bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Large-scale characterization of sex pheromone communication sys- tems in Drosophila

Mohammed A. Khallaf 1‡, Rongfeng Cui 2, Jerrit Weißflog 3, Aleš Svatoš 3, Hany K.M. Dweck 1,4, Dario Ric- cardo Valenzano 2, Bill S. Hansson 1†, and Markus Knaden 1†‡

Insects use sex pheromones as a reproductive isolating mechanism to attract conspecifics and repel het- erospecifics. Despite the profound knowledge of sex pheromones, little is known about the coevolutionary mechanisms and constraints on their production and detection. Using whole-genome sequences to infer the kinship among 99 drosophilids, we investigate how phylogenetic and chemical traits have interacted at a wide evolutionary timescale. Through a series of chemical syntheses and electrophysiological recordings, we identify 51 sex-specific compounds, many of which are detected via olfaction. Behavioral analyses reveal that many of the 42 male-specific compounds are transferred to the female during copulation and mediate female receptivity and/or male courtship inhibition. Measurement of phylogenetic signals demonstrates that sex pheromones and their cognate olfactory channels evolve rapidly and independently over evolutionary time to guarantee efficient intra- and inter-specific communication systems. Our results show how sexual isolation barriers between species can be reinforced by species-specific olfactory signals.

Main blends of shared chemical compounds as a result of genetic similarities and biosynthetic pathways shared rganisms communicate with each other through by ancestry 7,8. This diversity in sex pheromone Oexchanging signals that include visual, acoustic, communication can become further affected by tactile and chemical (smell and taste) senses. The factors like geographical or host variations. For chemical sense is common in all organisms, from example, sympatric species develop pronounced bacteria to mammals, and therefore, regarded from an divergent communication systems to overcome the evolutionary perspective as the oldest one. Animals risk of hybridization, while the unimpeded divergence are surrounded by a world full of odors emitted from due to geographic barriers may lead to relaxed conspecific or heterospecific individuals, as well as accumulation of differences 9. Moreover, colonization from the environment. The ability to exchange and of a different host – an ecological adaptation – could decipher these signals has significant impact on a also lead to differential sex pheromones and new species’ success as odors help to avoid imminent ways of signal transmission and perception 10,11. threats and localize and judge food or potential Although many studies have reported the diversity mates. Olfactory systems have, therefore, evolved in of sex pheromones among related species, the a sophisticated way to meet the challenge of detecting evolutionary phylogenetic history of these traits and and discriminating a countless number of odorants. their detection systems remains obscure. While it is well established that, and how animals use Flies, like most animals, rely on odors for intra- and interspecific communication, the chemical cues to locate and choose an appropriate evolution of olfactory systems with respect to signal mating partner 1,12-14. For several reasons, flies within production and perception is poorly understood. the genus Drosophila represent ideal species to study One of the most crucial channels that the evolution and diversity of sex pheromones, as have been suggested to contribute to speciation are well as their associated behaviors. First, Drosophila the sex pheromone-sensing channels 1. Volatile sex species live in an extensive range of diverse habitats pheromones – airborne chemicals that stimulate across all climatic conditions, from deserts and caves sexual behaviors in the opposite sex – are the to mountains and forests 15. In these environments, primary signals that reinforce the isolation barriers drosophilids feed and breed on varied hosts such as between different species. These species-specific decaying fruits, slime fluxes, mushrooms, flowers, as signals often provide a full biography written in scent well as frog spawn 16. Second, sexual behaviors of molecules about the sender, such as information about drosophilid flies differ quantitatively and qualitatively reproductive and internal status. Diversification of sex 17, which may include nuptial gift donation 18, pheromones among species arises either via neutral partners’ song duet 19,20, territorial dating 20, or the drift, or via sexual, and/or natural selection 2-6. Closely release of an anal fluidic droplet 21. Third, the neural related species tend to use different pheromone processing of pheromones in the brain of some

1Department of Evolutionary Neuroethology, Max Planck Institute for Chemical Ecology, D-07745, Jena, Germany. 2Max Planck Institute for Biology of Ageing and, CECAD Research Center at University of Cologne, D50931, Cologne, Germany. 3Group of Mass Spectrometry and Proteomics, Max Planck Institute for Chemical Ecology, D-07745, Jena, Germany. 4Present address: Department of Molecular, Cellular, and Developmental Biology, Yale University, CT 06520, New Haven, USA. †These authors share senior authorship; ‡Correspondence: [email protected] (M.K.); [email protected] (M.A.K.) bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. drosophilids, especially D. melanogaster, is well we recovered four main clusters within the genus understood 22. Fourth, pheromone receptors are Drosophila. First, the Drosophila subgenus that narrowly tuned to fly odors 23 and expected to evolve contains five main groups (repleta, virilis, melanica, at fast rates to match the dramatic diversity of cardini, and immigrans groups); second, the Zaprionus pheromones among closely related species 5,24. Lastly, subgenus that includes Zaprionus indianus; third, out of the 52 classes of olfactory sensory neurons the Dorsilopha subgenus that includes D. busckii; (OSNs) in D. melanogaster 25, only four respond to fourth, the Sophophora subgenus that includes fly odors and are localized in a specific sensillum melanogaster, obscura, willistoni, and saltans groups type 26. Hence, the restricted number of orthologues, (Fig. 1a). that are always expressed in an easily identifiable roup melanica group a lis g and accessible sensillum type, represent promising viri ca RAxML bootstrap support r D.pseudotalamancana din = 100 i g candidates to study the coevolution of Drosophila ro up D.montana

D.littoralis D.americana pheromones and their corresponding receptors. D.nannoptera D.lummei D.virilis

D.bromeliae

D.melanica D.euronotus

D.gaucha D.paramelanica D.micromelanica D.robusta D.polychaeta D.repletoides Olfactory sexual communication in D. D.hydei Z.indianus D.mettleri D.acutilabella D.stalkeri D.arawakana p D.dunni u D.mercatorum D.ornatifrons o D.repleta melanogaster is arguably one of the best studied r D.tripunctata g D.pallidipennis a D.buzzatii i 27 t D.testacea m e D.hamatofila l D.putrida m p systems in animals , and is by a large part carried i e D.macrospina g r D.mulleri r D.quinaria a D.wheeleri D.palustris n s

out through a single molecule, cis-vaccenyl acetate D.navojoa D.albomicans g r D.nasuta o 28 D.arizonae u Zaprionus 95 D.sulfurigaster p D.moj.mojavensis sophila (cVA) . This compound is produced exclusively by Dro 87 47 99 D.immigrans D.moj.wrigleyi 97 D.hypocausta D.moj.baja Root 99 D.sternopleuralis males and transferred to females during copulation, 97 D.funebris D.moj.sonorensis Colocasiomyini Dorsilopha D.busckii S.latifasciaeformis D.simulans which then reduces the attractiveness of the freshly 99 D.mauritiana S.lebanonensis a or D.sechellia 29 oph C.pararufithorax Soph D.melanogaster mated females . Moreover, cVA regulates multiple 92 C.procnemis 98 D.yakuba 96 95 D.sturtevanti 96 D.santomea behaviors: it induces sexual receptivity in virgin s D.saltans D.erecta a 98 98 l D.eugracilis t a D.suzukii 30-32 33 n D.neocordata s D.biarmipes females , elicits aggression in males , modulates g D.emarginata D.takahashii r o D.nebulosa D.fuyamai u 34 p D.willistoni D.elegans D.ficusphila oviposition behaviors , and acts as aggregation D.tsacasi w D.sucinea D.greeni il D.equinoxialis D.birchii li D.serrata s D.kikkawai 30,35 t D.asahinai o D.imaii D.lacteicornis D.rufa n D.auraria D.triauraria pheromone in presence of food . Despite the i D.paulistorumD.bifasciata g ro u p p D.azteca u D.guanche D.affinis ro D.subobscura g profound knowledge of cVA-induced behaviors in D. er st D.persimilis ga D.ananassae o

D.bipectinata n ob la melanogaster, little is known about analogous stimuli sc e u D.pseudoobscura D.malerkotliana m ra gro 0.04 that regulate social and sexual behaviors in other up drosophilids. Here, we identify the sex pheromones b b’ and their roles in 99 species within the family 1 0.5 Drosophilidae, explore the evolution of pheromone 0 signaling systems with respect to phylogenetic −0.5 Correlation coefficient relationships, and highlight how sexual isolation −1 barriers between species are reinforced by olfactory signals.

Whole-genome information-based phylogeny Fig. 1. Phylogenetic relationships and chemical varia- of 99 drosophilids tions among drosophilids (a) Phylogeny of 99 species within the family Drosophilidae inferred The genus Drosophila is arguably one of the most from 13,433,544 amino acids sites that represent 11,479 genes (See Supplementary Table 1 and Materials and Methods for details). extensively studied model systems in evolutionary bi- Using 4 species in the Colocasiomyini subgenus as outgroups (pur- ology 16,36-38. However, the phylogenetic relationships ple), 95 species are distributed in four subgenera belonging to the Drosophilini tribe (Drosophila, light green branches; Zaprionus, grey among drosophilids have suffered from low supports branch; Dorsilopha, brown branch; Sophophora, dark green branch- 39-41 . We therefore investigated the relationships of 99 es). Species names are color coded according to their relationships species within the family Drosophilidae, 95 of which in nine different species groups, with black species depicting individ- ual representatives of species groups. Scale bar for branch length span the diversity of flies across the genusDrosophila . represents the number of substitutions per site. Maximum likelihood Whole-genome sequences (WGS) for 41 of these 99 (ML) phylogenetic analyses display strong rapid bootstrap support species are available (Supplementary Table 1), thus, (100% support indicated by black circle at the nodes) for most rela- tionships among the different species. we generated WGS for the other 58 species (See Ma- (b) Heat map showing pairwise correlations between male chem- terial and Methods for accession numbers). We, then, ical profiles of the 99 species (ordered on each axis according to reconstructed the phylogeny of these 99 species using their phylogenetic relationships from Fig. 1a). Overall peak areas of 248 male chemical features across the 99 species were compared 13,433,544 amino acids sites from 11,479 genes (Fig. using Pearson correlation coefficient (R2); Color codes in the heat 1a). Maximum likelihood (ML) phylogenetic analyses map illustrate the pairwise correlations, which range from dark blue revealed strong support for the relationships among (Perfect correlation between chemical profiles) through white (no correlation) to dark red (perfect anticorrelation). The diagonal of the the different species (Fig. 1a). Using the four spe- correlation matrix is the correlations between each species and itself cies in the Colocasiomyini subgenus as outgroups, (values of 1). Note that the male correlation matrix displays frequent

2 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. dark blue cells mainly around the diagonal, i.e., high correlation coefficients are observed mostly between closely related species. (b’) Pairwise correlation analysis between female chemical profiles of the 99 species arranged according to their phylogeny from Fig. 1a. Overall peak areas of 256 female chemical features across the 99 species were compared using Pearson correlation coefficient (R2). See Supplementary Table 2 for the statistical Pagel’s lambda correlations.

