Drosophila species learn dialects through communal living

Abstract

Many species are able to share information about their environment by communicating through auditory, visual, and olfactory cues. In Drosophila melanogaster, exposure to parasitoid wasps leads to a decline in egg laying, and exposed females communicate this threat to naïve flies, which also depress egg laying. We find that species across the genus Drosophila respond to wasps by egg laying reduction, activate cleaved caspase in oocytes, and communicate the presence of wasps to naïve individuals. Communication within a species and between closely related species is efficient, while more distantly related species exhibit partial communication. Remarkably, partial communication between some species is enhanced after a cohabitation period that requires exchange of visual and olfactory signals. This interspecies "dialect learning" requires neuronal cAMP signaling in the mushroom body, suggesting neuronal plasticity facilitates dialect learning and memory. These observations establish Drosophila as genetic models for interspecies social communication and evolution of dialects.

Fig 1. A predator threat is communicated through visual cues within species across the genus Drosophila, modulating reproductive behavior and caspase activation.

(A) Standard experimental design. (B) Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Wild-type D. melanogaster (Canton S) exposed to wasps lay fewer eggs than unexposed flies. (C) Phylogeny of 8 species tested across the genus Drosophila that demonstrate the ability to communicate through visual cues. Green boxes indicate social learning is present in species tested. Representative ovary of control and wasp exposed Drosophila showing caspase activation (D. melanogaster). DAPI (D, H), activated Dcp-1 (E, I), WGA (F,J), and the merged images (G, K) are shown. Arrows denote apoptotic egg chambers. Error bars represent standard error (n = 12 biological replicates) (*p < 0.05).

Fig 2. Interspecies communication of predator threats.

Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Flies exposed to wasps lay fewer eggs than unexposed flies. Communication between D. melanogaster and: D. simulans (A, B), D. ananassae (C, D), D. kikkawai (E,F), D. willistoni (G,H), D. equinoxialis (I,J) and D. virilis (K,L) shows varying communication abilities. Communication between D. virilis and D. mojavensis occurs (M,N). Error bars represent standard error (n = 12 biological replicates) (*p < 0.05).

Fig 3. Flies raised in isolation have difficulty learning from socialized teachers.

(A) Experimental design of isolation of L1 larvae. Flies are allowed to lay for 24 hours, after which the adults are removed. Eggs are allowed to hatch and L1 larvae are isolated in an individual tube and allowed to eclose. A single isolated female fly is paired with a single isolated male and used as students to a single male and single female fly teacher that are socialized throughout life. Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Communication between socialized D. melanogaster teachers and: D. melanogaster, socialized students showing strong communication (B) and D. melanogaster students raised in isolation showing partial communication ability (C). Error bars represent standard error (n = 24 biological replicates) (*p < 0.05).

Fig 4. Species cohabitation enables inter-species communication.

(A) Experimental design of dialect training for flies that are used as students. Two species are cohabitated for one week prior to being used as students for naive, untrained teacher flies of the opposite species. Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Communication between trained students D. melanogaster and: D. ananassae showing strong communication following cohabitation (B, C), D. kikkawai showing strong communication following co-incubation (D, E), D. willistoni showing partial communication following co-incubation (F,G), D. equinoxialis showing partial communication following co-incubation (H,I), and D. virilis showing no communication following co-incubation (J,K). Error bars represent standard error (n = 12 biological replicates except for (C), n = 24 replicates) (*p < 0.05).

Fig 5. Flies can learn multiple dialects.

(A) Experimental design of dialect training for flies that are used as students using multiple three unique species. Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Communication between D. melanogaster students trained by D. ananassae and D. willistoni, shows that D. melanogaster learn each species dialect even in the presence of more than one species (B, C). Communication between D. ananassae students trained by D. melanogaster and D. willistoni, shows that D. ananassae learn each species dialect even in the presence of more than one species (D, E). Communication between D. willistoni students trained by D. melanogaster and D. ananassae, shows that D. willistoni learn each species dialect even in the presence of more than one species (F, G). Error bars represent standard error (n = 12 biological replicates) (*p < 0.05).

