Genetic Feminization of Pheromones and Its Behavioral Consequences in Drosophila Males

See allHide authors and affiliations

Science  06 Jun 1997:
Vol. 276, Issue 5318, pp. 1555-1558
DOI: 10.1126/science.276.5318.1555


Pheromones are intraspecific chemical signals important for mate attraction and discrimination. In the fruit fly Drosophila melanogaster, hydrocarbons on the cuticular surface of the animal are sexually dimorphic in both their occurrence and their effects: Female-specific molecules stimulate male sexual excitation, whereas the predominant male-specific molecule tends to inhibit male excitation. Complete feminization of the pheromone mixture produced by males was induced by targeted expression of the transformer gene in adult oenocytes (subcuticular abdominal cells) or by ubiquitous expression during early imaginal life. The resulting flies generally exhibited male heterosexual orientation but elicited homosexual courtship from other males.

In many animal species, sex- and species-specific bouquets of odors elicit subtle changes in potential sexual partners, which in turn may respond by appropriate behavior (1). In the fruit fly Drosophila, the stereotyped courtship behavior exhibited by male flies is induced largely by chemical cues, or pheromones, produced by his mate (2). These pheromones—the most abundant hydrocarbon molecules present on the fly cuticle (3)—are sensed principally by contact and are thought to play a crucial role in sexual isolation, tending to prevent interspecific mating (4, 5).

In D. melanogaster, pheromones are strikingly sexually dimorphic (6) and have very different effects on male courtship behavior (7, 8) (Table1). Female flies produce dienes (two double bonds) with 27 and 29 carbons [cis,cis-7,11-heptacosadiene (7,11HD) and cis,cis-7,11-nonacosadiene (7,11ND)]. A few tens of nanograms of both dienes together can elicit vigorous male precopulatory behavior (7, 8). Male flies synthesize monoenes (one double bond) with 23 and 25 carbons [cis-7-tricosene (7-T) and cis-7-pentacosene (7-P)]. 7-T can inhibit dose-dependent male excitation (8,9), whereas 7-P stimulates males of some strains (4, 7, 8).

Table 1

Effects of UAS-tra expression on 4-day-old males. For the production of pheromones, the percentage of 7-tricosene (%7-T), 7-pentacosene (%7-P), and 7,11-dienes (%7,11-heptacosadiene and %7,11-nonacosadiene pooled) were calculated from the total quantities of cuticular hydrocarbons (ΣHc). 7,11-Dienes were pooled because their respective contributions were approximately the same in all strains. Percentages (mean ± SE) were obtained by gas chromatography of extracts from 20 individual flies (21). Amounts of cis-vaccenyl acetate (cVA) were estimated from ΣHc: (++) male-like, (+) reduced, (0) absence. For behavioral tests, PGAL4 UAS-tra males were examined both as objects with courting males of the 55B-GAL4 strain and as subjects with Canton-S (Cs) male or with shibire (shi) female objects (34). All flies were 4 days old, and the target flies were decapitated before the 10-min experiment. Decapitation prevents reciprocal courtship and allows measurement of unidirectional behavior. The percentage of courting males only includes males that courted for more than 20 s. The courtship index is the mean fraction of time (±SE in parentheses) spent actively courting by all males (wing vibration, licking, and attempt to copulate) (8), with at least 20 trials per strain.

View this table:

One of the few genetic factors known to control the production of sex pheromones in D. melanogaster (10,11) is the gene transformer (tra), which controls the sexual dimorphism of pheromones (8,12) as part of its larger influence on somatic sex determination. When the feminizing transgene UAS-tra, made with the female cDNA of the tra gene, is expressed in certain regions of the male brain, the male exhibits a bisexual orientation (13, 14). The tra gene also affects downstream sex-determination genes likefruitless and doublesex, which in turn control the sex pheromones or the male sexual orientation (15). Here, we expressed the UAS-tra transgene at different stages of development and in a particular group of abdominal cells, with the aim of producing a male fly with an unaltered sexual orientation, but with a female pheromonal profile.