Closely related drosophilids exhibit highly in the populations that produce it 21. Similarly, in the divergent chemical profiles melanogaster group, 7,11-heptacosadiene, a female- specific compound, induces male courtship in the We then asked whether the phylogenetic relationships producing species, but serves as an isolation barrier could reflect the differences in cuticular chemicals for the closely-related non-producing species 42,43. among these species. We, therefore, analyzed the In search for analogous compounds all along the chemical profiles of males and virgin females of Drosophila genus, we analyzed the chemical profiles all 99 species, with five replicates or more of each of the 99 species and compared the chromatograms sex yielding more than 580 and 520 replicates, of both sexes within each species (Fig. 2a). Males respectively. Chemical analyses recognized the and females of only 18 species exhibited sexually presence of 248 and 256 compounds (i.e., features monomorphic chemical profiles, while 81 species with distinct m/z (mass-to-charge ratios)) across exhibited sexually dimorphic cuticular chemicals male and female replicates, respectively (Extended (Fig. 2b). All the 81 dimorphic species unveiled male- Data Fig.1a and a’). Principal component analyses specific compounds (in total 42 compounds), while revealed that females of the different species groups only 15 species exhibited female-specific ones (9 exhibit closer distances than male species groups, compounds) (Fig. 2b and Extended Data Fig.2a and indicating that females exhibit more similar chemical B). Of note, all the female-specific compounds, are profiles across the species groups (Extended Data long-chain unsaturated hydrocarbons (Extended Fig.1b and b’). Similarly, in a cluster analyses of the Data Fig.2a), display high boiling temperature chemical signals most males belonging to the same (Supplementary Table 3), and hence are likely to be species group are clustered, while species groups in non-volatile 42. However, male-specific compounds females are dispersed across the chemometric tree range between 10 to 32 carbon atoms, and belong (Extended Data Fig.1c and c’). Next, we assessed to different chemical classes such as esters, ketones whether the phylogenetic and chemometric distances and alkenes, as well as an ether and an alcohol are correlated. Pairwise correlation analyses imply (Fig. 2c and Supplementary Table 3). Notably, when that indeed closely related species exhibit more analyzing, in addition, the chemical profiles of freshly- similar chemical profiles in males (Fig. 1b and mated females, we found that many of the male- b’). For example, male species of the repleta and specific compounds were transferred to females the melanogaster group display high correlation during mating (green cells in Fig. 2b), reminiscent coefficients to other male species of their own group, of the transfer of male-specific compounds in D. but negative correlation coefficients to males of melanogaster and D. mojavensis 21,44,45. On the heterospecific groups (i.e., blue cells are frequently contrary, none of the female-specific compoundswas present around the diagonal) (Fig. 1b). However, transferred to males during mating (Extended Data female species generally display high correlation Fig.2b). coefficients (>0.75) randomly to each other apart from The proportion of male-specific their phylogenetic relationships (Fig. 1b’). Indeed, chemicals with low phylogenetic signals is significantly males have significantly more chemicals with higher higher than those with high phylogenetic signals phylogenetic signals (i.e., males of related species (p = 0.025), signifying that closely related species tend to resemble each other more than males of possess dissimilar male-specific compounds (Fig. unrelated species) than females as measured by 2d and Supplementary Table 2). Indeed, mapping Pagel’s λ (p = 0.006) (Extended Data Fig.1d and the sex-specific compounds onto the phylogenetic Supplementary Table 2). Together, our data suggests tree revealed that many of them are repeatedly that a subset of the male chemicals evolve neutrally present across different species in different groups, in relation to phylogeny, while other male chemicals while few are exclusively species- or group-specific diversify rapidly and could be under positive selection. compounds (Fig. 2b and Extended Data Fig.2b). For Previously unidentified potential sex example, several -specific compounds, including pheromones undergo rapid evolution cVA that has been thought to be restricted to the melanogaster and the immigrans groups 24, are In drosophilids sex-specific compounds typically present across several species in different groups in serve as short range communication signals that both subgenera Sophophora and Drosophila, while induce or inhibit sexual behaviors 22. For example, only methyl myristoleate is specific for the willistoni in the mojavensis complex, (Z)-10-heptadecen-2-yl group (Fig. 2b). Similarly, consistent with the rapid acetate, the male-specific sex pheromone, is detected evolution of pheromone-producing enzymes in by all populations, but only induces female receptivity drosophilid females 46, 7,11-heptacosadiene and 3 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Monomorphic sp. Dimorphic sp. O a 7 2-Pentadecanone c (Z )-11-Octadecen-1-yl acetate (cVA) Not transferred Transferred 1 AcO OAc rac -2-Pentadecyl acetate 8 (Z )-9-Hexadecen-1-yl acetate * 2 AcO O 2-Tridecanone (Z )-10-Heneicosene 9 3 10 1-Pentadecene OAc rac -2-Tridecyl acetate 4 * rac -(Z )-10-Heptadecen-2-yl acetate OAc O 11 5 (Z )-10-Heptadecen-2-one * Methyl nonyl ether O O 12 6 2-Heptadecanone Myristyl acetate Male-specific compounds 13 AcO b Monomorphic / Dimorphic 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930313233343536373839404142 Palmityl acetate D.moj.sonorensis 1 14 AcO D.moj. baja 1 D.moj.wrigleyi 3 (Z )-7-Tricosene D.moj.mojavensis 3 15 9 D.arizonae 1 D.navojoa 3 (Z )-7-Pentacosene D.wheeleri 2 16 14 D.mulleri 1 D.hamatofila 3 (Z )-11-Hexadecen-1-yl acetate D.buzzatii 3 17 AcO D.mercatorum 7 OAc rac -2-Heptadecyl acetate D.repleta 4 18 D.stalkeri 5 * D. mettleri 3 D.hydei 2 19 Unidentified 1 D.gaucha 6 D.bromeliae 2 D.nannoptera 4 20 Unidentified 2 (28:1-Ac) D.pseudotalamancana 1 D.americana 1 (Z,Z )-19,22-Octacosadien-1-yl acetate D.lummei 0 21 AcO D.virilis 1 15 D. ittoralis 1 (Z )-11-Eicosen-1-yl acetate D.montana 0 AcO 22 7 D.paramelanica 4 D.melanica 4 All-trans-Geranylgeraniol D.euronotus 2 23 OH D.micromelanica 0 D.robusta 1 O 24 Ethyl myristate D.polychaeta 5 O D.repletoides 5 O Ethyl palmitate Z.indianus 3 25 D.acutilabella 3 O D.arawakana 3 O D.dunni 4 26 Ethyl oleate D. ornatifrons 4 O D.tripunctata 4 Farnesyl acetate D.pallidipennis 3 27 OAc D.testacea 3 D.putrida 1 (E)-7-Tetradecene D.macrospina 4 28 D.quinaria 4 D. palustris 1 (Z )-7-Hexadecene D.albomicans 0 29 D.nasuta 0 D.sulfurigaster 0 Stearylacetate D.immigrans 5 30 AcO D.hypocausta 3 D.sternopleuralis 1 (E )-ß-Farnesene D.funebris 2 31 D.busckii 4 D.simulans 3 (E,E )-α-Farnesene D.mauritiana 3 32 D.sechellia 3 O (Z )-11-Docosen-1-yl acetate D.melanogaster 3 33 O D.yakuba 1 6 D.santomea 1 D.erecta 3 34 OAc Geranyl acetate D.eugracilis 1 D.suzukii 0 D.biarmipes 1 35 1-Tridecene D.takahashii 0 D.fuyamai 1 O rac-2-Heptyl butanoate D.elegans 0 36 O * D.ficusphila 0 D.tsacasi 0 D.greeni 0 37 Unidentified 3(26:2-Ac ) D.birchii 0 D.serrata 2 1-Heptacosene D.kikkawai 2 38 17 D.asahinai 3 D.lacteicornis 0 O Geranyl palmitoleate 39 O D.rufa 0 7 D.triauraria 0 (Z,Z,Z )-21,24,27-Triacontatrien-1-yl acetate D.auraria 1 40 AcO D.malerkotliana 4 17 D.bipectinata 2 O Methyl myristoleate D.ananassae 1 41 D.persimilis 4 O D.pseudoobscura 3 O D.affinis 2 12-Methyltetradecyl acetate 42 O D.azteca 0 D.guanche 4 D.subobscura 2 d 38 D.imaii 0 8 35 p = 0.025 D.bifasciata 2 D.paulistorum 1 7 34 D.sucinea 3 6 24 D.equinoxialis 3 D.willistoni 3 4 23 26 41 D.nebulosa 2 D.emarginata 1 1 9 16 33 D.neocordata 1 22 10 25 D.saltans 6 14 31 D.sturtevanti 5 2 5 C.procnemis 2 3 15 C.pararufithorax 4 S.lebanonensis 1 S.latifasciaeformis 2 0.0 0.2 0.4 0.6 0.8 1.0 34 1 4 6 6 10 7 3 8 3 5 2 4 12 8 11 9 4 1 1 6 13 1 1 1 1 5 2 3 6 5 5 11 4 1 1 1 5 1 1 3 1 0.04 Pagel's lambda Fig. 2. Newly-identified potential sex pheromones display rapid evolution (a) Representative gas chromatograms of virgin male (♂), and virgin (v♀) and mated (m♀) female flies obtained by solvent-free thermal desorption-gas chromatography-mass spectrometry (TD-GC-MS) 47. Five replicates or more of each sex were analyzed, yielding more than 4 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 580, 520, and 500 replicates of males, and virgin and mated females of all 99 species, respectively. Left panel, example of a monomorphic species, whose males exhibit a chemical profile identical to that of virgin and mated females. Right panel, example of a dimorphic species that displays sexually dimorphic profiles. Colored peaks indicate male-specific compounds (green, compounds transferred to females during mating; red, non-transferred compounds). (b) Distribution of 42 male-specific compounds among different drosophilids; 81 species are dimorphic species (in black), while 18 species (in grey) are monomorphic species. Phylogeny on the left side is identical to the tree in Fig. 1a; the branches are colored according to group identities. Numbers on the right side represents the sum of male-specific compounds present per species, while numbers at the bottom of the table represent number of times each male-specific compound appeared in the different species. Cell colors refer to transferred (green) and non-transferred (red) compounds. See Extended Data Fig.2 for female-specific compounds. (c) Chemical structures and names of the male-specific compounds according to the International Union of Pure and Applied Chemistry (IUPAC). Out of 42 male-specific compounds, 39 compounds were chemically identified. Compound size ranges between 10 to 32 carbon atoms, with 23 esters, 4 ketones, 8 alkenes, 2 terpenes, 1 ether and 1 alcohol. See Supplementary Table 3 for Kovat’s Index, chemical for- mula, exact mass, mass spectrum (M/Z), and boiling temperature of these compounds. (d) Frequency histogram of Pagel’s lambda estimates, which measures the phylogenetic signal exhibited by the chemical trait 48. A value of 0 suggests the male chemical trait cannot be explained by phylogenetic relatedness, while a value close to 1 suggests strong phylogenetic sig- nal consistent with neutral evolution. See Materials and Methods for details on statistical analyses. Note that the proportion of male-specific chemicals with low phylogenetic signals is significantly higher than those with high phylogenetic signals; * P = 0.025, one-sample Wilcoxon signed rank test (n = 24). See Supplementary Table 2 and Extended Data Fig.1c for Pagel’s lambda estimates for male and female chemicals.

7,11-nonacosadiene, the female-specific compounds D. pseudotalamancana and D. robusta, whose at1- in D. melanogaster 22,43, are not restricted to a specific like sensilla could not be identified (Extended Data group (Extended Data Fig.2b). This pattern supports Fig.3a; for identification,see Materials and Methods). the presence of strong compounds in D. melanogaster We, next, recorded the responses of both trichoid 22,43, are not restricted to a specific group (Extended sensillum classes in the females of 54 species to an Data Fig.2b). This pattern supports the presence of array of chemicals (Fig. 3a), which includes 28 male- strong selection on the sex-specific compounds to specific compounds and 8 compounds that were evolve fast and deviate from expectations based on previously described as drosophilid sex pheromones neutral evolution. Additionally, our analyses revealed (Supplementary Table 4) 47,50,51. Electrophysiological that 58 of the 81 dimorphic species have a blend of recording revealed that females of 36 of 49 dimorphic multiple compounds that could reach up to seven species detect their conspecific males’ compounds compounds, as in D. mercatorum, while the other (Fig. 2b) by olfactory neurons (Fig. 3b and c). Of 23 species employ single male-specific compounds note, flies are also able to detect many male-specific (Fig. 2b). Overall, we identified 51 potential sex compounds of other species (Fig. 3b, Extended pheromones (Supplementary Table 3), which seem to Data Fig.3a and b’, and Supplementary Table 5). evolve rapidly and independently from phylogenetic To analyze the olfactory-based communication constraints across drosophilids. between the different species, we performed network analyses, which revealed a higher olfactory Drosophilids communicate intra- and inter- clustering coefficient (i.e., the number of olfactory specifically through rapidly evolving olfactory communications between the species divided by channels number of communications that could possibly exist) of interspecific communication through at1-like (for The volatility (i.e., low boiling points due to their identification, see Materials and Methods) compared shorter chain length compared to female-specific to at4-like sensilla (Fig.3d, and Extended Data Fig.3b compounds; Supplementary Table 3) of most male- and B’). However, self-loops of the intraspecific specific compounds suggests that they could be communication (i.e., the ability of a female to detect potential olfactory signals. We, therefore, screened her conspecific male) are comparable through at1-like for olfactory sensory neurons (OSNs) that detect and at4-like sensilla (Extended Data Fig.3b and b’). male-specific compounds in Drosophila species Pairwise correlation and statistical analyses revealed via single sensillum recordings (SSR). We focused that electrophysiological responses of the different our attention on 54 species – 49 dimorphic and 5 species exhibited low phylogenetic signals (Fig. 3e- monomorphic species – because they could be f), indicating that olfactory channels of the different successfully reared on artificial food under our lab species evolve rapidly apart from their phylogenetic conditions. In D. melanogaster, sex pheromone- relationships. responsive neurons are localized in antennal trichoid (at) sensilla, which are morphologically distinct from We further asked whether pheromone other sensillum types and belong to two classes (at1 receptors in drosophilids have evolved under positive and at4) that are known to be located on different selection. We, therefore, queried the genomic data antennal regions 49. The at1 sensillum houses a single for the orthologs of the known pheromone receptors neuron (Or67d) that responds to cVA 31, while at4 in drosophilid flies. Of these receptors, we found in houses 3 neurons (Or47b, Or65a/b/c, and Or88a) that our WGS data 42, 41, and 36 orthologs of Or47b, respond to methyl laurate, cVA, and methyl palmitate, Or67d, and Or88a, respectively, which displayed full- respectively 23,47. Indeed, we found the at1-like and length sequences. We next assessed the selection at4-like sensillum classes in all tested species except pressures on these genes by computing the ratio

5 (a) Fig. 3.Drosophilids communicateintra- andinter-specifically throughrapidlyevolving olfactorychannels b c a

