environmental concentration. f The clustering tree (Bray-Curtis dissimilarity) showed their similarity of community structure. Group P represents the sorted microbial fraction that successfully takes up DiO-labeled EVs, while group N represents the fraction that did not exhibit EV fluorescence. In the microcosmic experiments, group A, B, C, D, and E represent the different group which were treated with strain XL92 bacterial cells, intact EVs derived from XL92, ruptured EVs derived from XL92, XL92 bacterial supernatant depleted of EVs, and control with PBS, respectively. g The bubble diagram showed the pathway enrichment analysis from functional predictions for prokaryotes by PICRUSt2. The color and the size of the bubbles represented the P value and the abundance of pathways, respectively.