evolution and sexual behavior in are shaped by male sensory exploitation of other males

Soon Hwee Nga, Shruti Shankara,b, Yasumasa Shikichic, Kazuaki Akasakad, Kenji Moric, and Joanne Y. Yewa,b,1

aTemasek Life Sciences Laboratory, National University of Singapore, Singapore 117604; bDepartment of Biological Sciences, National University of Singapore, Singapore 117546; cPhotosensitive Materials Research Center, Toyo Gosei Co., Ltd, Inzai-shi, Chiba 270-1609, Japan; and dShokei Gakuin University, Natori-shi, Miyagi 981-1295, Japan

Edited* by Edward A. Kravitz, Harvard Medical School, Boston, MA, and approved December 31, 2013 (received for review July 19, 2013) exhibit a spectacular array of traits to attract mates. are taste and olfactory cues that, in many , Understanding the evolutionary origins of sexual features and play an important role in mate selection (15). As with courtship preferences is a fundamental problem in evolutionary biology, and cues detected by other sensory modalities, pheromones are the mechanisms remain highly controversial. In some species, shaped by and, thus, may exhibit enormous choose mates based on direct benefits conferred by the structural diversity and exquisite stereochemical specificity. In male to the and her offspring. Thus, female preferences are , exogenously secreted lipids advertise status, avail- thought to originate and coevolve with male traits. In contrast, ability, and reproductive fitness (16). In some cases, male pher- sensory exploitation occurs when expression of a male trait takes omones serve as a nuptial gift, thus providing direct reproductive advantage of preexisting sensory biases in females. Here, we benefits to females and offspring in the form of either nutritive document in Drosophila a previously unidentified example of sen- or defensive compounds (17). Little is known, however, about sory exploitation of males by other males through the use of the the mechanisms underlying the diversification and the origin of sex pheromone CH503. We use mass spectrometry, high-perfor- chemical specificity. Here, we provide an example of a phero- mance liquid chromatography, and behavioral analysis to demonstrate mone that has evolved from sensory exploitation. In Drosophila that an antiaphrodisiac produced by males of the melanogaster melanogaster, CH503 [formally, (3R,11Z,19Z)-3-acetoxy-11,19- subgroup also is effective in distant Drosophila relatives that do octacosadien-1-ol; Fig. 1A] functions as an antiaphrodisiac (18). not express the pheromone. We further show that species that The pheromone is secreted in the anogenital region, is transferred produce the pheromone have become less sensitive to the com- to females during mating, and suppresses courtship from males. pound, illustrating that sensory occurs after sensory exploitation. Our findings provide a mechanism for the origin of Our findings indicate CH503 evolved from males exploiting the a sex pheromone and show that sensory exploitation changes male preexisting sensory biases of other males to gain mating advan- sexual behavior over evolutionary time. tage by limiting access to females. Moreover, the use of CH503 has altered male sexual behavior over evolutionary time such that sexual selection | laser desorption/ionization | supernormal stimulus | males have adapted by becoming less sensitive to the pheromone. male–male competition | chiral pheromone Results and Discussion Evolutionary Origin of CH503 Expression. To determine the evolu- exual selection is widely regarded as an important mecha- tionary origins of CH503, we examined eight species of Drosophila Snism for the origin of new traits and species. Darwin first proposed that the elaboration of male secondary sexual traits is Significance driven by female preferences (1, 2). This concept has been re- fined by models suggesting that females select male traits that indicate genetic quality or confer direct reproductive benefits (3– How sexual features and preferences originate and evolve is one 7). In contrast, sensory exploitation occurs when expression of of the most important and contentious problems in evolutionary biology. In some species, females choose male traits that indicate a male trait takes advantage of preexisting sensory biases in genetic quality or confer benefits. Thus, male traits and female females (8). In this case, female preference does not coevolve preferences are assumed to originate at the same time and to with the male trait but rather precedes it. In one of the first coevolve. In contrast, sensory exploitation occurs when males examples documenting sensory exploitation, female Physalaemus take advantage of females’ preexisting sensory biases. We show coloradorum were shown to prefer male calls that contain in Drosophila sensory exploitation of males by other males alow-frequency“chuck” component despite the absence of this through the use of a pheromone to gain sexual advantage. feature in calls from conspecifics. The sensory bias for chucks was Notably, sensory exploitation leads to male sensory adaptation. shown to have its mechanistic basis in the tuning properties of the These findings provide a mechanism for the evolutionary origins inner ear, a physiological feature that predated the appearance of of a pheromone and are a previously unidentified example of chucks (9). Similarly, female platyfish exhibit a preference for sensory exploitation between males. males with swordtails despite the absence of swordtails in male platyfish. Females consistently chose to spend more time with Author contributions: S.H.N., S.S., Y.S., K.A., K.M., and J.Y.Y. designed research; S.H.N., S.S., Y.S., K.A., K.M., and J.Y.Y. performed research; Y.S., K.A., and K.M. contributed new conspecific males exhibiting an artificially attached plastic sword reagents/analytic tools; S.H.N., S.S., Y.S., K.A., K.M., and J.Y.Y. analyzed data; and S.H.N., (10). In both these examples, female preference predates expres- K.A., K.M., and J.Y.Y. wrote the paper. sion of the trait. Sensory exploitation has since been documented The authors declare no conflict of interest. for numerous other visual cues, across a diversity of taxa (11–14). *This Direct Submission article had a prearranged editor. In each case, females prefer traits that are not found naturally in Freely available online through the PNAS open access option. their own species but appear in males of other species. Moreover, 1To whom correspondence should be addressed. E-mail: [email protected]. both the sensory bias and behavioral response to the trait already This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. were present before expression of the trait. 1073/pnas.1313615111/-/DCSupplemental.