Male-specific c.#1717 Single Sensillum

At4 At1 125 100

Left: schematics of recordings (SSR)from single sensillum the antennaltrichoid(at1andat4) sensilla.Right:Namesof the different Response (spikes/s)

D.moj.s. 75 50 25 Recording 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 10 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 10 11 11 At1 At4 9 8 7 6 5 4 3 2 1 9 8 7 6 5 4 3 2 1

35 17 16 11 D.moj. b. bioRxiv preprint not detected Compound# 35 17 16 11 D.moj.w. 11 3420111 01 22033142181 05433117733803010601 4 16 0 6 0 1 0 3 0 8 3 3 7 7 1 1 3 3 4 5 10 7 12 8 1 2 4 1 3 3 0 2 12 1 0 6 13 10 2 2 5 3 18 1 1 0 2 4 13 12 11 8 5 9 123362213829026002706521 4514231 2 0 4 3 1 3 2 4 3 3 4 7 1 2 1 2 3 3 1 3 3 14 3 2 4 1 5 14 1 11 2 5 6 0 7 2 0 0 6 2 0 9 2 8 3 1 2 2 6 3 3 2 11 1 preprint (whichwasnotcertifiedbypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplayin

31 30 D.moj.m. D.moj.sonorensis 4 Transferred Male-specificcompound Untransferred Male-specificcompound Non-specific compound 35 34 33 16 11 D.arizonae

8 D.moj.baja

31 30 D.navojoa 4 D.mulleri D.moj.wrigleyi 9 7 5 D.moj.mojavensis Compound #

D.hamatofila 7 6 5 4 3 2 1 9 6 5 2-P ( 2-H ( ( ( rac-2-Tri

18 13 D.buzzatii Z Z Z Z 9 7 6 5 D.arizonae detected byat1 )-10-H )-10-H acetate adecen-1-yl )-9-Hex )-11-Oct 31 30

D.mercatorum ent ept 9 4

D.navojoa doi: D.repleta adecanone 9 7 6 D.mulleri adecanone ept ene eneicos decylacetate

19 D. mettleri 9 -1-yl acetate acetate adecen-1-yl https://doi.org/10.1101/2020.09.21.305854

D.hydei D.hamatofila adecen-2- 16 35 17 11 D.gaucha

6 5 D.buzzatii

22 D.bromeliae 7 6 D.mercatorum D.nannoptera 6 D.pseudotala. D.repleta one detected byat4 3

D.americana D. mettleri perpetuity. Itismadeavailableundera 3

21 10 D.virilis 6 D.hydei D.melanica (cVA) 24 22 20 D.polychaeta D.gaucha 18 1

D.bromeliae 12 11 10 9 8 14 13

13 12 D.repletoides 1 D.dunni 35 11 Myristyl acetate Myristyl (R- 1-P 2-Tri rac-2-P D.nannoptera ( acetate Palmityl 5 1 Z

D.pallidipennis ene )-7-Tricos 18 Z 1 D.pseudotalamancana ent )-10-H D.testacea decanone 1 1 detected byboth adecene

D.putrida D.americana ent 18 15 13 cyl acetate adecyl

20 15 13 D.macrospina 12 D.virilis ept 1 D.immigrans 18 15 -2-yl acetate adecen-2-yl

1 D.melanica D.hypocausta 15 1 D.funebris D.robusta 18 34 33 31 30 13 8 7 6 4 D.busckii D.polychaeta 14 1 D.simulans 14

1 D.repletoides D.mauritiana 14 1 D.sechellia D.dunni 14 ; 1

D.melanogaster D.pallidipennis this versionpostedSeptember22,2020. 1 1 18 17 16 15 21 20 19

6 D.yakuba

14 D.testacea ( rac-2-Heptadecyl acetate ( ( ( Farnesyl acetate All-trans-Geranylgeraniol Z Z D.erecta Z E 1 , )-11-Hexadecen-1-yl acetate D.putrida )-11-Eicosen-1-yl acetate )-7-Tetradecene Z

D.eugracilis CC-BY 4.0Internationallicense 18 D.malerkotliana D.macrospina )-19,22-Octacosadien-1-yl acetate 1 D.ananassae 18 D.nasuta 1 D.pseudoobscura

2 1 D.immigrans

14 D.affinis 1 D.hypocausta

222 22 D.subobscura D.willistoni D.funebris D.nebulosa D.busckii

d D.simulans

0.00 Clustering0.25 0.50 coefficient0.75 1.00 D.mauritiana sp. D.sechellia

At1 D.melanogaster 24 23 22 28 27 26 25

*** D.yakuba rac-2-Heptyl butanoate Geranyl acetate ( Methyl palmitate Methyl myristate Methyl laurate ( E Z At4 )-ß-Farnesene

D.erecta )-9-Tricosene

D.eugracilis . D.suzukii The copyrightholderforthis e D.takahashii

0.0 Pagel's0.2 0.4 lambda0.6 0.8 1.0 D.elegans D.ficusphila

At4 At1 D.malerkotliana 36 35 34 33 32 31 30 29 D.ananassae ( S S R rac-3-Pentadecyl acetate S R 2-Hexyl acetate Z -(Z)-10-Heptadecen-2-yl acetate -2-Pentadecyl acetate -2-Tridecyl acetate -2-Pentadecyl acetate -2-Tridecyl acetate D.pseudoobscura )-9-Octadecen-1-yl acetate f D.affinis Correlation coefficient D.subobscura −1 0 1 At4 At1 D.willistoni D.nebulosa D.neocordata 0.04 D.saltans D.sturtevanti 5 3 10 8 7 11 12 5 8 14 22 3 9 7 7 5 6 4 4 3 10 10 4 1 4 8 5 8 6 8 4 20 5 2 7 11 6 0 5 2 3 1 1 4 4 6 3 10 8 3 4 5 4 7 6 2 4 11 3 2 7 3 6 4 2 5 3 9 1 9 5 36 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. chemicals used to screen the trichoid sensilla. Note that all chemicals are male-specific compounds identified in this study (Fig. 2b), except compounds# 25, 26, 27, 28, 29, 32, and 36, which were described as flies’ pheromones in (47, 50, 51). (b) Color-coded electrophysiological responses towards heterospecific compounds (grey bubbles) and conspecific compounds (colored bubbles) in at1 (top) and at4 (bottom) sensilla of 54 species. Compound names are depicted in Fig. 3a. Red and green bubbles represent species-specific male untransferred and transferred compounds, respectively. Bubble size corresponds to the response values ranging from 25 to 125 spikes per second. Responses less than or equal to 10 spikes per second were excluded from the bubble chart (see Extended Data Fig.3a and a’). Species names are arranged on the top according to their phylogenetic relationship; the tree branches are colored according to the group identities. Numbers on the right side represent the sum of species that can detect each of the male-specific compounds, while the number below the table represent the sum of chemicals that can be detected by each species. (c) A summary of female’s abilities to detect their own male-specific compounds through olfaction in 47 dimorphic species (two species, D. robusta and D. neocordata, whose compounds were not included among the 36 compounds, were excluded). Black numbers, undetected male-specific compounds; orange, detected by at1 neuron(s); blue, detected by at4 neurons; green, detected by both. See Extended Data Fig.3a for more details about the number of neurons in at1 sensillum. Note that, out of 47, females of 36 species detect their conspecific male cues through at1 and/or at4. (d) Top: A schematic example of how to calculate the olfactory clustering coefficient of a given species (yellow circle) to communicate with heterospecific species (black circles) through at1 (orange lines) and at4 (blue lines). The olfactory clustering coefficient is the number of other species that are detected by a given species through at1 or at4 (colored lines) divided by the total number of detected and undetected species (colored + grey dashed lines). The clustering coefficient of a species is a number between 0 (i.e., no species detected) and 1 (i.e., all species detected). Below: scatter plot indicates olfactory clustering coefficients of the 54 species and their mean through at1 (orange) and at4 (blue); Mann Whitney U test, *** P < 0.001 (n = 54 species). Note that species exhibit more olfactory intra- and inter-specific communication through at1 than at4. See Extended Data Fig.3b and b’ and Materials and Methods for more details on communication network analyses. (e) Frequency histogram of Pagel’s lambda estimates, which explain the correlation between the olfactory responses of at1 (orange) and at4 (blue) among the different species and their phylogenetic relationships. Note that responses of both at1 and at4 display low phylogenetic signals (i.e., do not correlate with the phylogeny). Additionally, their phylogenetic signals are comparable to each other; Mann Whitney U test, ns P = 0.27 between at1 and at4 responses. (f) Heat map showing the pairwise correlations between the electrophysiological responses of at1 (top) and at4 (bottom) sensilla in 54 spe- cies, which are ordered on each axis according to their phylogenetic relationships in Fig. 3b. Overall responses to 36 compounds across the 54 species were compared using Pearson correlation coefficient (R2); Color codes in the heat map illustrate the pairwise correlations, which range from dark blue (Perfect correlation between species’ responses) through white (no correlation) to dark red (perfect anticorrelation). The diagonal of the correlation matrix depicts the correlations between each species and itself (values of 1). The phylogenetic trees on each axis are similar to Fig. 3b; the branches are colored according to the group identities. Note that the male correlation matrix displays frequent dark blue cells around the diagonal and overall the matrix, i.e., high correlation coefficients only between the closely related species. Unlike the at1 correlation matrix, species’ responses display lower correlation coefficients, indicating that olfactory channels in at4 evolve rapidly and independently from the phylogeny. of nonsynonymous (dN) to synonymous (dS) Many species displayed different species-specific substitutions per gene (see Materials and Methods). behaviors (Movies1 to 427, available on Edmond; in Statistical analyses revealed an evidence of positive total 1467 replicates, 16-48 replicates per species). selection on all tested pheromone receptors with For example, males of D. elegans and D. suzukii the highest pressures on the Or47b and the Or67d dance and spread their wings in front of females 52, loci (Or47b locus, p-value < 0.0001; Or67d locus, D. mojavensis and D. virilis males release fluidic p-value < 0.0001; Or88a locus, p-value = 0.014). droplets while courting the females 21, D. subobscura males extend their proboscis to gift females with Lastly, using two different model-tuning regurgitated drop of their gut contents 18, and D. criteria, we performed a phylogenetically corrected nannoptera couples tend to re-mate as many as two to correlation between the evolution of male chemical three times within the recording time frame of 60 min phenotypes and the associated olfactory responses 53. We further quantified copulation success, latency, of their females. Despite the high intraspecific match and duration (Extended Data Fig.4a), which varied – females of 36 out of 49 species detect their males largely among different species. Unlike the prolonged – (Fig. 3c), the evolution of females’ responses does copulation time in the species of the melanogaster not fit the evolution of their male-specific compounds group, copulation lasts for less than two minutes (Supplementary Table 6). This implies presence of in members of the repleta group (Extended Data low interspecific correlation between detection and Fig.4a). Together, courtship recordings (available production. Indeed, for example, females of 19 out on Edmond; Link) reveal numerous quantitative and of 20 species, whose males produce cVA, detect qualitative differences in sexual behaviors among the this compound (i.e., cVA functions as a conspecific Drosophila species. signal), while females of 21 out of 34 species are still able to detect cVA (Fig. 3b), although their males do We next focused on Drosophila species not produce it (i.e., cVA functions as a heterospecific that detect their male-specific compounds via signal). olfaction – 36 out of 49 species (Fig. 3c and Fig. 4a) – and asked whether these compounds induce Male-specific compounds regulate intra- and female receptivity. In a competition-mating assay, inter-specific sexual behaviors virgin females of each species were allowed to To examine the intra- and inter-specific behaviors choose between two conspecific males perfumed governed by the male-specific olfactory signals and to with a male-specific compound (Fig. 4b) or solvent gain a better understanding of the courtship rituals of [consistency of perfuming and correspondence to these 54 species, we recorded the sexual behaviors biologically relevant amounts were confirmed by of conspecific couples in a single-pair courtship arena. chemical analyses; see Materials and Methods]. 7 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

a 1 (Z)-11-Octadecen-1-yl ac. (cVA) b c d 3 (Z)-10-Heneicosene 4 rac-2-Tridecyl ac. 5 (Z)-10-Heptadecen-2-one 6 2-Heptadecanone Hexane 7 2-Pentadecanone 8 rac-2-Pentadecyl ac. / Headless 9 2-Tridecanone 101-Pentadecene cVA (Hetero-specific C.) 11 rac-(Z)-10-Heptadecen-2-yl ac. 15(Z)-11-Hexadecen-1-yl ac. 16rac-2-Heptadecyl ac. Compound# 18(Z)-11-Eicosen-1-yl ac. 19All-trans-Geranylgeraniol 20Farnesyl ac. Hexane Untransferred c. Transferred c. 22(E)-ß-Farnesene 24rac-2-Heptyl butanoate : Aphrodisiac to : Anti-aphrodisiac to : Anti-aphrodisiac to cVA C# at1 at4 n n at1 at4 n 11 108 112 57 D.moj.wrigleyi 16 96 76 D.moj.wrigleyi 11 56 84 D.moj.mojavensis D.moj.mojavensis 62 16 64 104