To assess the critical period during which the tra gene product regulates pheromone expression, we transiently expressedUAS-tra throughout the organism at different developmental stages by crossing it to a line in which GAL4 is fused to aheat shock 70 promoter (16). The tragene, fused to a promoter containing a GAL4-dependent upstream activation sequence (UAS), was therefore expressed with the same temporal pattern as GAL4(17). Heat shock induced ubiquitous traexpression at different developmental stages from embryo to 4-day-old adults (Fig. 1). The extent of feminization of pheromone production (the replacement of 7-monoenes by 7,11-dienes) reached a peak when UAS-tra expression was induced by a single heat shock between 12 and 48 hours of adult life. No pheromonal feminization was observed with control males expressing UAS-lacZ under the same heat shock conditions. This result suggests that the gene product or products being synthesized in these flies, after a 2-hour heat shock, have a sufficiently long-lasting effect to enable the production of female pheromones up to 4 days later and confirms that early imaginal life is the critical period during which sexually dimorphic hydrocarbons replace immature hydrocarbons on the fly cuticle (18).

Figure 1

Production of sex pheromones in 4-day-old male flies as a function of temporal activation ofUAS-tra or of UAS-lacZ. A single pulse of heat shock (37°C) was applied for 2 hours, at various times (or 6 hours before pupariation). Each data point represents the mean percentage (±SE) of 7-monoenes (%7-T + %7-P) and of 7,11 dienes (%7,11-HD + %7,11-ND) for 20 hsp-GAL4 UAS-tra individuals and for 10 hsp-GAL4 UAS-lacZindividuals. Control, non–heat-shocked hsp-GAL4 UAS-traand hsp-GAL4 UAS-lacZ males yielded 52.8 ± 1.5 and 57.5 ± 2.3% 7-monoenes, and 0.9 ± 0.5 and 0% 7,11 dienes, respectively. Values were measured as in (21).

Mosaic studies have localized the origin of pheromonal sexual dimorphism in the fly abdomen (19). To precisely map the cells that control the production of sex pheromones, we generatedPGAL4 UAS-tra strains in which males show different patterns of regional feminization in their abdomen. The PGAL4 system uses enhancer detection to express the GAL4 transcriptional activator in different cellular patterns (16,20). The feminizing UAS-tra gene is therefore expressed with the same tissue specificity as GAL4(14).

Out of 50 PGAL4 UAS-tra lines originally screened, we identified five lines (A through E) in which male flies exhibited a female pattern of pheromones (Table 1). These regionally feminized flies, chromosomally XY, produced high amounts of female dienes (7,11HD and 7,11ND) and low amounts of male monoenes (7-T and 7-P). The UAS-tra expression was responsible for the feminization of sex pheromones because neither PGAL4 UAS-lacZ males nor PGAL4 UAS-tra females from these five PGAL4 strains showed any substantial variation of their male or female pheromonal pattern (21).

We examined the pattern of GAL4 expression in the five feminized strains (A through E) to seek a relation between their expression patterns and pheromonal feminization. The GAL4expression patterns were revealed by a cross to a UAS-lacZreporter strain (Fig. 2). The adult expression patterns were of varying complexity, but they overlapped in two cell types: the oenocytes and the midgut (22). Oenocytes are subcuticular abdominal cells found in segmentally repeated rows that form crescent-shaped strands on the tergites and small clusters on the sternites (23). Oenocytes were the only cells able to change the production of pheromones because males of the other PGAL4 UAS-tra strains that were not feminized for their pheromones (Table 1) often showed strong expression in the midgut but not in the adult oenocytes (Fig. 2, F and G). A correlation between oenocyte expression and pheromone feminization was confirmed by analysis of a larger number of PGAL4 lines (24). Together with previous studies of such unrelated insects as the desert locust (25) and the mosquito Culicoı̈des nubeculosus (26), this result suggests that pheromones may be synthesized in the oenocytes of many insect species (27).