3056–3061 | PNAS | February 25, 2014 | vol. 111 | no. 8 www.pnas.org/cgi/doi/10.1073/pnas.1313615111 Downloaded by guest on September 29, 2021 for production of the pheromone CH503 and tested whether cVA A males of these species respond to CH503 as an antiaphrodisiac. 349.24 We first used UV laser desorption/ionization mass spectrometry (3R,11Z,19Z)-CH503 (natural stereoisomer) (UV-LDI MS) to analyze the chemical profiles of the male OH n-C8H17 anogenital region in other species. UV-LDI MS revealed sig- OAc nals matching the expected molecular weight for CH503 in the anogenital region of D. melanogaster, Drosophila simulans, Dro- CH503 sophila yakuba, Drosophila sechellia,andDrosophila ananassae 503.38 (Fig. 1A). A signal for cis-vaccenyl acetate (cVA), another known antiaphrodisiac, also was found in these species (19). In contrast, no signal for CH503 could be detected from the anogenital region D. melanogaster of Drosophila willistoni, Drosophila mojavensis,orDrosophila virilis (Fig. 1A). To determine the double-bond geometries and absolute D. simulans configuration of CH503 from melanogaster group , chemical D. yakuba derivatization and high-performance liquid chromatography (HPLC) separation were used to compare the retention times D. sechelia of the derivative of naturally occurring CH503 with synthetic 003 m/z 056 standards of the eight possible derivatized stereoisomers. Each of the derivatized stereoisomers could be differentiated based on D. ananassae D. willistoni their distinct retention times (Fig. 1B). D. melanogaster previously 501.37 405.32 was shown to express (3R,11Z,19Z)-CH503 [hereafter abbreviated 471.43 as (R,Z,Z)-CH503] (20). HPLC analysis of derivatized thin-layer 503.39 chromatography-purified fractions revealed a single peak corre- cVA sponding to (R,Z,Z)-CH503, indicating that as with D. melanogaster, 349.25 407.26 D. simulans, D. yakuba, and D. sechellia express a single stereo-

485.40 515.42 isomer (Fig. 1B). Although a compound with a mass signal cor- responding to (R,Z,Z)-CH503 was observed by UV-LDI MS analysis in D. ananassae, the retention time did not match any D. mojavensis 559.34 D. virilis 577.37 of the eight stereoisomers. Taken together, chemical profiling with 499.45 563.35 UV-LDI MS and HPLC reveals that all tested drosophilids of the 487.39 515.42 551.34 melanogaster subgroup express (R,Z,Z)-CH503. In contrast, other 471.41 drosophilids representing outgroup taxa from Sophophora 507.28 443.38 (D. willistoni, D. ananassae)andDrosophila (D. virilis, D. mojavensis) 415.37 do not express (R,Z,Z)-CH503 or any other stereoisomer. The

phylogenetic distribution of the trait suggests a single origin in EVOLUTION 003 m/z 056 003 m/z 056 the melanogaster group of Drosophila.