D.arizonae 4 56 36 D.arizonae 56

8 68 88 D.navojoa D.navojoa 65 16 68 72

D.mulleri 4 84 D.mulleri 68

D.hamatofila 5 40 D.mercatorum 68

D.mercatorum 5 52 D.bromeliae 60 9 40 112 D.nannoptera 64 D.hydei 19 76

D.bromeliae 5 64 D.pseudotalamancana 57

D.nannoptera 7 68 44 D.virilis 61

D.pseudotalamancana 6 44 D.busckii 52

D.virilis 3 104 100 D.malerkotliana 76

D.melanica 10 72 80 D.sturtevanti 74

76 D.polychaeta 20 24 88

D.repletoides 1 104 48 18 92 52

D.dunni 1 20 20

32 64 D.pallidipennis 1 5 56

D.testacea 1 24 84

D.putrida 1 56 40

D.macrospina 1 24 40

D.immigrans 1 44 56

D.hypocausta 1 41 53

44 48 D.funebris 1 15 32 52 64 D.busckii 4 8 56 80

D.simulans 1 92 63

D.mauritiana 1 72 92

D.sechellia 1 52 48

D.melanogaster 1 80 64

D.yakuba 1 80 32

D.erecta 1 72 16

D.malerkotliana 18 48 108

56 92 D.pseudoobscura 1 18 80 D.affinis 1 56 56

D.subobscura 1 44 32

D.saltans 22 68

0.04 D.sturtevanti 22 60 Fig. 4. Male-specific compounds regulate intra- and inter-specific sexual behaviors (a) Left: Names of the compounds (top) that are exclusively produced by males of 54 species (left below) and detected by conspecific females through at1 or/and at4 (right below). These compounds were used for the behavioral experiments in Fig. 4b, C, and D. (b) Top: Schematic of a mating arena where females of each species had the choice to mate with two conspecific males perfumed with their olfactory-detected male-specific compound (indicated, in Fig. 4a, on the left side of the horizontal dashed stoke) or solvent (dichloromethane, DCM). For consistency of perfuming and correspondence to biologically relevant amounts see Materials and Methods. Below: bar plots rep- resent the percentages of copulation success of the rival males. Results from females that were only courted by one male were excluded. In 8 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. this and other panels, filled bars indicate significant difference between the tested groups; nsP > 0.05; ** P < 0.01; *** P < 0.001, chi-square test. Number of replicates are stated on the left side on the bar plots. See Extended Data Fig.4a for details regarding the differences and sim- ilarities of sexual behaviors among the 54 species. Note that in 11 instances, females displayed a preference to copulate with the male-specif- ic compound-perfumed males over the control ones, while 6 compounds resulted in avoidance, and 29 turned out to be neutral. See Supple- mentary Table 7 for raw data and statistical analyses. See Extended Data Fig.4b for the effect of perfuming on the males’ courtship behavior. (c) Top: Competition courtship arenas where a male of each species had the choice to court two decapitated conspecific females perfumed with the male-transferred compound. Note that we tested only transferred compounds (green). Below: bar plots represent the percentage of the first copulation attempts towards perfumed and control females. Results from males that only courted one female were excluded; see Materials and Methods. Note that 15 compounds inhibited courtship, 1 compound increased courtship and 16 compounds turned out to be neutral. (d) Top: Schematic of a mating arena where a female of each species had the choice to mate with two conspecific males perfumed with olfactory-detected heterospecificc VA or solvent (DCM). Note that we only tested the species that do not produce but still detect cVA. Below: bar plots represent the percentages of copulation success of the rival males. Results from females that were only courted by one male were excluded.

In 11 instances, females displayed a preference Many of these olfactory-detected male-specific to copulate with males perfumed with the male- compounds are transferred to females during specific compound over the control ones (Fig. 4b and copulation (Fig. 2b; 28 out of the 36 species in Fig. Supplementary Table 7). However, females of six 4b). We, therefore, asked whether transfer of these species avoided copulating with the males perfumed compounds contributes to a general post-copulation with the male-specific compound (Fig. 4b). Notably, mate-guarding strategy [as described for transferred perfuming an additional amount of cVA on males of pheromones in D. melanogaster and D. mojavensis the melanogaster clade did not increase the males’ 21,44,55]. To distinguish male sexual behaviors from copulation success (Fig. 4b), indicating that the built- female acceptance, males were offered the choice to in amount of cVA in the control males is already court two headless females perfumed with a male- sufficient for females’ acceptance. To assure that high specific compound or solvent. We scored the first copulation success of perfumed males was not due to attempt to copulate with one of the rival females as a an increased intensity of male courtship 54, we recoded choice. However, we ensured that males do choose their courtship activities. Courtship indices did not between perfumed females and not simply attempt differ between perfumed and control males (Extended copulation with the first female they court. Males in Data Fig.4b), indicating that these compounds almost half of the tested species that exhibit male- influence exclusively the females’ sexual decisions. transferred compounds displayed a preference to copulate with the solvent-perfumed females, including We next focused on Drosophila species that males of many species of the melanogaster group detect their male-specific compounds via olfaction – (Fig. 4c and Supplementary Table 7). On the contrary, 36 out of 49 species (Fig. 3c and Fig. 4a) – and asked males of D. hydei exhibited copulation preference for whether these compounds induce female receptivity. the perfumed females over the control ones, indicating In a competition-mating assay, virgin females of that in different species male-transferred compounds each species were allowed to choose between two can result in reverse effects. Together, out of the 36 conspecific males perfumed with a male-specific Drosophila species that produce and detect male- compound (Fig. 4b) or solvent [consistency of specific compounds, the compounds of 24 species perfuming and correspondence to biologically relevant regulated sexual behaviors in our experiments. amounts were confirmed by chemical analyses; see Materials and Methods]. In 11 instances, females Lastly, we examined why females displayed a preference to copulate with males are still able to detect the chemical signals perfumed with the male-specific compound over the of heterospecific males and whether these control ones (Fig. 4b and Supplementary Table 7). heterospecific signals could act as reproduction However, females of six species avoided copulating isolation barriers. We focused our analysis on cVA with the males perfumed with the male-specific due to its presence in many cosmopolitan species compound (Fig. 4b). Notably, perfuming an additional (e.g., D. melanogaster, D. simulans, D. funebris, amount of cVA on males of the melanogaster clade and D. immigrans) that have a high chance to meet did not increase the males’ copulation success other non-cVA-producing species (Fig. 4d). Females (Fig. 4b), indicating that the built-in amount of cVA of Drosophila species whose females detect, but in the control males is already sufficient for females’ their males don’t produce cVA, had the choice to acceptance. To assure that high copulation success mate with two conspecific males perfumed with of perfumed males was not due to an increased cVA or solvent. All these females detected cVA intensity of male courtship 54, we recoded their with OSNs that did not detect the compounds of courtship activities. Courtship indices did not differ their conspecific males (Fig. 3b). Notably, females’ between perfumed and control males (Extended preference for their conspecific males in 8 out of 13 Data Fig.4b), indicating that these compounds species was significantly reduced byc VA (Fig. 4d and influence exclusively the females’ sexual decisions. Supplementary Table 7). Overall, many male-specific

9 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. compounds seem to regulate intra-sexual behaviors changes of sexual signaling are likely to occur all along the Drosophila phylogeny and promote between closely-related sympatric species 69,70 to inter-sexual isolation for heterospecific species. overcome the homogenizing effects of gene flow. The high proportion – ~82% – of species Discussion that exhibit sexually dimorphic chemical profiles (Fig. 2b) indicates the significance of chemical Sexual selection imposed by coevolution of female communication in the genus Drosophila. This sexual preferences to particular male traits leads to rapid and dimorphism seems to positively correlate with flies’ dramatic evolutionary divergence 56,57 and potentially ability to mate under low light conditions 15,71, while contributes to speciation processes 58. Using whole- many of the chemically monomorphic species cannot genome sequences of 99 drosophilid species, we mate in the dark 15. The latter often display sexually- investigated how phylogenetic constraints impact dimorphic color patterns, implying that they rely on the evolution of cuticular hydrocarbons and potential visual cues during sexual communication 71,72. Of note, sex pheromones per se. By linking the chemical many of the sex-specific compounds exist across many variations and phylogenetic relationships on the one species in different groups, indicating that the gaining hand with the physiological responses and behavioral of a sexual chemical trait has occurred independently functions on the other, we provide large-scale multiple times during the evolution of drosophilid flies evidence for the rapid coevolution of sex pheromone (Fig. 2b and Extended Data Fig.2b). For example, production and detection among drosophilid flies. The cVA, (Z)-11-eicosen-1-yl acetate, and palmityl acetate characterizations of sex pheromones, their cognate are present in 34, 13, and 12 species that belong to olfactory channels and behavioral significances different groups, respectively. Instead, the production provide several insights into the evolution of chemical of 2-heptadecanone and (Z)-7-pentacosene are communication systems and its role in speciation. present in different groups but restricted to higher In general, cuticular chemistry varies taxonomic levels (i.e., subgenre Drosophila and between closely related species in relation to their Sophophora, respectively). Moreover, 13 compounds genetic relationships 59,60, geographical locations 61, have appeared on only one occasion across the 99 and environmental factors 62. Environmental factors species. The observed saltational changes are not are thought to have stronger impact on the evolution necessarily unexpected as minor mutations suffice of cuticular chemicals than genetic relatedness 59,63. to induce large-scale changes in the biosynthetic 66,73,74 Our findings reveal that males have significantly pathways of sex pheromones . Likewise, more chemicals with higher phylogenetic signals than gene families involved in biosynthesis of cuticular females, i.e. there is a better correlation between chemicals have been shown to evolve rapidly and 75 genetic and chemical distance in males (Extended independently among closely related drosophilids . Data Fig.1c and d). Notably, the picture changes One key observation of our study when we focus on sex-specific compounds, which is the diversity and abundance of male-specific evolve at rapid rates (i.e., display low phylogenetic compounds compared to female ones – 42 compared signals) resulting in dramatic differences between to 9, respectively – across the dimorphic species closely related species (Fig. 2b and d, and Extended (Fig. 2b and Extended Data Fig.2b). Surprisingly, Data Fig.2b) and indicating the divergence of all the 81 dimorphic species exhibit male-specific sexual communication among them. Consistent compounds, while only 15 species have female with our results, nonsexual chemical hydrocarbons specific compounds (Fig. 2b and Extended Data in ants 59, aphids 64, ladybird beetles 65, moths 66, Fig.2b). This could be attributed to the fact that and drosophilids 24 exhibit gradual evolution, while drosophilid females are regarded as “the choosy aggregation pheromones in beetles – used as sex”, which rely on volatile male sex pheromones sexual signals – display saltational (i.e., sudden to find a high quality conspecific male 31,35,50,76 and and large) shifts 67{Symonds, 2016 #13123}. By to avoid costly interspecific mating 21,31,77. Moreover, contrast, our observed saltational shifts in male- males are found to court heterospecific females in specific compounds contradicts previous studies on equal vigor as conspecific females 21,78,79, even after the gradual mode of evolution of some of the sex learning the conspecific females’ chemical profiles80 . pheromones in Bactrocera 68 and some aggregation Similarly, males exhibit higher preference for females pheromones in Drosophila 24. We identified some of that exhibit no cuticular hydrocarbons (i.e., females these previously-identified aggregation pheromones lacking oenocytes (“ oe−” females)) over wildtype as potential sex pheromones (Figs. 2b and 4b). females, while females are less receptive to oe− The discrepancy of the mode of evolution of these males 81. Furthermore, male cuticular hydrocarbons sex pheromones could be explained due to binary are modulated more easily by lab-induced natural and encoding (i.e., presence or absence) of these traits sexual selection than female cuticular hydrocarbons among a limited number of species 24. The saltational 82. All these reasons, aside from the females’

