Fig. 3

Investigation of the role of extracellular symbionts on D. melanogaster repellency and D. suzukii oviposition avoidance. (a) Axeny, the removal of extra-cellular microorganisms, had different effects on D. melanogaster and D. suzukii (n = 16, 16, 17, and 27 D. suzukii individuals, from left to right). Oviposition sites exposed to axenic D. melanogaster were not avoided by D. suzukii, showing the importance of symbionts in D. melanogaster for repellency. By contrast, axenic D. suzukii behaved like conventionally reared flies; D. suzukii microorganisms were therefore not required for perceiving the repellent. (b) Tests of candidate bacteria in association with D. melanogaster (n = 14, 13, 11, and 17 D. suzukii individuals, from left to right) revealed the bacterium Lactobacillus brevis can restore repellency in formerly axenic flies (note the axenic and Acetobacter pomorum treatments were marginally non-significant, p = 0.071 and p = 0.075, respectively). We hypothesized D. suzukii avoidance loss was due to their colonization with D. melanogaster symbionts. (c) As expected, D. suzukii females experimentally associated with the bacterium L. brevis did not avoid oviposition on sites exposed to L. brevis-associated D. melanogaster (n = 29 and n = 28 D. suzukii individuals for axenics and gnotobiotics, respectively). (d) Direct inoculation of medium with L. brevis cells in large numbers or at a dose similar to that naturally shed by D. melanogaster (i.e., 1,000,000 vs 5000) produced different results (n = 29, 29, 22, and 11 D. suzukii individuals, from left to right). The low, natural dose of deposited bacteria failed to elicit avoidance, suggesting D. melanogaster repellency is largely due to the production of unidentified molecules when in symbiosis. Symbols indicate means and error bars standard errors. Statistical tests produced by Wilcoxon signed rank tests; * for p < 0.05; ** for p < 0.01; *** for p < 0.001