Conserved Function of CH503 as an Antiaphrodisiac. We next tested B 37.3 whether the function of CH503 as an antiaphrodisiac is con- 3 2 32.5 served across the different species. Socially isolated virgin males 5 D. simulans were placed with a virgin female perfumed with various amounts 6 1 of CH503. Surprisingly, all Drosophila species tested suppressed 4 7 8 courtship initiation in a dose-dependent manner in response to 37.6 CH503, although the pheromone is produced only by a subgroup 32.7 D. yakuba of these species. Male courtship behavior was significantly inhibi- ted in D. melanogaster, D. simulans, D. yakuba,andD. sechellia,all 37.4 species that produce CH503 (Fig. 2A). The latency to courtship 32.6 initiation also increased significantly in a dose-responsive manner to increasing amounts of CH503 (Fig. S1). Notably, D. ananassae, min. D. sechellia 0 08 D. willistoni, D. mojavensis,andD. virilis, none of which expresses Derivatized CH503 isomer and RT (min) 24.4 CH503, also responded to the compound by suppressing courtship 1. (3S,11Z,19Z) 34.8 5. (3R,11E,19Z) 46.7 behavior and delaying courtship initiation (Fig. 2B, Fig. S1,and 2. (3R,11Z,19Z) 37.5 6. (3R,11Z,19E) 48.8 Table S1). These findings indicate that the behavioral response to D. ananassae 3. (3S,11E,19Z) 42.8 7. (3S,11E,19E) 68.3 CH503 predates the expression of CH503. We hypothesize that 4. (3S,11Z,19E) 44.7 8. (3R,11E,19E) 74.8 08min. 0males of the ancestral species that gave rise to the melanogaster Fig. 1. Characterization of CH503 expression in Drosophila.(A) Chemical struc- subgroup used sensory exploitation to inhibit courtship from male ture of CH503 and representative UV-LDI mass spectra measured from the male competitors. anogenital region of different Drosophila species. Each spectrum is recorded from To test whether production of CH503 might provide an ad- a single . Signals corresponding to the mass-to-charge ratio (m/z)forcVA(m/z vantage in male–male competition for access to females, we ex- 349.24) and CH503 (m/z 503.38) were detected in D. melanogaster, D. simulans, amined the effect of introducing (R,Z,Z)-CH503 into the mating D. yakuba, D. sechellia,andD. ananassae. No signal for CH503 was detected from system of species that do not produce the pheromone. D. ananassae D. willistoni, D. mojavensis,orD. virilis. Potassium-bearing molecular compounds and D. virilis males were given a choice of mating with CH503- + + [M K] constitute the major ion species in all cases. (B) The HPLC chromatogram perfumed or solvent-perfumed females. In both species, males shows distinct retention times (RT) for each of the eight synthetic CH503 stereo- E isomers following derivatization. HPLC analysis of derivatized CH503 isolated showed a significant aversion to courting the former (Fig. 2 ). Thus, from D. simulans, D. yakuba, and D. sechellia reveals that (3R,11Z,11Z)-CH503 is the use of a potent antiaphrodisiac by males could significantly shift the only expressed stereoisomer. The retention times for the major peaks are courtship choice of rival males. By limiting access to mated females, noted in each chromatogram. The compound isolated from D. ananassae has the male producers of the pheromone might potentially gain a mating same m/z and elemental composition as CH503 but a different structure. advantage by reducing sperm competition.

Ng et al. PNAS | February 25, 2014 | vol. 111 | no. 8 | 3057 Downloaded by guest on September 29, 2021 ABResponse to (R,Z,Z)-CH503 Response to (R,Z,Z)-CH503 100 100

80 ** 80 ns ns ns *** ** * ns 60 ns 60 * ** **** ** *** **** **** *** ** *** 40 40 **** **** **** **** **** 20 20 Courtship initiation (%) Courtship initiation (%) 0 0 0 83 249 415 0 0 580 700 830 1245 1660 2075 580 830 83 249 415 0 83 166 249 0 0 83 415 830 83 166 249 0 1000 0 Dmel Dsim Dyak Dsec Dana Dwil Dmoj Dvir CDResponse to (S,Z,Z)-CH503 Response to (S,Z,Z)-CH503 100

100 ns ** 80 80 * **** **** ns 60 * 60 *** **

** * **

*** 40

40 **** **** *** **** **** **** **** 20 **** **** 20 **** **** Courtship initiation (%) Courtship initiation (%) 415 415 415 249 249 0 83 415 830 0 249 332 0 83 249 415 415 83 249 0 0 0 0 0 0 83 249 415 83 83 249 415 83 0 Dmel Dsim Dyak Dsec Dana Dwil Dmoj Dvir E Courtship choice D. ananassae D. virilis D. melanogaster ns ns **** ns **** ** * 1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2 Courtship vigor Courtship vigor Courtship vigor 0.0 0.0 0.0 Control 250ng Control 250ng Control 83ng Control 83ng Control 83ng Control 83ng 83ng 83ng TB-CH503 (R,Z,Z)- TB-CH503 (R,Z,Z)- TB- (S,Z,Z)- (R,Z,Z)- (S,Z,Z)- CH503 CH503 CH503 CH503 CH503 CH503