10 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. strong preferences for male sexual traits, seem to correlates of intra- and inter-sexual communication. have resulted in stronger selection pressures on A future challenge will be to investigate the genetic the cuticular hydrocarbons of drosophilid males. basis of the rapid evolutionary rate of sex pheromone production and detection and how these chemicals, To match the diverse male chemical together with the other sensory signals, collaborate traits, females are expected to coevolve cognate to result in the birth of new species. sensory detection systems to permit mate recognition Materials and Methods 1. Drosophilid chemoreceptor genes evolve rapidly 83 and single point mutations can result in species- Drosophila lines and chemicals specific variance of receptor tuning84 . Such specificity has shown to be not random and principally occurred Fly stocks. Wild-type flies used in this study were to match chemical divergence associated to host obtained from the National Drosophila Species Stock 83-87 88 selection or mate recognition . Similarly, we Centre (NDSSC; http://blogs.cornell.edu/drosophila/) found that pheromone-responsive olfactory channels and Kyoto stock center (Kyoto DGGR; https://kyotofly. evolve high selectivity that permits an extreme fit kit.jp/cgi-bin/stocks/index.cgi). Stock numbers and to the evolution of sex pheromones in conspecific breeding diets are listed in Supplementary Table partners (Fig. 3b). Females of more than 75% of 1. All flies were reared at 25°C, 12 h Light:12 h 85 the (Fig. 3e and ). Moreover, we found that many Dark and 50% relative humidity. For more details species detect other heterospecific male compounds, on the food recipes see Drosophila Species Stock highlighting the broad potential for interspecific Centre (http://blogs.cornell.edu/drosophila/recipes/). olfactory communication among the different drosophilids (Fig. 3b). Behavioral experiments Chemicals. Male- and female-specific compounds revealed that heterospecific signals reduce the are listed in Supplementary Table 3, while dimorphic species are able to detect their diverse compounds used for SSR and behavior, their sources conspecific male-specific compounds through the and CAS numbers are listed in Supplementary same olfactory channels (Fig. 3b), suggesting that Table 4. All odors were diluted in dichloromethane their cognate olfactory receptors are under positive (DCM) for SSR and behavioral experiments. selection that has acted strongly to modify their functional capabilities. Similar to the evolution of Whole-genome sequencing and phylogenetics male-specific compounds, the functional divergence of these olfactory channels among the different Sequencing library preparation drosophilids is not correlated with their phylogeny Genomic DNA was extracted from a single fly per likelihood of hybridization through different olfactory each species (for more details see Supplementary channels from those specialized to detect conspecific Table 1) using qiagen DNeasy blood and tissue kit pheromones (Figs. 3b and d). For example, the (cat. nos. 69504). Extracted DNA (~20ng/μl) was subspecies of D. mojavensis, as well as D. arizonae quantified with Qubit broad range dsDNA kit, and and D. navojoa detect their own pheromones diluted to a concentration of 1ng/µL. Tagmentation through their at4-like sensilla, while they detect the was performed with in-house Tn5 transposase heterospecific cVA through the at1-like sensilla (Fig. prepared with a previously described method 4d). Interestingly, contrary to that, at1-sensilla are used (Picelli et al., 2014; Genome Research). Tagmented in D. melanogaster to detect conspecific pheromones fragments were purified with 1 volume of SPRI beads 31 . These results reveal that species retain – at the (1mL SeraMag GE Healthcare, 65152105050250 peripheral level – the ability to detect the chemicals beads in 100mL of PEG8000 20%, NaCl 2.5M, Tris- no longer produced by conspecifics, but reverse – HCl pH=8.0 10mM, EDTA 1mM, Tween20 0.05%), at the central level – the hedonic value of these and subjected to 20 cycles of Kapa HiFi PCR 89 signals . In line with our findings, previous studies enrichment with barcoded primers using the following have shown that heterospecific sex pheromones cycling conditions: 72°C 3min, 98°C 1min, 20 cycles could reinforce the sexual isolation among sympatric of 98°C 45s, 65°C 30s, 72°C 30s. An equal volume species or recently-diverged populations through of PCR products was pooled and purified with SPRI 21,43 conserved peripheral olfactory pathways . beads with a two-sided size selection protocol, Unlike cVA-induced behaviors in D. using 0.55X (of the PCR pool volume) SPRI beads melanogaster, which are encoded mainly through for the first selection and 0.2X SPRI beads for the a single olfactory channel 31, sexual behaviors of second. Library pool was quantified with Qubit broad many other drosophilids seem to be mediated by range dsDNA kit and sized with TapeStation D1000. different compounds through multiple channels Sequencing was performed on two HiSeq X lanes. (Figs. 3b and 4b). The lack of genetic tools for most of the drosophilid species currently precludes Gene annotations and determination of orthologs further investigations of the genetic and neuronal Nine draft assemblies deposited on NCBI genbank

11 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. (D. albomicans, D. americana, D. montana, D. nasuta, branch-length stealing algorithm using the parameters D. pseudoobscura, D. robusta, D. subobscura, S. “--branches s --threshold 0.5”. Due to computational lebanonensis, P. variegata) were not annotated. We constraints, only the top 500 most informative genes lifted over annotation information from Drosophila were used to re-optimize branch lengths. melanogaster for these genomes by performing blast, followed by exonerate and genewise alignments Chemical analyses as previously described. We classified annotated genes by clustering protein-coding sequences from Thermal Desorption-Gas Chromatography- 31 species using UPhO as previously described. Mass Spectrometry (TD-GC-MS). Individual male Together, 11575 orthologs were identified from the and female flies in different mating status (virgin or annotated genomes. Together with already annotated freshly-mated) were prepared for chemical profile genomes (n=22), they serve as reference genomes to collection as described previously (47), with some which short reads from other species were mapped. modifications. Briefly, the GC-MS device (Agilent GC 7890A fitted with an MS 5975C inert XL MSD unit; Read processing and generation of pseudogenome www.agilent.com) was equipped with an HP5-MS UI assemblies. Raw reads were demultiplexed with column (19091S-433UI; Agilent Technologies). After dual barcodes by the sequencing facility, and desorption at 250°C for 3 min, the volatiles were trimmed to remove any adapter sequences using trapped at -50°C using liquid nitrogen for cooling. In Trimmomatic 0.32 using the following parameters: order to transfer the components to the GC column, ILLUMINACLIP:illumina-adaptors.fa:3:7:7:1:true the vaporizer injector was heated gradually to 270°C LEADING:25 TRAILING:25 SLIDINGWINDOW:4:20 (12°C/s) and held for 5 min. The temperature of the GC MINLEN:50. We next determined the optimal oven was held at 50°C for 3 min, gradually increased reference genome to use by mapping the first (15°C/min) to 250°C and held for 3 min, and then to 10,000 paired-reads with BWA-MEM to each of the 280°C (20°C/min) and held for 30 min. For MS, the 31 reference genomes, followed by computing the transfer line, source and quad were held at 260°C, proportion of properly mapped read pairs Pproper and 230°C and 150°C, respectively. Eluted compounds the averaged mapping quality MAPQ. We designed for this and the following analyses were ionized in an ad hoc index to maximize data usage, reference electron ionization (EI) source using electron beam quality index = (completeness of reference genome operating at 70 eV energy and their mass spectra annotation) * Pproper * (1-10^MAPQ). The reference were recorded in positive ion mode in the range from with the highest reference quality index was chosen m/z 33 to 500. The structures of the newly identified for each short-read dataset. Pseudogenomes were compounds were confirmed by comparing their mass produced as previously described, by mapping reads spectrum with synthesized or commercially available to the best reference, realigning around gaps and standards (for more details see Supplementary substituting bases of the reference genome and Table 4. Age of males and females is 10 days. masking regions with no mapped reads (MAPQ<20). Body extract analysis by GC-MS. Fly body extracts Alignment and phylogenetics. Orthologous were obtained by washing single flies of the respective protein-coding sequences were extracted from sex and mating status in 10 μl of hexane for 30 min. For reference genomes and pseudogenomes by using GC stimulation, 1 μl of the odor sample was injected the GFF annotations of the corresponding reference in a DB5 column (Agilent Technologies; www.agilent. species. TranslatorX was used to align the coding- com), fitted in an Agilent 6890 gas chromatograph, sequencings by codon, and cleaned with GBlocks and operated as described previously (90). The inlet (MinSeqConsv=0.5, MinSeqFlank = 0.55). Aligned temperature was set to 250°C. The temperature of protein-coding sequences were concatenated the GC oven was held at 50°C for 2 min, increased for each species, resulting in the final alignment gradually (15°C/min) to 250°C, which was held for 3 matrix with 11,479 genes, 13,433,544 sites in 99 min, and then to 280°C (20°C/min) and held for 30 species (5 samples were excluded based on a min. The MS transfer-line, source and quad were preliminary tree, due to their clear contradiction with held at 280°C, 230°C and 150°C, respectively. well-established taxonomy, suggesting potential Chiral chromatography. To check the presence problems in mislabeling or strain contamination). of different stereoisomers of some compounds, Data completeness ranges from 4.46%-97.27% hexane body extracts of male flies were injected (mean=58.59%). Partitioning the full alignment into 3 into a CycloSil B column (112-6632, Agilent codon positions, we inferred a maximum likelihood Technologies; www.agilent.com) fitted in Agilent tree by using RAxML 8.2.4 with 100 rapid bootstrap 6890 gas chromatograph and operated as follows: supports. Because branch length may not be The temperature of the GC oven was held at 40°C accurate with extensive missing data, we then further for 2 min and then increased gradually (10°C/ optimized the branch lengths with ForeSeqs using a min) to 170°C, then to 200°C (1°C/min), and finally 12 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. to 230°C (15°C/min) which was held for 3 min. excluded. All courtship and copulation data were Perfuming flies with hexane or male-specific acquired by a researcher blind to the treatment. compounds. Male and female flies were perfumed with the compounds singly diluted in DCM or DCM Electrophysiological experiments alone as previously described (21). Briefly, 10 μL of Single sensillum recording (SSR). Adult flies were a 50 ng/μL stock solution was pipetted into a 1.5-mL immobilized in pipette tips, and the third antennal glass vial. After evaporating the DCM under nitrogen segment was placed in a stable position onto a glass gas flow, ten flies were transferred to the vial and coverslip (91). Trichoid sensilla were identified based subjected to three medium vortex pulses lasting for on their sensillum morphology under a microscope 30 s, with a 30-s pause between each pulse. Flies (BX51WI; Olympus) at 100´ magnification. The two were transferred to fresh food to recover for 2 h and different classes of trichoid sensilla were identified on then introduced to the courtship arenas or subjected the basis of their anatomical location (at1 sensilla in to GC-MS analysis to confirm the increased amount the central region, while at4 sensilla in the distolateral of the perfumed acetate. Each fly was coated with ~2- region of the antenna) and spontaneous activities 10 ng of the compound of interest. (at1 sensilla house less neurons than at4 sensilla), which are known from D. melanogaster (49). The Chemical identification and synthesis. (provided extracellular signals originating from the OSNs were as a separate word file). measured by inserting a tungsten wire electrode in the base of a sensillum and a reference electrode into Behavioral experiments the eye. Signals were amplified (Syntech Universal AC/DC Probe; Syntech), sampled (10,667.0 Single and competitive mating assays. Males and samples/s), and filtered (300 – 3,000 Hz with 50/60 Hz females were collected after eclosion and raised suppression) via USB-IDAC connection to a computer individually and in groups, respectively. For each (Syntech). Action potentials were extracted using experiment, courtship assays were performed in a AutoSpike software, version 3.7 (Syntech). Synthetic chamber (1 cm diameter × 0.5 cm depth) covered compounds were diluted in dichloromethane, DCM, with a plastic slide. Courtship behaviors were (Sigma-Aldrich, Steinheim, Germany). Prior to each recorded for 60 min using a GoPro Camera 4 or experiment, 10 μl of diluted odor was freshly loaded Logitech C615 as stated in the figure legends. All onto a small piece of filter paper (1 cm2, Whatman, single mating experiments were performed under Dassel, Germany), and placed inside a glass Pasteur normal white light at 25°C and 70% humidity. Each pipette. The odorant was delivered by placing the tip of video was analyzed manually for copulation success, the pipette few millimeters away from the antennae, to which was measured by the percentage of males ensure the delivery of the low volatile chemicals (92). that copulated successfully, copulation latency, which Neuron activities were recorded for 10 s, starting 2 s was measured as the time taken by each male until before a stimulation period of 0.5 s. Responses from the onset of copulation, and copulation duration. In individual neurons were calculated as the increase (or all competition experiments, copulation success was decrease) in the action potential frequency (spikes/s) manually monitored for 1 h. Decapitated females relative to the pre-stimulus frequency. Traces were were used in the courtship assays to disentangle processed by sorting spike amplitudes in AutoSpike, male sexual behaviors from female acceptance. analysis in Excel and illustration in Adobe Illustrator In the competition mating assays, rival CS (Adobe systems, San Jose, CA). flies were marked by UV-fluorescent powder of different colors (red: UVXPBR; yellow: UVXPBB; Statistical analyses green: UVXPBY; purchased from Maxmax.com; https://maxmax.com/) 24 h before the experiments. Estimating phylogenetic signal with Pagel’s λ. Competition assays were manually observed for 1 h Raw peak signals were first standardized by dividing and copulation success was scored identifying the the area under each peak by the sum of areas under successful rival under UV light. Decapitated females all peaks. For each sex, the corresponding peaks were were used to observe the first copulation attempt aligned, and the standardized signals across samples of males in presence of the different compounds were logarithm-transformed to approximate normality, and DCM perfumed conspecific females. Data from followed by standardization with a z-transformation. competition experiments represents either females The phylogenetic signals contained in each chemical courted by both rival males or males courted with component were estimated by combining the both rival females to ensure that females or males transformed peak intensity with the DNA phylogeny, chose between rival pairs and did not simply copulate using the phylosig function in the phytools R package. or court with the first partner they encountered. We compared the distribution of Pagel’s λ between Results from females that were only courted by one sexes using the unpaired Wilcoxon rank sum test. male, or males that only courted one female were In order to test whether correlations exist between

