Copyright 0 1993 by the Genetics Society of America

Genetic Control of Pheromones inDrosophila simulans. 11. kete, a Locus on the X Chromosome

Jean-FranGois Ferveur and Jean-MarcJallon Laboratoire de Biologie et Ginitique Evolutives, CNRS, 91 198 Gij-sur-Yvette,France Manuscript received June 17, 1992 Accepted for publication October 28, 1992

ABSTRACT The production of cuticular hydrocarbons, including contact pheromones, is under polygenic control. To investigate X-linked loci, EMS mutations were induced in Drosophila simulans . A mutant strain was discovered which in both sexes showa reduction in the biosynthesis of both 7-tricosene (7-T)the species contact pheromoneand all other linear hydrocarbons. The locus controlling this effect, kiti, is recessive and was localized to I, 18.5. Unlike a previously identified gene on the second chromosome of thisspecies, Ngbo, kiti does not affect theratio of 7-T:7- pentacosene (7-P). Other reproductive characteristics are also affected, including egg-hatching. However, courtship behaviors in both sexes appear normal.

-TRICOSENE is the main hydrocarbon on the MATERIALS AND METHODS 7 cuticle of male and female flies in most strains of Drosophila stocks and crosses: All strains of D. simulans Drosophila simulans and of Drosophilamelanogaster were kept on standard cornmeal and yeast medium under a males of many strains (LUYTEN1982; PECHINEet al. 12: 12-hr 1ight:dark cycle at 25 O. Three strains were used: 1985).This substance has been shown to act as a (1) Seychelles,based on several inseminated femalescol- pheromone, able to inducewing display in D. simulans lected in the Seychelles archipelago in 1981; (2) ywmJ a strain carrying four recessive markers: yellow (I, O.O), white males UALLON 1984). The cuticle of D. simulans flies (I, 1.5), miniature (I, 36.1) and forked (I, 56.7) (a gift of J. contains at least 13 hydrocarbons which vary quanti- COYNE); and(3) C(Z)RM, a strain with attached X chromo- tatively according to sex and strain. These molecules somes carrying yellow and white (a gift of T. WATANABE). have 23, 25, 27 or 29 carbons and are linear hydro- For regular crosses, five pairs of virgin4-day-old flies carbons or 2-methyl-branched alkanes. The predom- were placed in vials with food. Progeny from these crosses inant linear hydrocarbons are monoenes with a single were analyzed continuously throughout the course of the experiments reported here. For backcross experiments, is- double bond in position 7: 7-tricosene (7-T, 23 car- ofemale lines were set up from females randomly collected bons) and 7-pentacosene(7-P, 25 carbons) UALLON over 14 generations and inaividually mated when 4- 14 days 1984; PECHINEet al. 1985). Apart from7-T, only one old with4-day-old males. In general, crossesset up for other D.simulans hydrocarbon, 7-P, is currently sus- studying the inheritance of hydrocarbon phenotype were pected of playing a role in mate recognition or stim- studied simultaneously. Mutagenesis: Forty Seychelles males, 12-24 hr old, were ulation in this species, perhaps via a synergetic inter- fed overnight on saccharose supplemented with 0.1% EMS action (COBBand JALLON 1990). (LEWISand BACHER1968). These males were then mass- In a previous article we reported the existence of mated with 50 virgin C(Z)RM females. This procedure was Ngbo, a majorsecond chromosome locus in D. simulans replicated 10 times. Six hundred and fifty F1 males were which controls the ratio of 7-T:7-P by reciprocally produced and were individually backcrossed to four virgin C(Z)RM females, producing F2males sharing the same X changing the levels of both hydrocarbons (FERVEUR chromosome. Each line was kept inbred, except in the case 1991). We also showed that females of most D. simu- oflow fertility, where maleswere backcrossed to virgin lans strains have higher levels of 7-T than their hom- C(Z)RM females. Cuticular hydrocarbon profiles were com- otypic males. This and other data suggest that genes pared for males from 295 lines. on the X chromosome may be involved (BENAMAR and Extraction and analysis of cuticular hydrocarbons: Cu- ticular hydrocarbons were extracted from individual flies JALLON 1983;FERVEUR, COBB and JALLON 1989). with the method described in FERVEUR (1991),except for SCOTTand RICHMOND(1988) argued that theX chro- mutagenized lines, for which four males from each line were mosome could modify levels of 7-T and 7-P in D. pooled at generations FP and Fs. This “pooling” procedure melanogaster males, although it is possible that genetic enabled us to exclude intraline hydrocarbon variability control of the 7-T:7-P ratio is different in the two which was not controlled by the X chromosome. Individuals species (FERVEUR1991). To further investigate di- from lines showing X-linked hydrocarbon variation were subsequently tested separately. rectly the role of the X chromosome on pheromone Hydrocarbon parameters: Thirteen hydrocarbons were production we have used EMS mutagenesis to induce detected ingas chromatography (GC) profiles and their X-linked mutations in D. simulans. quantities were measured. For the 295 mutagenized lines