Fig. 2. Comparative analysis of the behavioral response to natural CH503 and an artificial stereoisomer. (A and B) Courtship assays reveal that the natural pheromone (R,Z,Z)-CH503 functions as an antiaphrodisiac both in species that produce the natural pheromone (A) and in species that do not (B). Doses (in nanograms) indicated represent the approximate amount on the cuticular surface following perfuming. The courtship initiation percentage for each dose is compared with the percentage from control trials and statistically assessed using a Holm–Bonferroni-corrected Fisher exact test. Dana, D. ananassae; Dmel, D. melanogaster; Dmoj, D. mojavensis; Dsec, D. sechellia; Dsim, D. simulans; Dvir, D. virilis; Dwil, D. willistoni; Dyak, D. yakuba; ns, nonsignificant. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; n = 24–36 for each dose. (C and D) A dose-dependent response to the artificial stereoisomer (S,Z,Z)-CH503 is observed in all tested species. In species that produce CH503 (C), the artificial stereoisomer is more potent than the natural pheromone. In species that do not produce CH503 (D), the artificial stereoisomer and the natural pheromone are effective over a similar range of doses. Statistical analysis was performed as in A.(E) In courtship choice assays, D. ananassae and D. virilis males spent a greater amount of time courting nonperfumed females over (R,Z,Z)-CH503–per- fumed females. D. melanogaster males preferentially courted nonperfumed females over females perfumed with (S,Z,Z)-CH503. In the presence of both stereoisomers, males preferentially courted females perfumed with the natural pheromone. No effect on courtship was observed when females were per- fumed with an equivalent dose of a CH503 analog (TB-CH503). The ends of the box plot represent the spread of data between the 25th (bottom) and the 75th percentile (top), whereas the horizontal line represents the median and + represents the mean. Courtship vigor within each pairing was compared using a Wilcoxon rank-sum test. ****P < 0.0001; **P = 0.0032; *P = 0.0397; n = 21–29.

Adaptive Response to Sensory Exploitation. It is well established amount of the natural pheromone is needed to achieve the same that the biological activity of pheromones can be highly de- level of courtship inhibition produced by the artificial stereoiso- pendent on the stereostructure (21). To determine whether the mer. In contrast, males of D. ananassae, D. willistoni, D. virilis,and stereostructure of CH503 is important for its function as an D. mojavensis exhibited an equivalent or greater responsiveness to antiaphrodisiac, we tested the effect of a synthetic stereoisomer the wild-type form of CH503 compared with the artificial stereo- of CH503, (S,Z,Z)-CH503, a stereoisomer not known to be isomer (Fig. 2D, Fig. S2,andTable S2). Thus, species that produce expressed naturally by drosophilids (Fig. S2). Unexpectedly, the pheromone exhibit a weaker behavioral response to the nat- D. melanogaster, D. simulans, D. yakuba,andD. sechellia exhibited ural compound, whereas nonproducers of the pheromone are stronger courtship suppression and a longer latency to initiate more sensitive to the pheromone (Fig. 3). courtship in the presence of the artificial stereoisomer compared The lowered sensitivity exhibited by CH503-producing species with the natural pheromone at equivalent doses (Fig. 2C, Fig. S2, to the natural pheromone may be the result of sensory adapta- and Table S2). When D. melanogaster males were given a choice of tion or habituation from exposure to the pheromone from con- females perfumed with (S,Z,Z)-CH503 or a nonperfumed female, specifics. However, adaptation likely is not a contributing factor males preferred to court the latter (Fig. 2E). Even in the presence because males were individually housed during the late pupal of a female perfumed with an equivalent amount of the wild-type stage and remained isolated until testing at 4–5 d old. Further- stereoisomer, (R,Z,Z)-CH503, males continued to show a stronger more, the response of D. melanogaster males collectively raised in aversion to the unnatural pheromone (Fig. 2E). The disparity groups was not significantly different from that of individually between the strength of the natural vs. unnatural pheromone housed flies (Fig. S3). Thus, even in the presence of heightened is particularly striking in D. simulans, in which nearly 15 times the preexposure to CH503, the differential sensitivity exhibited by