13 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. chemical production and neuronal responses, we Acknowledgements: We thank Ibrahim Alali for applied phylogenetic generalized linear models fly rearing and helping in the courtship assays, and (PGLS). Raw neuronal response values were used Maide Erdogmus for helping in the electrophysiolog- as independent variables, and only z-transformed ical recordings. We are grateful to Kerstin Weniger, because visual inspection showed an approximately Sybille Lorenz, Silke Trautheim, Carolin Hoyer, and normal distribution of the dataset. Chemical levels Lisa Merkel for their technical support. Many thanks were transformed as described in the previous to Richard Benton for discussions and comments on section, and two encoding methods were used for the the manuscript, and to Marcus Stensmyr for discus- chemical levels – binary for presence or absence, or sions on the phylogeny. Wild-type flies were obtained continuous. When the chemical levels were binary- from the San Diego Drosophila Species Stock Center encoded, we used phylogenetic logistic regression (now The National Drosophila Species Stock Center, implemented in the R package phylolm and 2000 Cornell University) and KYOTO Stock Center. Fund- bootstraps to determine statistical significance. For ing: This research was supported through funding by continuous encoding of the chemical levels, we used the Max Planck Society. the PGLS method implemented in the R package caper, with the optimal branch transformation model Author Contributions: M.A.K, H.K.M.D., B.S.H., and determined by model selection with BIC as previously M.K. conceived of the project. All authors contributed described (93). to experimental design, analysis and interpretation of results. M.A.K. prepared all figures and collected all Selection pressure analysis. BUSTED (Branch- experimental data. R.C. and D.R.V. reconstructed the Site Unrestricted Statistical Test for Episodic phylogeny and performed all phylogenetic analyses. Diversification) was used to asses if a gene has A.S. identified and J.W. synthesized the sex-specific experienced a positive selection at any site at the compounds. Other experimental contributions were gene-wide level. BUSTED approach is available at the as follows: M.A.K (Fig.1, Fig.2a-C, Fig. 3a-D and datamonkey web server (https://www.datamonkey. F, Fig. 4, Extended Data Fig.1b-C’, Extended Data org/) (94). All branches of the three phylogenetic Fig.2, Extended Data Fig.3, Extended Data Fig.4, trees – including 42, 41, and 36 orthologs of Or47b, movies S1-427, and Tables S1,3-5, and 7), R.C. (Fig. Or67d, and Or88a, respectively – were entirely tested 1a, Fig.2d, Fig.3e, Fig.S1a-A’, and D, Tables S2, and for positive selection. 4), J.W. (Fig.2c, Fig.S2a). M.A.K. wrote the original manuscript, and M.K., R.C., D.R.V. and B.S.H. con- Statistics and figure preparations. Normality test tributed to the final manuscript. All coauthors contrib- was first assessed on datasets using a Shapiro uted to the subsequent revisions. test. Statistical analyses (see the corresponding legends of each figure) and preliminary figures were Competing interests: the authors declare that they conducted using GraphPad Prism v. 8 (https://www. have no competing interests. Data and biological graphpad.com). Figures were then processed with material availability: All relevant data supporting the Adobe Illustrator CS5. findings of this study and all unique biological mate- rials generated in this study are available from the corresponding authors upon request. Table Legends

Supplementary Table 1: List of the 99 species, their classification according to NCBI, their stock number, and their breeding media. Whole-genome sequences (WGS) for 58 of theses 99 species (written in blue) are generated in this study, while the genomes of the other 41 species (written in green) were available from the corresponding reference. Supplementary Table 2: Pagel’s lambda estimates that explain the correlation between the chemicals among the different male and female species and their phylogenetic relationships. Lambda values are plotted in Extended Data Fig.1d; male-specific compounds are identified and shown in Fig. 2d. Supplementary Table 3: List of the names (according to IUPAC), appearance times, Kovat's Index, chemical formula, exact mass, and m/z of male and female-specific compounds. Supplementary Table 4: Names and chemical classes of the compounds used in the SSR experiments. 21 of these compounds are synthesized in this study, while the other 15 are commercially available through the listed vendors and CAS numbers. Supplementary Table 5: Species-species interaction network through at1 and at4 sensillum. The number of edges represent the olfactory interaction between species pairs, while the number of self-loops represent the intraspecific interactions. Correlation coefficient of each species is given and indicate the number of edges between the species divided by number of edges that could possibly exist. Supplementary Table 6: Pagel’s lambda estimates that explain the phylogenetic corrected correlations between male-specific compounds and their neuronal response in at1 and at4. Supplementary Table 7: Values of sexual preferences of males and females of the different species. Statistical analy- sis between the tested groups is conducted via chi-square test.

14 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Extended data figures and tables

a 248 compounds a’ 256 compounds 99 species

−7.01 value 2.99 −5.59 value 3.48

b b’ 8 3

4 1.5

0

PC2 (21.7%) 0 PC2 (22.8 %)

-4 -1.5

-8 -3 -8 -4 0 4 8 -2 -1 0 1 2 3 PC1 (29.3%) PC1 (29.5%)

c c’

n=583 n=528

d Distribution of Pagel's lambda

p = 0.006

0.0 0.2 0.4 0.6 0.8 1.0 Pagel's lambda Extended Data Fig.1. Chemical analyses of 99 species within the family Drosophilidae (a) Heat map of normalized percentages of peak areas of 248 features as detected by XCMS across the males of 99 species (See Materials and Methods for details). Rows represent the species, which are ordered on each axis according their DNA phylogenetic relationships shown on the left side, while columns represent the 248 male chemical features present across the species sorted by lambda, highest Pagel’s lamb- da values to the left side of the panel (i.e., most correlated with phylogeny). Each cell represents the mean of the normalized percentages of the feature’s peak area in the different replicates of the same species. The feature’s percentage is calculated by dividing the peak area of a particular feature by the sum of areas of all features in the replicates of the same species. (a’) Normalized percentages of peak areas of 256 female chemical features across 99 female species. Similar to Extended Data Fig.1a, rows and columns represent the species and chemical features, respectively. (b) The first two principal components of male chemical profiles (data in Extended Data Fig.1a) of the 583 replicates across the 99 male species (>5 replicates per species) based on difference in peak areas of 248 male chemical features present across these replicates (see Materials and Methods for details). Data points of each group are enclosed within line. The lines’ fill is colored according to the group identities in Fig. 1a. (b’) The first two principal components of male chemical profiles (data in Extended Data Fig.1a’) of the 528 replicates across the 99 male species (>5 replicates per species) based on difference in peak areas of 256 male chemical features present across these replicates. (c) Chemometric hierarchical cluster analysis of 583 replicates across the 99 male species. Replicates are color coded according their spe- cies group in Fig. 1a. Cluster analysis was performed using neighbor joining (NJ) and correlation similarity index based on peaks’ quantities. S. latifasciaeformis was used to root the Chemometric tree. (c’) Chemometric hierarchical cluster analysis of 528 females (≥5 replicates per species) based on 256 female chemical features (see Mate- rials and Methods). Algorithm and parameter settings are similar to Fig.S1c. Note that male chemometric tree has recovered more monophy- letic groups than female chemometric tree (i.e., species groups cluster in males more than in females). (d) Frequency histogram of Pagel’s lambda estimates, showing more male chemicals (black) to be concordant with the phylogeny than in females (grey). * P = 0.006, Mann Whitney U test (n = 248 and 256 for male and female chemical compounds, respectively). 15 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

(Z,Z )-7,11-Pentacosadiene a 1 Extended Data Fig.2. The female-specific compounds in 99 species (Z )-7-Pentacosen 2 within the family Drosophilidae (Z,Z )-7,11-Heptacosadiene 3 (a) The chemical structure and names of the female-specific compounds according 2-Methyloctacosane to the International Union of Pure and Applied Chemistry (IUPAC). Out of 9 female- 4 (Z,Z )-7,11-Nonacosadiene specific compounds, 6 were identified. 5 (b) Distribution of 9 female-specific compounds among different drosophilids. Only 15 6 Unidentified 4 species display female-specific compounds. Species in black are dimorphic species, 2-Methyltriacontane 7 while in grey are monomorphic species (See Fig. 2b for more details). Phylogeny on 8 Unidentified 5 the left side is identical to the tree in Fig. 1a; the branches are colored according to (Z,Z )-9,23-tritriacontadiene the group identities. Numbers on the right side represent the sum of female-specific 9 compounds present per species, while numbers at the bottom of the table represent Not transferred number of times each female-specific compound appeared in the different drosophilids. b female-specific compounds Monomorphic / Dimorphic 1 2 3 4 5 6 7 8 9 D.moj.sonorensis 0 D.moj. baja 0 D.moj.wrigleyi 0 D.moj.mojavensis 0 D.arizonae 0 D.navojoa 0 D.wheeleri 0 D.mulleri 0 D.hamatofila 0 D.buzzatii 0 D.mercatorum 0 D.repleta 2 D.stalkeri 0 D. mettleri 1 D.hydei 2 D.gaucha 0 D.bromeliae 0 D.nannoptera 0 D.pseudotalamancana 0 D.americana 0 D.lummei 0 D.virilis 4 D. ittoralis 0 D.montana 0 D.paramelanica 0 D.melanica 0 D.euronotus 0 D.micromelanica 0 D.robusta 0 D.polychaeta 0 D.repletoides 2 Z.indianus 0 D.acutilabella 0 D.arawakana 0 D.dunni 0 D. ornatifrons 2 D.tripunctata 0 D.pallidipennis 0 D.testacea 0 D.putrida 0 D.macrospina 0 D.quinaria 0 D. palustris 0 D.albomicans 0 D.nasuta 0 D.sulfurigaster 0 D.immigrans 0 D.hypocausta 1 D.sternopleuralis 2 D.funebris 0 D.busckii 0 D.simulans 0 D.mauritiana 2 D.sechellia 5 D.melanogaster 3 D.yakuba 0 D.santomea 0 D.erecta 4 D.eugracilis 0 D.suzukii 0 D.biarmipes 0 D.takahashii 0 D.fuyamai 0 D.elegans 0 D.ficusphila 0 D.tsacasi 0 D.greeni 0 D.birchii 0 D.serrata 0 D.kikkawai 0 D.asahinai 0 D.lacteicornis 0 D.rufa 0 D.triauraria 0 D.auraria 0 D.malerkotliana 0 D.bipectinata 0 D.ananassae 0 D.persimilis 0 D.pseudoobscura 0 D.affinis 0 D.azteca 0 D.guanche 0 D.subobscura 0 D.imaii 0 D.bifasciata 0 D.paulistorum 0 D.sucinea 0 D.equinoxialis 0 D.willistoni 1 D.nebulosa 0 D.emarginata 0 D.neocordata 0 D.saltans 0 D.sturtevanti 1 C.procnemis 0 C.pararufithorax 0 S.lebanonensis 0 S.latifasciaeformis 1 2 1 4 5 4 3 4 4 6 33 0.04 16 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

a at1: 1 neuron at1: 2 neurons a’ at1: uncertain 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 At1 D.moj.sonorensis D.moj.sonorensis D.moj.baja At4 D.moj.baja D.moj.wrigleyi D.moj.wrigleyi D.moj.mojavensis D.moj.mojavensis D.arizonae D.arizonae D.navojoa D.navojoa D.mulleri D.mulleri D.hamatofila D.hamatofila D.buzzatii D.buzzatii D.mercatorum D.mercatorum D.repleta D.repleta 100 D.mettleri D.mettleri D.hydei D.hydei D.gaucha D.gaucha D.bromeliae D.bromeliae D.nannoptera D.nannoptera D.pseudotalamancana D.pseudotalamancana D.americana 50 D.americana D.virilis D.virilis D.melanica D.melanica D.robusta D.robusta D.polychaeta D.polychaeta D.repletoides D.repletoides D.dunni D.dunni D.pallidipennis D.pallidipennis D.testacea D.testacea D.putrida D.putrida 50 D.macrospina D.macrospina D.nasuta D.nasuta D.immigrans D.immigrans D.hypocausta D.hypocausta D.funebris D.funebris D.busckii D.busckii D.simulans D.simulans D.mauritiana D.mauritiana D.sechellia D.sechellia D.melanogaster D.melanogaster D.yakuba D.yakuba D.erecta D.erecta D.eugracilis D.eugracilis D.suzukii D.suzukii D.takahashii 0 D.takahashii D.elegans D.elegans 0 D.ficusphila D.ficusphila D.malerkotliana D.malerkotliana D.ananassae D.ananassae D.pseudoobscura D.pseudoobscura D.affinis D.affinis D.subobscura D.subobscura D.willistoni D.willistoni D.nebulosa D.nebulosa D.neocordata D.neocordata D.saltans D.saltans 0.04 D.sturtevanti 0.04 D.sturtevanti

b D.neocordata b’ At1 D.pseudotalamancana D.sturtevanti At4 D.subobscura D.ficusphila D.takahashii D.saltans D.elegans D.mauritiana D.nebulosa D.putrida D.hydei D.willistoni D.ananassae D.gaucha D.nannoptera D.moj.baja D.testacea D.mulleri D.repletoides D.dunni D.hypocausta D.melanogaster D.busckii D.virilis D.polychaeta D.willistoni D.testacea D.buzzatii D.navojoa D.virilis D.moj.moj. D.saltans D.malerkotliana D.ananassae D.macrospina D.bromeliae D.melanica D.mettleri D.arizonae D.gaucha D.funebris D.ficusphila D.mettleri D.pseudoobscura D.putrida D.hamatofila D.mulleri D.simulans D.polychaeta D.affinis D.navojoa D.pallidipennis D.arizonae D.subobscura D.buzzatii D.moj.wrig. D.macrospina D.immigrans D.sturtevanti D.repletoides D.mercatorumD.moj.wrig. D.pseudotalamancana D.americana D.melanica D.sechellia D.eugracilis D.pallidipennis D.sechellia D.repleta D.nebulosa D.simulans D.dunni D.hydei D.immigrans D.americana D.erecta D.moj.sono. D.moj.moj. D.hamatofila D.pseudoobscura D.nasuta D.moj.baja D.melanogaster D.erecta D.bromeliae D.takahashii D.mauritiana D.suzukii D.funebris D.nannoptera D.nasuta D.yakuba D.eugracilis D.mercatorum D.hypocausta D.repleta D.affinis D.yakuba D.busckii D.moj.sono. D.suzukii D.malerkotliana