Genetics 133: 561-567 (March, 1993) 562 andJ.-F. Ferveur J.-M. Jallon

the percentage of each hydrocarbon was measured relative ratio and the overall production of branched com- to the sum of all hydrocarbons (CHc) for that line and pounds (CBr) were not affected in strain 430 males. compared to the percentages obtained from all lines pooled. The analysis of putative mutant lines was carried out using To localize the character(s) involved, cuticular hy- percentages and absolute amounts (Q) of hydrocarbons, in drocarbons were analyzed from crosses between strain particular of the two main alkenes (7-T and 7-P) and two 430and pmf flies, which carrya multimarked X methyl branched hydrocarbonswith 27 and 29 carbons (27 chromosome (Table 2). pmf males and reciprocal F1 Br and 29 Br). Levels of other linear and branched hydro- carbons which showed lower absolute levels were separately males sharing the ywmf X chromosome all showed pooled (Elin and ZBr). The ratios of 7-T:7-P and 7-T:27 much higher levels of 7-monoenes thanstrain 430 Br were also measured. males. The Q7-T value was approximately six times Statistical tests: For mutagenesis screen each hydrocar- higher in males with a ywmf X chromosome than in bon parameter was compared to the mean f 2 SD calculated strain 430 males. from all the 295 mutagenized strains(P < 0.05). When one hydrocarbon parameter was compared between two inde- However, ywmf males produced about 30% more7- pendent samples, we used Student’s t test. For comparisons T than Seychelles flies while their 7-T:27 Br ratio was between two or more samples differing for two conditions very close (6.04 and 6.6,respectively). For this reason, (genotype and/or age and/or mating status), we used a two and because males with a pmf X chromosome show a factors ANOVA. 7-T:27 Br ratiofour times higherthan strain 430 Behavioral and fertility tests: Virgin pairs of flies were observed under a watch glass for 2 hr. Courtshipand males (Table2), we mapped the locus causing the copulation latencies and durations were notedeach for pair. hydrocarbon defect with a 7-T:27 Br ratio. Mated females were individually transferred to a vial con- Recombinant males from backcrosses and F2 strains taining food colored with neutral red (0.05%; to enhance were grouped according to their morphological phe- background contrast) and were allowed to oviposit for 24 notypes and were characterized by the two predomi- hr before being discarded. At least 24 hr later the number of unhatched eggswas noted. The sex ratio of the resulting nant hydrocarbons Q7-T and Q27 Br and by their imagos was also recorded. The hydrocarbon profile of all ratio (7-T:27 Br) (Table 3).The 7-T:27 Br parameter flies was subsequently analyzed. gave the most consistent result. The factor controlling this parameter appears to segregate with white and RESULTS miniature (6.42 & 0.14 for w-m- as against 2.57 +- 0.1 1 for w+m+). On thebasis of these distributions, an Genetic control of hydrocarbon variation: Males from 295 EMS-mutagenized lines were screened for empirical cutoff point was established at 4.2, separat- their cuticular hydrocarbon profile. One strain (430), ing the two “7-T:27 Br” phenotypes with a low mis- showed a significant quantitative variation: these flies classification probability (P = 0.056) (Figure 1). With showed the same number of GC peaks as wild-type