3058 | www.pnas.org/cgi/doi/10.1073/pnas.1313615111 Ng et al. Downloaded by guest on September 29, 2021 D. melanogaster to the natural and artificial pheromone remained sensory trap induced sensory adaptation (14). Interestingly, we did unchanged. not observe sensory adaptation to cVA. Responsiveness to the Courtship suppression in the presence of (R/S,Z,Z)-CH503 courtship inhibition properties of the pheromone showed no cor- might simply be the result of aversion to a foreign chemical or relation with whether the species produced the pheromone (Table masking of endogenous female pheromones. To address this S3). The multiple functions of cVA as an aggregation cue (26) and possibility, males of all species were tested with females per- aphrodisiac for females (27) likely prevent males from completely fumed with an equivalent dose of modified (S,Z,Z)-CH503 or losing sensitivity to the pheromone despite the costs of courtship (R,Z,Z)-CH503 analogs bearing either two triple bonds (instead suppression. of two double bonds) or a single double bond only. In each case, The robust behavioral response of melanogaster subgroup no significant courtship suppression was observed (Tables S1 and species to (S,Z,Z)-CH503 suggests future opportunities for sen- S2; Fig. 2E). These findings indicate that first, melanogaster sory exploitation. A subtle change in stereochemistry due to al- subgroup males have evolved partial resistance to the antiaphro- lelic variations in genes underlying pheromone production might disiac effects of the natural pheromone, and second, the behavioral allow for expression of a stereoisomer variant that would cir- response is specific to the stereostructure of (R/S,Z,Z)-CH503. cumvent male adaptation. Behavioral experiments are consistent Our findings support the hypothesis that in Drosophila,an with this prediction: use of the unnatural pheromone results in antiaphrodisiac pheromone evolved as a result of sensory ex- significantly greater courtship aversion, even in the presence of ploitation. All tested species, regardless of their endogenous the natural pheromone. It will be interesting to consider whether pheromone expression, suppressed courtship in response to the persistent use of (S,Z,Z)-CH503 over the long term may con- natural pheromone, (R,Z,Z)-CH503. Because (R,Z,Z)-CH503 tribute to physiological desensitization and also result in dis- originates only in the melanogaster subgroup, response to the advantages for the population, such as lowered mating frequency, pheromone predates expression of the pheromone. In contrast to resulting in reduced offspring fitness (28). previous findings that identified attractive superstimulants, these Several mechanisms underlying the evolution of antiaphrodisiacs results show that preexisting sensory systems for highly aversive have been proposed, arguing that chemical diversity in pheromone cues also may be exploited and that sensory exploitation also takes structures arises as a result of male–female cooperation (23, 29), place between males. male–female sexual conflict (30), or male–male competition (31). In D. melanogaster, multiple pheromones, such as cVA and Our study indicates that sensory exploitation is another mechanism (7Z)-tricosene, and accessory gland proteins are used to sup- for the evolution of antiaphrodisiacs and traits used in male–male press female mating frequency (22–25). We speculate that initial competition. A similar phenomenon whereby synthetic nonnatural use of CH503 helped reinforce courtship inhibition from rival stereoisomers elicited a behavioral response stronger than that of males because, in contrast to other male-transferred cues, CH503 the natural pheromone was reported previously in the German persists on female cuticles for at least 10 d (18). The low vapor cockroach, indicating that sensory exploitation may underlie the pressure of the pheromone and its behavioral potency would place evolution of other pheromone systems (32). These findings strong selective pressure for males to become less sensitive to the contrast with an alternative mechanism shown in Nasonia wasps, in pheromone to benefit from increased mating opportunities. Our which novel pheromone compounds appear before a preexisting results are consistent with this prediction: members of the (R,Z,Z)- response (33). Currently, the chemosensory receptor(s) and un- EVOLUTION CH503–expressing melanogaster subgroup have adaptively evolved derlying neural circuits mediating CH503 detection are unknown. to become less sensitive to the natural pheromone. In this way, Once they are identified in D. melanogaster,itwillbepossibleto a trait that originated from sensory exploitation drives the evo- correlate evolution of the receptor(s) structure with behavioral lution of sensory pathways. A similar phenomenon occurs in responses of various species and in this way, to understand the un- Goodeidae splitfin fish, where a visual cue originating as a female derlying neurophysiological mechanisms. Furthermore, determining

(R,Z,Z)-CH503 CH503 producing species 0.20 non-CH503 producing species D. simulans Courtship suppression index: 0.14 to 0.19 0.39 to 0.44 0.64 to 0.69 CH503 0.20 to 0.25 0.45 to 0.50 0.70 to 0.75 0.26 to 0.32 0.51 to 0.56 0.76 to 0.82 D. sechellia 0.33 to 0.38 0.57 to 0.64 0.24

D. melanogaster 0.14

melanogaster group D. yakuba 0.34

Sophophora D. ananassae 0.76 willistoni group D. willistoni 0.54

repleta group D. mojavensis 0.63 Drosophila D. virilis virilis group 0.59

50 40 30 20 10 0 Divergence time (million years)

Fig. 3. Evolution of CH503 expression and the behavioral response to CH503. CH503 expression (left) and the behavioral response to natural CH503 (right) are mapped onto the Drosophila phylogeny using a linear parsimony model. The values at the branch termini (right phylogram) reflect the courtship sup- pression index, a measure of the strength of courtship suppression induced by the natural pheromone compared with the artificial stereoisomer. All tested species respond to both stereoisomers. However, species that produce CH503 respond more weakly to the natural pheromone. Hatch marks indicate the likely point of origin of each trait.