Extended Data Fig.3. The intra- and inter-specific communication through trichoid sensilla (a) Responses of single sensillum recordings of at1 sensillum (Left: schematic drawing) in 54 species to 36 compounds used in Fig. 3b. Spe- cies names are arranged according to their phylogenetic relationship; the tree branches are colored according to the group identities. Species names are colored to reflect the presence of one (green) or two (red) neurons in at1 sensilla (95). Due to technical difficulties, neurons in at1 of species colored in black could not be counted by spike sorting. Note that at1 in D. pseudotalamancana and D. robusta could not be found. Color codes in the heat map illustrate the response values, which range from blue (inhibition) through white (no response) to red (activation). (a’) Electrophysiological responses of at4 neurons in 54 species to 36 compounds used in Fig. 3b. (b) Species-species interaction network through at1 sensillum. The directed edges (i.e., arrows that connect the 53 species (i.e., nodes)) represent the olfactory interaction between species pairs (2076 pairs). Direction of the arrow points to the species which is detected. Number of species that detect itself through at1 (i.e., self-loops) is 27 out of 49. The average of correlation coefficient is 0.51 (for more details seeFig. 3d and Supplementary Table 5). The network was analyzed by Cytoscape (for more details see Materials and Methods). (b’) Species-species interaction network through at4 sensillum. Number of edges is 1559 that connect 52 species, 21 of them are able to detect their own odors by at4 (i.e., self-loops). The average of correlation coefficient is 0.292.

17 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

a D.moj.sonorensis 20 D.moj. baja 24 D.moj.wrigleyi 20 D.moj.mojavensis 24 D.arizonae 24 D.navojoa 22 D.mulleri 24 D.hamatofila 20 D.buzzatii 20 D.mercatorum 18 D.repleta 46 D. mettleri 24 D.hydei 23 D.gaucha 16 D.bromeliae 24 D.nannoptera 40 D.pseudotalamancana 20 D.americana 19 D.virilis 24 D.melanica 32 D.robusta 24 D.polychaeta 16 D.repletoides 19 D.dunni 24 D.pallidipennis 24 D.testacea 28 D.putrida 22 D.macrospina 24 D.nasuta 44 D.immigrans 24 D.hypocausta 24 D.funebris 23 D.busckii 31 D.simulans 36 D.mauritiana 24 D.sechellia 31 D.melanogaster 32 D.yakuba 32 D.erecta 24 D.eugracilis 20 D.suzukii 28 D.takahashii 32 D.elegans 24 D.ficusphila 24 D.malerkotliana 32 D.ananassae 24 D.pseudoobscura 32 D.affinis 24 D.subobscura 22 D.willistoni 44 D.nebulosa 30 D.neocordata 20 0.04 D.saltans 24 D.sturtevanti 40

0 25 50 75 0 0 100 600 1200 1800 2400 3000 360048005400 600 1200 1800 2400 3000

Copulation success % Copulation latency (S) Copulation duration (S)

b 1200 de s de s an s penn is pu trida .f uneb ris .f uneb ris D. D.s al t D.busckii D.busckii D Headless D pa lli di 900 D.r ep letoi D.r ep letoi D.m oj.w r igle yi D. D.m oj. m oja v en sis

600

Pheromone Courtship index % DCM 300

0

DCM DCM DCM DCM DCM DCM DCM DCM DCM DCM DCM

(E)-ß-Farnesene

rac-2-Tridecyl acetate (Z)-10-Heptadecen-2-one rac-2-Pentadecyl acetate (Z)-11-Eicosen-1-yl acetate (Z)-11-Octadecen-1-yl acetate (Z)-11-Octadecen-1-yl(Z)-11-Octadecen-1-yl acetate (Z)-11-Hexadecen-1-yl acetate acetate

rac-(Z)-10-Heptadecen-2-ylrac-(Z)-10-Heptadecen-2-yl acetate acetate Extended Data Fig.4. Sexual behaviors of 54 species within the genus Drosophila (a) Left: Copulation success [%] of virgin couples in the different species (illustrated in schematics on left side) within a 1-hour time window. For more details on the courtship behaviors of a particular species, watch its movies (Movies1 to 427, available on Edmond). Species are arranged in respect to phylogeny; color of the tree branches corresponds to the species group. Number of replicates is indicated right of spe- cies names. Middle: Copulation latency in seconds (s). Males exhibiting no courtship behavior were excluded from analysis. Right: Copulation duration of the different species in seconds (s). Unlike the prolonged copulation time in melanogaster group (≥ 15 min) (53), copulation lasts for ~2-3 min in most species of repleta group. In this and the below panels, age of males and females is 10 days. (b) Left: Schematic of courtship arena where a decapitated female is courted by a conspecific male perfumed with DCM or one of the other compounds. Right, y-axis represents courtship index [%] (equal the time a male exhibits courtship behaviors / total amount of recording time (10 minutes)). Ns P > 0.05, Mann Whitney U test (n = 10 replicates).

18 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. References

1 Smadja, C. & Butlin, R. K. On the scent of speciation: the chemosensory system and its role in premating isolation. Heredity 102, 77-97, doi:10.1038/ hdy.2008.55 (2009). 2 Andersson, M. Sexual selection. (Princeton University Press Princeton, 1994). 3 Boughman, J. W. How sensory drive can promote speciation. Trends Ecol Evol 17, 571-577, doi:10.1016/S0169-5347(02)02595-8 (2002). 4 Coyne, J. A. & Orr, H. A. Speciation. (Sinauer Associates, 2004). 5 Symonds, M. R. & Elgar, M. A. The evolution of pheromone diversity. Trends Ecol Evol 23, 220-228, doi:10.1016/j.tree.2007.11.009 (2008). 6 Dobzhansky, T. Genetics of the evolutionary process. (Columbia University Press, 1970). 7 Cardé, R. T. & Haynes, K. F. Structure of the pheromone communication channel in moths. (Cambridge University Press, 2004). 8 de Bruyne, M. & Baker, T. C. Odor detection in insects: volatile codes. J Chem Ecol 34, 882-897, doi:10.1007/s10886-008-9485-4 (2008). 9 Turelli, M., Barton, N. H. & Coyne, J. A. Theory and speciation. Trends Ecol Evol 16, 330–343 (2001). 10 Funk, D. J., Filchak, K. E. & Feder, J. L. Herbivorous insects: model systems for the comparative study of speciation ecology. Genetica 116, 251-267, doi:Doi 10.1023/A:1021236510453 (2002). 11 Endler, J. A. & Basolo, A. L. Sensory ecology, receiver biases and sexual selection. Trends Ecol Evol 13, 415-420, doi:Doi 10.1016/S0169-5347(98)01471-2 (1998). 12 Johansson, B. G. & Jones, T. M. The role of chemical communication in mate choice. Biol Rev 82, 265-289, doi:10.1111/j.1469-185X.2007.00009.x (2007). 13 Wyatt, T. D. Pheromones and signature mixtures: defining species-wide signals and variable cues for identity in both invertebrates and vertebrates.J Comp Physiol A Neuroethol Sens Neural Behav Physiol 196, 685-700, doi:10.1007/s00359-010-0564-y (2010). 14 Thomas, M. L. & Simmons, L. W. Crickets detect the genetic similarity of mating partners via cuticular hydrocarbons. J Evolution Biol 24, 1793-1800, doi:10.1111/j.1420-9101.2011.02319.x (2011). 15 Jezovit, J. A., Levine, J. D. & Schneider, J. Phylogeny, environment and sexual communication across the Drosophila genus. J Exp Biol 220, 42-52, doi:10.1242/jeb.143008 (2017). 16 Markow, T. A. & O’Grady, P. M. Evolutionary genetics of reproductive behavior in Drosophila: connecting the dots. Annual Review of Genetics 39, 263–291, doi:10.1146/annurev.genet.39.073003.112454 (2005). 17 Spieth, H. T. Mating behavior within the genus Drosophila (Diptera). B Am Mus Nat Hist 99, 401-474 (1952). 18 Steele, R. H. Courtship feeding in Drosophila subobscura .1. The nutritional significance of courtship feeding.Anim Behav 34, 1087-1098, doi:Doi 10.1016/ S0003-3472(86)80168-3 (1986). 19 LaRue, K. M., Clemens, J., Berman, G. J. & Murthy, M. Acoustic duetting in Drosophila virilis relies on the integration of auditory and tactile signals. Elife 4, doi:ARTN e0727710.7554/eLife.07277 (2015). 20 O’Grady, P. M. & Markow, T. A. Rapid morphological, behavioral, and ecological evolution in Drosophila: comparisons between the endemic Hawaiian Dro- sophila and the cactophilic repleta species group. Rapidly Evolving Genes and Genetic Systems 1, 176-186 (2012). 21 Khallaf, M. A. et al. Mate discrimination among subspecies through a conserved olfactory pathway. Sci Adv 6, eaba5279 (2020). 22 Auer, T. O. & Benton, R. Sexual circuitry in Drosophila. Current opinion in neurobiology 38, 18-26, doi:10.1016/j.conb.2016.01.004 (2016). 23 van Naters, W. V. G. & Carlson, J. R. Receptors and neurons for fly odors in Drosophila. Curr Biol 17, 606-612, doi:10.1016/j.cub.2007.02.043 (2007). 24 Symonds, M. R. E. & Wertheim, B. The mode of evolution of aggregation pheromones in Drosophila species. J Evolution Biol 18, 1253-1263, doi:10.1111/ j.1420-9101.2005.00971.x (2005). 25 Grabe, V., Strutz, A., Baschwitz, A., Hansson, B. S. & Sachse, S. Digital In Vivo 3d Atlas of the Antennal Lobe of Drosophila melanogaster. J Comp Neurol 523, 530-544, doi:10.1002/cne.23697 (2015). 26 Couto, A., Alenius, M. & Dickson, B. J. Molecular, anatomical, and functional organization of the Drosophila olfactory system. Curr Biol 15, 1535-1547, doi:DOI 10.1016/j.cub.2005.07.034 (2005). 27 Dickson, B. J. Wired for sex: The neurobiology of Drosophila mating decisions. Science 322, 904-909, doi:10.1126/science.1159276 (2008). 28 Ejima, A. Pleiotropic actions of the male pheromone cis-vaccenyl acetate in Drosophila melanogaster. J Comp Physiol A 201, 927-932, doi:10.1007/s00359- 015-1020-9 (2015). 29 Vander Meer, R. K., Obin, M. S., Zawistowski, S., Sheehan, K. B. & Richmond, R. C. A reevaluation of the role of cis-vaccenyl acetate, cis-vaccenol and esterase 6 in the regulation of mated female sexual attractiveness in Drosophila melanogaster. Journal of Insect Physiology 32, 681-686, doi:10.1016/0022- 1910(86)90109-5 (1986). 30 Bartelt, R. J., Schaner, A. M. & Jackson, L. L. Cis-vaccenyl acetate as an aggregation pheromone in Drosophila melanogaster. Journal of Chemical Ecology 11, 1747-1756, doi:Doi 10.1007/Bf01012124 (1985). 31 Kurtovic, A., Widmer, A. & Dickson, B. J. A single class of olfactory neurons mediates behavioural responses to a Drosophila sex pheromone. Nature 446, 542-546, doi:10.1038/nature05672 (2007). 32 Ha, T. S. & Smith, D. P. A pheromone receptor mediates 11-cis-vaccenyl acetate-induced responses in Drosophila. J Neurosci 26, 8727-8733, doi:10.1523/ Jneurosci.0876-06.2006 (2006). 33 Wang, L. & Anderson, D. J. Identification of an aggression-promoting pheromone and its receptor neurons in Drosophila. Nature 463, 227–231 (2010). 34 Dumenil, C. et al. Pheromonal cues deposited by mated females convey social information about egg-laying sites in Drosophila melanogaster. J Chem Ecol 42, 259-269, doi:10.1007/s10886-016-0681-3 (2016). 35 Lebreton, S. et al. Feeding regulates sex pheromone attraction and courtship in Drosophila females. Sci Rep-Uk 5, doi:Artn 1313210.1038/Srep13132 (2015). 36 Gompel, N. & Carroll, S. B. Genetic mechanisms and constraints governing the evolution of correlated traits in drosophilid flies. Nature 424, 931-935, doi:10.1038/nature01787 (2003). 37 Kuntz, S. G. & Eisen, M. B. Drosophila embryogenesis scales uniformly across temperature in developmentally diverse species. PLoS Genet 10, e1004293, doi:10.1371/journal.pgen.1004293 (2014). 38 Zhao, Z. & McBride, C. S. Evolution of olfactory circuits in insects. J Comp Physiol A Neuroethol Sens Neural Behav Physiol, doi:10.1007/s00359-020- 01399-6 (2020). 39 O’Grady, P. M. & DeSalle, R. Phylogeny of the genus Drosophila. Genetics 209, 1-25, doi:10.1534/genetics.117.300583 (2018). 40 Throckmorton, L. H. The phylogeny, ecology and geography of Drosophila. 421–469 (Plenum Publishing Corporation, 1975). 41 van der Linde, K., Houle, D., Spicer, G. S. & Steppan, S. J. A supermatrix-based molecular phylogeny of the family Drosophilidae. Genetics research 92, 25-38, doi:10.1017/S001667231000008X (2010). 42 Toda, H., Zhao, X. L. & Dickson, B. J. The Drosophila female aphrodisiac pheromone activates ppk23(+) sensory neurons to elicit male courtship behavior. Cell Rep 1, 599-607, doi:10.1016/j.celrep.2012.05.007 (2012). 43 Seeholzer, L. F., Seppo, M., Stern, D. L. & Ruta, V. Evolution of a central neural circuit underlies Drosophila mate preferences. Nature 559, 564-569, doi:10.1038/s41586-018-0322-9 (2018). 44 Yew, J. Y. et al. A new male sex pheromone and novel cuticular cues for chemical communication in Drosophila. Curr Biol 19, 1245-1254, doi:10.1016/j. cub.2009.06.037 (2009). 45 Guiraudie-Capraz, G., Pho, D. B. & Jallon, J. M. Role of the ejaculatory bulb in biosynthesis of the male pheromone cis-vaccenyl acetate in Drosophila melanogaster. Integr Zool 2, 89-99, doi:10.1111/j.1749-4877.2007.00047.x (2007). 46 Shirangi, T. R., Dufour, H. D., Williams, T. M. & Carroll, S. B. Rapid evolution of sex pheromone-producing enzyme expression in Drosophila. Plos Biol 7, doi:10.1371/journal.pbio.1000168 (2009). 47 Dweck, H. K. M. et al. Pheromones mediating copulation and attraction in Drosophila. P Natl Acad Sci USA 112, E2829-E2835, doi:10.1073/pnas.1504527112 (2015). 48 Pegel, M. Inferring the historical patterns of biological evolution. Nature 401, 877–884. (1999). 49 Miller, C. J. & Carlson, J. R. Regulation of Odor Receptor Genes in Trichoid Sensilla of the Drosophila Antenna. Genetics 186, 79-U166, doi:10.1534/ge- netics.110.117622 (2010).