Figure 2

Photomicrographs showinglacZ expression pattern (blue) in the abdominal oenocytes of various PGAL4 strains. All flies are 4-day-old males. (A) Sagittal frozen section of the thorax and abdomen, and (B and D to G) horizontal frozen sections (10 μm) in the abdomen of PGAL4 UAS-lacZ males. (C, H, and I) Dorsal views of the abdominal cuticule of PGAL4 UAS-lacZ males (C and H) andPGAL4 UAS-tra; UAS-lacZ males (I). (A to C) Strain C; (D) strain B; (E) strain E; (F) strain F; (G) strain G; and (H and I) strain D. Arrowheads (A and E) indicate the oenocytes. Bars, 50 μm (D, E, F, and G; H and I have the same magnification).

Oenocytes have multiple endocrine functions, including the regulation of ecdysteroids (28), one of which, 20-OH-ecdysone, controls an elongase required for the synthesis of 23 and 27 C hydrocarbons in Musca domestica (29). In D. melanogaster, according to the biosynthetic scheme proposed by Jallon (6), an elongase, perhaps coupled with a desaturase, would be sufficient to replace 7-monoenes by 7,11-dienes. In the mutant Drosophila ecdysoneless 1 tsfemales, 7,11-dienes are to a large extent replaced by 7-monoenes (30).

Drosophila adult oenocytes show a slight sexual dimorphism (23, 31), but this does not seem to underlie the pheromonal difference between the sexes. To visualize directly whether the ectopic feminization of the oenocytes by the tra gene could have changed their sex-specific pattern, we simultaneously expressed both UAS-tra and UAS-lacZ transgenes (32). Resulting XY flies (PGAL4 UAS-tra UAS-lacZ) did not differ in their segmental pattern oflacZ expression, as compared with PGAL4 UAS-lacZmales (Fig. 2), nor in their production of sex pheromone, as compared with XY PGAL4 UAS-tra flies (33).

The sex pheromones produced by the feminized XY flies from the five strains (A through E) functioned as female pheromones and elicited a more vigorous courtship response in control males than in males from F-tra, G-tra, and control strains (34). The variation in these male courtship responses may reflect variability in controlling signals other than female pheromones such as the chemicals 7-T, 7-P, and cis-vaccenyl-acetate (cVA) (35) and visual cues like the abdominal and genital morphology of target PGAL4 UAS-tra males, the phenotypes of which seem to be independent of oenocyte feminization (36).

When tested as subjects against control male and female flies, feminized males from C-tra, D-tra, and E-tra strains retained a strong and typical male heterosexual behavior (Table 1), suggesting no relation between the feminization of their hydrocarbons and their sexual orientation. However, A- and B-tra males exhibited some bisexual behavior, possibly because they were feminized in the calyces of their mushroom bodies (strain A) and in a dorso-medial subset of their antennal lobes (strains A and B). However, these two brain structures, which function in mate recognition (13, 14), were not feminized in the other GAL4-tra strains (strains C through G), showing that they are not required for feminization of the pheromonal profile.

Our analysis shows that in D. melanogaster, two aspects of individual sexual identity—the perception of others and the presentation of self to others—are under separate genetic and anatomical control. Homosexual courtship may take place either because of factors in the courter's brain (13,14) or because of factors in the courted fly's pheromonal profile. The interactive aspect of courtship and the complex nature of sexual identity in an animal as relatively simple as the fruitfly indicate that simplistic explanations of the genetic bases of sexuality are unlikely to be true.

  • * To whom correspondence should be addressed. E-mail: ferveur{at}

  • Present address: The Neurosciences Institute, 10640 John Jay Hopkins Drive, San Diego, CA 92121, USA.


View Abstract

Navigate This Article