Ng et al. PNAS | February 25, 2014 | vol. 111 | no. 8 | 3059 Downloaded by guest on September 29, 2021 molecular markers for the relevant sensory pathways will provide Z,Z)-CH503 was performed by gently vortexing live females in a glass vial a way to map the evolutionary origins of predisposed sensory coated with 2–100 μg of CH503, as previously described (18, 38). The lowest biases and the shaping of pheromone structural specificity. dose tested was 83 ng, corresponding to the approximate amount expressed by males from laboratory Canton-S stocks (37). The same procedures were Materials and Methods used for perfuming control flies, except that the vials were coated with hexane alone and the solvent was allowed to evaporate before vortexing. Drosophila Stocks and Husbandry. Flies were obtained from the Bloomington Stock Centre and Ehime University Stock Center, and maintained on cornmeal Courtship assays were performed at 23.3 °C, 60% humidity. Perfumed × × agar food at 25 °C, 60% humidity, with a 12:12-h light:dark cycle. D. mojavensis female flies were decapitated and placed in 16 9-mm or 35 10-mm were wild-caught from Las Bocas, Sonora, Mexico (in March 2009) by W. J. Etges courtship chambers containing moistened filter paper to maintain humidity. (University of Arkansas) and raised on standard cornmeal agar food supple- Control trials with hexane-perfumed females were conducted in parallel mented with banana (∼110 g per 20 half-pint bottles) and cactus powder (∼2.3 with experimental trials. For D. sechellia and D. mojavensis, live females g per 20 half-pint bottles). Adult flies used in courtship assays were isolated at were necessary to induce male courtship behavior. Males were aspirated the pupal stage or collected within 4 h after pupal eclosion. Male virgins were into the chambers, and features of courtship behavior [latency to initiate raised individually in a test tube containing 2 mL of cornmeal agar food, courtship, orienting, wing extension and vibration, attempted , whereas females were kept in groups of 10 per vial. For grouped-male con- and copulation (for live females)] were scored for 30 min. Courtship initia- ditions, adult males were collected at the pupal stage and kept in groups of 10 tion percentage is defined as the number of trials in which males displayed per vial for 7 d. Flies were tested at the following ages, corresponding to sexual continuous courtship behavior for at least 1 min. For D. mojavensis, 20 s was maturity: D. melanogaster and D. simulans,4–7d;D. sechellia and D. yakuba,5– used as the threshold because of the rapid copulation time (usually between 7d;D. ananassae,7d;D. virilis and D. mojavensis,10d;andD. willistoni,15d. 1 and 2 min). Courtship latency is defined as the amount of time elapsed between aspiration of the male into the chamber and the first display of Pheromone Profiling of Drosophila Species Using UV-LDI MS Analysis. UV-LDI courtship behavior sustained for at least 1 min. When no courtship is ob- MS analysis and the procedures for preparing the flies were described in detail served during the 30-min trial, the maximum score of 1,800 s is given. previously (18). Measurements were performed on a Q-Star Elite (AB SCIEX) orthogonal time-of-flight mass spectrometer equipped with an intermediate Two-Choice Courtship Behavior Assays. A single CH503-perfumed female and λ= pressure oMALDI2 source and an N2 laser ( 337 nm, 40-Hz repetition rate, a single hexane-perfumed female were decapitated and placed 10–15 mm μ 200- m beam diameter, 3-ns pulse duration). Ions were generated in a buffer apart in a 35 × 10-mm courtship chamber. A socially naïve male then was gas environment by using 2 mbar of N2. Individual flies were attached to aspirated into the chamber, and courtship behavior was scored for 30 min. a coverslip with adhesive tape and mounted onto a custom-built sample plate. Courtship vigor was calculated by normalizing the amount of time spent During data acquisition, the anogenital region was irradiated for 30 s, corre- displaying courtship behaviors toward one target to the total time males sponding to 1,200 laser shots. For each species, 10–15 socially naïve male flies spent