19 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.21.305854; this version posted September 22, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 50 Lin, C. C., Prokop-Prigge, K. A., Preti, G. & Potter, C. J. Food odors trigger Drosophila males to deposit a pheromone that guides aggregation and female oviposition decisions. Elife 4, doi:10.7554/eLife.08688 (2015). 51 El-Sayed, A. M. The pherobase: Database of pheromones and semiochemicals. (2009). 52 Revadi, S. et al. Sexual behavior of Drosophila suzukii. Insects 6, 183-196, doi:10.3390/insects6010183 (2015). 53 Markow, T. A. Evolution of Drosophila mating systems. Evolutionary Biology 29, 73–106 (1996). 54 Spieth, H. T. Courtship behavior in Drosophila. Annu Rev Entomol 19, 385-405, doi:DOI 10.1146/annurev.en.19.010174.002125 (1974). 55 Zawistowski, S. & Richmond, R. C. Inhibition of courtship and mating of Drosophila melanogaster by the male-produced lipid, cis vaccenyl acetate. J. Insect Physiol 32, 189–192, doi:10.1016/0022-1910(86)90057-0 (1986). 56 Lande, R. Models of speciation by sexual selection on polygenic traits. P Natl Acad Sci USA 78, 3721–3725 (1981). 57 Ritchie, M. G. Sexual selection and speciation. Annual Review of Ecology, Evolution, and Systematics 38, 79-102, doi:10.1146/annurev.ecol- sys.38.091206.095733 (2007). 58 Mayr, E. Systematics and the origin of species. (Columbia University Press, 1942). 59 van Wilgenburg, E., Symonds, M. R. & Elgar, M. A. Evolution of cuticular hydrocarbon diversity in ants. J Evol Biol 24, 1188-1198, doi:10.1111/j.1420- 9101.2011.02248.x (2011). 60 Martin, S. J., Helantera, H. & Drijfhout, F. P. Evolution of species-specific cuticular hydrocarbon patterns in Formica ants. Biological Journal of the Linnean Society 95, 131-140, doi:10.1111/j.1095-8312.2008.01038.x (2008). 61 Akino, T., Yamamura, K., Wakamura, S. & Yamaoka, R. Direct behavioral evidence for hydrocarbons as nestmate recognition cues in Formica japonica (Hymenoptera : Formicidae). Applied Entomology and Zoology 39, 381–387, doi:10.1303/aez.2004.381 (2004). 62 Martin, S. J., Vitikainen, E., Shemilt, S., Drijfhout, F. P. & Sundstrom, L. Sources of variation in cuticular hydrocarbons in the ant Formica exsecta. J Chem Ecol 39, 1415-1423, doi:10.1007/s10886-013-0366-0 (2013). 63 Leonhardt, S. D., Rasmussen, C. & Schmitt, T. Genes versus environment: geography and phylogenetic relationships shape the chemical profiles of sting- less bees on a global scale. P Roy Soc B-Biol Sci 280, doi:10.1098/rspb.2013.0680 (2013). 64 Lang, C. & Menzel, F. Lasius niger ants discriminate aphids based on their cuticular hydrocarbons. Anim Behav 82, 1245-1254, doi:10.1016/j.anbe- hav.2011.08.020 (2011). 65 Magro, A. et al. Oviposition deterring infochemicals in ladybirds: the role of phylogeny. Evol Ecol 24, 251-271, doi:10.1007/s10682-009-9304-6 (2010). 66 Baker, T. C. Mechanism for saltational shifts in pheromone communication systems. P Natl Acad Sci USA 99, 13368-13370, doi:10.1073/pnas.222539799 (2002). 67 Symonds, M. R. E. & Gitau-Clarke, C. W. The evolution of aggregation pheromone diversity in bark beetles. Adv Insect Physiol 50, 195-234, doi:10.1016/ bs.aiip.2015.12.003 (2016). 68 Symonds, M. R. E., Moussalli, A. & Elgar, M. A. The evolution of sex pheromones in an ecologically diverse genus of flies.Biological Journal of the Linnean Society 97, 594-603, doi:DOI 10.1111/j.1095-8312.2009.01245.x (2009). 69 Saveer, A. M. et al. Mate recognition and reproductive isolation in the sibling species Spodoptera littoralis and Spodoptera litura. Front Ecol Evol 2, doi:10.3389/fevo.2014.00018 (2014). 70 Higgie, M., Chenoweth, S. & Blows, M. W. Natural selection and the reinforcement of mate recognition. Science 290, 519-521, doi:DOI 10.1126/sci- ence.290.5491.519 (2000). 71 Keesey, I. W. et al. Inverse resource allocation between vision and olfaction across the genus Drosophila. Nat Commun 10, 1162, doi:10.1038/s41467-019- 09087-z (2019). 72 Fuyama, Y. Visual stimulus in the courtship of Drosophila suzukii. Experientia 35, 1327-1328, doi:Doi 10.1007/Bf01963987 (1979). 73 Dyer, K. A., White, B. E., Sztepanacz, J. L., Bewick, E. R. & Rundle, H. D. Reproductive character displacement of epicuticular compounds and their con- tribution to mate choice in Drosophila subquinaria and Drosophila recens. Evolution 68, 1163-1175, doi:10.1111/evo.12335 (2014). 74 Roelofs, W. L. et al. Evolution of moth sex pheromones via ancestral genes. P Natl Acad Sci USA 99, 13621-13626, doi:10.1073/pnas.152445399 (2002). 75 Finet, C., Slavik, K., Pu, J., Carroll, S. B. & Chung, H. Birth-and-death evolution of the fatty acyl-CoA reductase (FAR) gene family and diversification of cuticular hydrocarbon synthesis in Drosophila. Genome Biol Evol 11, 1541-1551, doi:10.1093/gbe/evz094 (2019). 76 Grillet, M., Dartevelle, L. & Ferveur, J. F. A Drosophila male pheromone affects female sexual receptivity. P Roy Soc B-Biol Sci 273, 315-323, doi:10.1098/ rspb.2005.3332 (2006). 77 Partridge, L. & Fowler, K. Non-mating costs of exposure to males in female Drosophila melanogaster. Journal of Insect Physiology 36, 419–425 ( 1990). 78 Coyne, J. A. Genetics of sexual isolation between 2 sibling species, Drosophila simulans and Drosophila mauritiana. P Natl Acad Sci USA 86, 5464-5468, doi:DOI 10.1073/pnas.86.14.5464 (1989). 79 Noor, M. A. F. Absence of species discrimination in Drosophila pseudoobscura and D. persimilis males. Anim Behav 52, 1205-1210, doi:DOI 10.1006/ anbe.1996.0268 (1996). 80 Kandul, N. P., Wright, K. M., Kandul, E. V. & Noor, M. A. F. No evidence for learned mating discrimination in male Drosophila pseudoobscura. Bmc Evol Biol 6, doi:10.1186/1471-2148-6-54 (2006). 81 Billeter, J. C., Atallah, J., Krupp, J. J., Millar, J. G. & Levine, J. D. Specialized cells tag sexual and species identity in Drosophila melanogaster. Nature 461, 987-U250, doi:10.1038/nature08495 (2009). 82 Sharma, M. D., Hunt, J. & Hosken, D. J. Antagonistic responses to natural and sexual selection and the sex-specific evolution of cuticular hydrocarbons in Drosophila simulans. Evolution 66, 665-677, doi:10.1111/j.1558-5646.2011.01468.x (2012). 83 McBride, C. S. Rapid evolution of smell and taste receptor genes during host specialization in Drosophila sechellia. P Natl Acad Sci USA 104, 4996-5001, doi:10.1073/pnas.0608424104 (2007). 84 Auer, T. O. et al. Olfactory receptor and circuit evolution promote host specialization. Nature 579, 402-408, doi:10.1038/s41586-020-2073-7 (2020). 85 Ramasamy, S. et al. The evolution of olfactory gene families in Drosophila and the genomic basis of chemical-ecological adaptation in Drosophila suzukii. Genome Biol Evol 8, 2297-2311, doi:10.1093/gbe/evw160 (2016). 86 Prieto-Godino, L. L. et al. Evolution of acid-sensing olfactory circuits in drosophilids. Neuron 93, 661-676, doi:10.1016/j.neuron.2016.12.024 (2017). 87 Karageorgi, M. et al. Genome editing retraces the evolution of toxin resistance in the monarch butterfly. Nature 574, 409-412, doi:10.1038/s41586-019- 1610-8 (2019). 88 Leary, G. P. et al. Single mutation to a sex pheromone receptor provides adaptive specificity between closely related moth species. P Natl Acad Sci USA 109, 14081-14086, doi:10.1073/pnas.1204661109 (2012). 89 Rosenthal, G. G. Evaluation and hedonic value in mate choice. Curr Zool 64, 485-492, doi:10.1093/cz/zoy054 (2018). 90 Stokl, J. et al. A deceptive pollination system targeting drosophilids through olfactory mimicry of yeast. Curr Biol 20, 1846-1852, doi:10.1016/j.cub.2010.09.033 (2010). 91 Olsson, S. B. & Hansson, B. S. Electroantennogram and single sensillum recording in insect antennae. Methods Mol Biol 1068, 157-177, doi:10.1007/978- 1-62703-619-1_11 (2013). 92 Ng, R., Lin, H. H., Wang, J. W. & Su, C. Y. Electrophysiological recording from Drosophila trichoid sensilla in response to odorants of low volatility. Jove-J Vis Exp, doi:ARTN e5614710.3791/56147 (2017). 93 Cui, R. F. et al. Relaxed selection limits lifespan by increasing mutation load. Cell 178, 385-399, doi:10.1016/j.cell.2019.06.004 (2019). 94 Murrell, B. et al. Gene-wide identification of episodic selection. Mol Biol Evol 32, 1365-1371, doi:10.1093/molbev/msv035 (2015). 95 Prieto-Godino, L. L. et al. Functional integration of “undead” neurons in the olfactory system. Sci Adv 6, eaaz7238, doi:10.1126/sciadv.aaz7238 (2020).

20