Fig 6. Dialect training requires multiple sensory inputs.

(A) Experimental design of dialect training for flies that are used as students using only visual cues (panels B,C). Flies only see each other through the duplex, with no direct interaction. Two species are co-incubated for one week prior to being used as students. Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Communication between trained students D. melanogaster and D. ananassae with training through visual cues only, shows that visual cues are not sufficient (B, C). Communication between trained students D. melanogaster and D. ananassae with training through monochromatic, red light only, shows a lack of dialect training (D, E). Communication between trained students ewg^NS4, mutant flies, and D. ananassae shows that moving wings are necessary (F, G). Communication between trained students Orco¹ and D. ananassae shows that olfactory cues are necessary (H, I). Communication between trained students Ir8a¹ and D. ananassae shows that Ir8a is a necessary receptor (J, K). Communication between trained students D. melanogaster and D. ananassae with training by male D. melanogaster only or by female D. melanogaster only, is not sufficient for dialect training (L,M). Error bars represent standard error (n = 12 biological replicates) (*p < 0.05).

Fig 7. Genetic perturbations reveal a critical role of the mushroom body and memory proteins for dialect learning.

(A) Experimental design of dialect training for flies being fed RU486 or methanol that are used as students. Both species are fed either RU486 or methanol during dialect training. Two species are co-incubated for one week prior to being used as students for naive, untrained teacher flies of the opposite species. Standard Drosophila media is used once the training period is over. Percentage of eggs laid by exposed flies normalized to eggs laid by unexposed flies is shown. Communication between trained students D. melanogaster and D. ananassae trained by flies expressing tetanus toxin (UAS-TeTx) in the mushroom body (MB) shows that the MB serves a critical role during the training period. D. ananassae learn from D. melanogaster with an inhibited MB, demonstrating that a functional MB is not needed to confer information during the training period (B, C). Communication between trained students Orb2^ΔQ and D. ananassae shows that Orb2 is required in students, but is dispensable for teachers to D. ananassae (D, E). Communication between D. ananassae and students co-incubated with D. ananassae that have RNAi-mediated Orb2 knockdown in the MB through RU486 feeding shows that the MB requires Orb2 during the training period (F). Communication between D. ananassae and students co-incubated with D. ananassae that have RNAi-mediated FMR1 knockdown (strain #24944) in the MB through RU486 feeding shows that FMR1 is not required in the MB during the training period (G). Communication between D. ananassae and students co-incubated with D. ananassae that have RNAi-mediated PTEN knockdown in the MB through RU486 feeding shows that PTEN is required in the MB during the training period (H). Error bars represent standard error (n = 12 biological replicates) (*p < 0.05).

Fig 8. Phylogenetic summary of dialect learning and pathway model for interspecies social learning.

Utilizing species across the genus Drosophila (A) demonstrates conservation of oviposition depression following wasp exposure, mediated by activated Dcp-1 to varying degrees and with varying expression patterns. The ability to communicate with D. melanogaster and the ability to demonstrate interspecies communication varies across the genus, with species closely related to D. melanogaster able to communicate without barriers. More distantly related species have difficulty communicating, though the barrier can be alleviated with dialect training. Finally, some species are too distantly related to communicate even after dialect training. Double boxes in a given row and column indicate multiple wild-type strains were tested. Interspecies communication is dependent on the presence of both male and female flies, the visual and olfactory systems, the mushroom body, and various long-term memory gene products (B). This model is based of the use of D. melanogaster and D. ananasse. Alleles tested in (B) are: Orco[1], Ir8a[1];Ir25a[2];Orco[1], Ir8a[1], Ir25a[2], ninaB[P315], Orb2ΔQ, ewg[NS4], Orb2[RNAi], PTEN[RNAi], FMR1[RNAi], and UAS-TeTx.