courting either target. Trials in which courtship lasted for less than 10 s were measured. Mass accuracy for the mass spectrometer was ∼20 ppm. More were not considered. details about instrumentation conditions and analytical parameters important for the chemical imaging of insects are provided in ref. 34. Statistical Analysis. Courtship percentages were compared using a Holm– Bonferroni corrected Fisher exact test. For courtship latencies, a Kruskall– Purification of CH503 from Drosophila Species. Purification of CH503 by Wallis test followed by a Wilcoxon rank sum test was applied. All analyses thin layer chromatography (TLC) was performed using the conditions de- scribed by Yew et al. (18). Approximately 800–1,000 males and females of were performed using GraphPad Prism 6. D. melanogaster, D. simulans, D. yakuba, D. sechellia,andD. ananassae were extracted with hexane (10 mL) for 20 min at room temperature. The extract Mapping of CH503 Expression and Behavioral Response onto the Phylogeny. Character mapping was performed with Mesquite 2.75 (39) using a linear par- was concentrated to dryness using a stream of N2. The residue was dissolved in hexane before separation by TLC. TLC separation was performed on glass- simony model. Presence or absence of CH503 expression was treated as an backed silica gel plates (10 × 10 cm, coated with 0.2 mm of silica gel 60; unordered character state. The behavioral response to natural CH503 was rep- Merck) using a running solvent consisting of hexane/diethyl ether/acetic acid resented in terms of a courtship suppression index. The index is a measure of the (66:33:1, each by volume). The major component of the isolated fraction was behavioral response elicited by the natural pheromone (R,Z,Z)-CH503 relative confirmed as CH503 by Direct Analysis in Real Time (DART) mass spectrom- to the artificial stereoisomer (S,Z,Z)-CH503, and was calculated as follows: etry. The DART ion source was operated in positive-ion mode with helium gas, with the gas heater set to 200 °C. The glow discharge needle potential RCP ð + Þ, was set to 3.5 kV. Electrode 1 was set to +150 V, and electrode 2 (grid) was RCP SCP set to +250 V. where RCP is the courtship initiation percentage for (R,Z,Z)-CH503 and SCP is the courtship initiation percentage for (S,Z,Z)-CH503 at the lowest or second HPLC Analysis. The separation of eight stereoisomers was determined using lowest effective dose. To determine whether there is significant phyloge- a previously described method of derivatization with (1R,2R)- or (1S,2S)-2- netic signal from the behavioral response to the natural pheromone, the (2,3-anthracenedicarboximido)cyclohexanecarboxylic acid and HPLC (20, 35). HPLC separation was performed using a Tosoh DP-8020 pump equipped with terminal taxa of the tree were shuffled 10,000 times. The likelihood of the = a Rheodyne 7125 sample injector, Jasco FP-920 fluorescence detector, and character distribution was greater than expected by chance alone (P 0.017), column heater (Cryocool CC100 II). Data analysis was performed with Chro- indicating that the courtship suppression index exhibits significant phyloge- matocorder 21 (System Instruments). netic signal. Drosophila phylogeny was reconstructed based on information from FlyBase (40, 41). Synthesis of CH503 Analogs. CH503 analogs (R)- and (S)-3-acetoxy-11,19- octacosadiyn-1-ol (triple-bond or TB-CH503), and (3R,11Z)- and (3S,11Z)-3- ACKNOWLEDGMENTS. We thank D. P. Araujo, Y. N. Chiang, J. Chin, acetoxy-11-octacosen-1-ol (11Z)-dihydro-CH503 were synthesized according J. Y. Chua, Q. L. Koh, W. C. Ng, A. Pirkl, M. Rozenbaum, K. Su, and K. J. Tan for technical support and helpful suggestions and W. J. Etges for D. mojavensis to published procedures (36). flies. Fly illustrations were drawn by J. Y. Chua. R. Meier, S. Pletcher, and K. Dreisewerd provided critical comments on the manuscript. J.Y.Y. is sup- Single-Fly Courtship Behavior Assay. (R,Z,Z)- and (S,Z,Z)-CH503 were synthe- ported by a Singapore National Research Foundation fellowship and a fellow- sized previously by Mori et al. (37). Perfuming of females with synthetic (R/S, ship from the Alexander von Humboldt research foundation.

1. Darwin C (1859) On the Origin of Species by Means of Natural Selection, or the 6. Hamilton WD, Zuk M (1982) Heritable true fitness and bright : A role for para- Preservation of Favoured Races in the Struggle for Life (John Murray, London). sites? Science 218(4570):384–387. 2. Darwin C (1871) The Descent of Man, and Selection in Relation to Sex (John Murray, 7. Andersson M (1986) Evolution of condition-dependent sex ornaments and mating prefer- London). ences—sexual selection based on viability differences. Evolution 40(4):804–816. 3. Fisher RA (1930) The Genetical Theory of Natural Selection (Clarendon, Oxford). 8. Ryan MJ (1998) Sexual selection, receiver biases, and the evolution of sex differences. 4. Kirkpatrick M (1982) Sexual selection and the evolution of female choice. Evolution Science 281(5385):1999–2003. 36(1):1–12. 9. Ryan MJ, Fox JH, Wilczynski W, Rand AS (1990) Sexual selection for sensory exploi- 5. Zahavi A (1975) Mate selection—a selection for a handicap. J Theor Biol 53(1):205–214. tation in the Physalaemus pustulosus. Nature 343(6253):66–67.

3060 | www.pnas.org/cgi/doi/10.1073/pnas.1313615111 Ng et al. Downloaded by guest on September 29, 2021 10. Basolo AL (1990) Female preference predates the evolution of the sword in swordtail 26. Bartelt RJ, Schaner AM, Jackson LL (1985) cis-Vaccenyl acetate as an aggregation fish. Science 250(4982):808–810. pheromone in Drosophila melanogaster. J Chem Ecol 11(12):1747–1756. 11. McClintock WJ, Uetz GW (1996) Female choice and pre-existing bias: Visual cues 27. Kurtovic A, Widmer A, Dickson BJ (2007) A single class of olfactory neurons mediates during courtship in two Schizocosa wolf spiders (Araneae: Lycosidae). Anim Behav behavioural responses to a Drosophila sex pheromone. Nature 446(7135):542–546. 52(1):167–181. 28. Priest NK, Galloway LF, Roach DA (2008) Mating frequency and inclusive fitness in 12. Jones IL, Hunter FM (1998) Heterospecific mating preferences for a feather ornament Drosophila melanogaster. Am Nat 171(1):10–21. in least auklets. Behav Ecol 9(2):187–192. 29. Ayasse M, Paxton RJ, Tengö J (2001) Mating behavior and chemical communication in 13. Burley NT, Symanski R (1998) “A taste for the beautiful”: Latent aesthetic mate the order Hymenoptera. Annu Rev Entomol 46:31–78. preferences for white crests in two species of Australian grassfinches. Am Nat 152(6): 30. Andersson J, Borg-Karlson AK, Wiklund C (2004) Sexual conflict and anti-aphrodisiac 792–802. titre in a polyandrous butterfly: Male ejaculate tailoring and absence of female – 14. Garcia CM, Ramirez E (2005) Evidence that sensory traps can evolve into honest sig- control. Proc Biol Sci 271(1550):1765 1770. nals. Nature 434(7032):501–505. 31. Estrada C, Schulz S, Yildizhan S, Gilbert LE (2011) Sexual selection drives the evolution – 15. Wyatt TD (2003) Pheromones and Behavior: Communication by Smell and of antiaphrodisiac pheromones in butterflies. Evolution 65(10):2843 2854. 32. Eliyahu D, Mori K, Takikawa H, Leal WS, Schal C (2004) Behavioral activity of ster- Taste (Cambridge Univ Press, Cambridge, UK). eoisomers and a new component of the contact sex pheromone of female German 16. Howard RW, Blomquist GJ (2005) Ecological, behavioral, and biochemical aspects of cockroach, Blattella germanica. J Chem Ecol 30(9):1839–1848. insect hydrocarbons. Annu Rev Entomol 50:371–393. 33. Niehuis O, et al. (2013) Behavioural and genetic analyses of Nasonia shed light on the 17. Gwynne DT (2008) Sexual conflict over nuptial gifts in insects. Annu Rev Entomol 53: evolution of sex pheromones. Nature 494(7437):345–348. 83–101. 34. Yew JY, Soltwisch J, Pirkl A, Dreisewerd K (2011) Direct laser desorption ionization of 18. Yew JY, et al. (2009) A new male sex pheromone and novel cuticular cues for chemical endogenous and exogenous compounds from insect cuticles: Practical and meth- communication in Drosophila. Curr Biol 19(15):1245–1254. odologic aspects. J Am Soc Mass Spectrom 22(7):1273–1284. 19. Brieger G, Butterworth FM (1970) Drosophila melanogaster: Identity of male lipid in 35. Imaizumi K, Terasima H, Akasaka K, Ohrui H (2003) Highly potent chiral labeling reproductive system. Science 167(3922):1262. reagents for the discrimination of chiral alcohols. Anal Sci 19(9):1243–1249. 20. Shikichi Y, et al. (2012) Pheromone synthesis. Part 250: Determination of the ster- 36. Shikichi Y, Shankar S, Yew JY, Mori K (2013) Synthesis and bioassay of the eight eostructure of CH503 a sex pheromone of male Drosophila melanogaster, as analogues of male pheromone CH503 (3-acetoxy-11,19-octacosadien-1-ol) of the fruit (3R,11Z,19Z)-3-acetoxy-11,19-octacosadien-1-ol by synthesis and chromatographic fly Drosophila melanogaster. Biosci Biotechnol Biochem 77(9):1931–1938. – analysis of its eight isomers. Tetrahedron 68(19):3750 3760. 37. Mori K, Shikichi Y, Shankar S, Yew JY (2010) Pheromone synthesis. Part 244: Synthesis 21. Mori K (2007) Significance of chirality in pheromone science. Bioorg Med Chem of the racemate and enantiomers of (11Z,19Z)-CH503 (3-acetoxy-11,19-octacosadien- – 15(24):7505 7523. 1-ol), a new sex pheromone of male Drosophila melanogaster to show its (S)-isomer 22. Ejima A, et al. (2007) Generalization of courtship learning in Drosophila is mediated and racemate as bioactive. Tetrahedron 66(35):7161–7168. by cis-vaccenyl acetate. Curr Biol 17(7):599–605. 38. Billeter JC, Atallah J, Krupp JJ, Millar JG, Levine JD (2009) Specialized cells tag sexual 23. Scott D (1986) Sexual mimicry regulates the attractiveness of mated Drosophila and species identity in Drosophila melanogaster. Nature 461(7266):987–991. melanogaster females. Proc Natl Acad Sci USA 83(21):8429–8433. 39. Maddison WP, Maddison DR (2011) Mesquite: A modular system for evolutionary 24. Ferveur JF, Sureau G (1996) Simultaneous influence on male courtship of stimulatory analysis (http://mesquiteproject.org), Version 2.75. and inhibitory pheromones produced by live sex-mosaic Drosophila melanogaster. 40. Marygold SJ, et al.; FlyBase consortium (2013) FlyBase: Improvements to the bibli- Proc Biol Sci 263(1373):967–973. ography. Nucleic Acids Res 41(Database issue, D1):D751–D757. 25. Chen PS, et al. (1988) A male accessory gland peptide that regulates reproductive 41. Powell JR (1997) Progress and Prospects in Evolutionary Biology: The Drosophila behavior of female D. melanogaster. Cell 54(3):291–298. Model (Oxford Univ Press, New York). EVOLUTION

Ng et al. PNAS | February 25, 2014 | vol. 111 | no. 8 | 3061 Downloaded by guest on September 29, 2021