this classification criterion, the hydrocarbon profileof individuals, with similar retention times, but the pro- males showing only one of the two markers was stud- portions of 7-T and of 7-P (41.6% and 2.5%, respec- ied. Overall, males expressing either w-or m- showed tively) in males were lower than for all 295 mutagen- intermediate values; recombination between the two markersand the two 7-T:27 Br values (< 4.2 <) ized lines pooled (59.8 f 6.5%and 5.4 & 2.8%, respectively). Conversely, strain430 males showed enabled us to map a single locus or a tightly linked much higher proportions of both methyl branched group ofloci on the X chromosome at 18.5 f 1.8, hydrocarbons 27 Br and 29 Br (25.8% and 9%, re- which controls the ratio of7-T:27 Br. We have called spectively) as compared tomales from all mutagenized this locus kite‘, a word meaning “small quantity” in a strains pooled (9.0 & 3.4% and 3.8 f 2.1%). Thus West African dialect, following the mutant phenotype strain 430 males have a much lower ratio of 7-T:27 which shows low 7-T:27 Br value. Br than the other mutagenized strains (1.6 1and 6.65 kiti is the main factor involved in the control of 7- +- 0.20, respectively). T level. F2 recombinant kiti+ males showed signifi- To assess whether the main hydrocarbon variation cantly much larger amounts of 7-T than F2 recombi- was linked to the X chromosome or not, males from nant kiti’ males (1478 2 37 ng and 606 +- 21 ng, the strain430 were crossed with females from the respectively; F = 428.4; d.f. = 1, 458; P < 0.001). strain C(Z)RM. Absolute amounts of hydrocarbons Q27 Br does not seem to be affected by kiti in F2 were compared in Seychelles, strain 430 and C(I)RM males (239 & 6 ng and 226 & 5 ng, respectively, for flies (Table 1). Strain 430 males produced approxi- kiti’ and kiti+; F = 1.82; d.f. = 1, 458; P > 0.05). mately 30% of the total amount of hydrocarbons A similar effect of kiti on the linear hydrocarbon (CHc) shown by Seychelles males. There was a similar production was also shown in females (see next sec- decrease in the amount of 7-T and 7-P (Q7-T and tion). Reciprocal female hybrids between strain 430 Q7-P) and of all other linear compounds (Clin). Strain and ywmf strain showed no significant difference in 430 females carrying C(Z)RM attached X chromosomes Q7-T, norwere they significantly different frompmf showed no such effect, suggesting that an X-linked females, showing that the characteris recessive (Table factor is implicated in this difference. The 7-T:7-P 4a). Genetic Control of Pheromones 563

TABLE 1 Ratio and absolute amounts of certain cuticular hydrocarbon in various strains of D. simulans

Seychelles Male Seychelles 97018 f 65 125 f 12 265 f 18 285 f 15 1780 f 2.05120 Female 15 1300f 110 130f 18 3255 24 290f20 2300f 148 2.30 Strain 430 Male 17 300 f 36 35 f 4 85 f 7 265 f 18 670 -C 28 2.15 Female 24 2250 f 102 120 f 8 455 f 28 350f 16 3450 f 172 2.93 C(I)HM Female 178939 f 113 108 f 7 389 f 24 342 f 15 2811 f 158 2.81

Unless otherwise specified, hydrocarbons were extracted from individual 4-5-day-old flies. Data shownare mean (~sE)of absolute amounts of 7-tricosene (Q7-T), 7-pentacosene (Q7-P), the sum of linear alkanes with 23 C, 25 C, and 27 C (ZLin), the sum of branched (-C2-) compounds with 27 C and 29 C (CBr) and the sum of all hydrocarbons (CHc); 7-T:7-P ratio was logarithmically transformed. EMS-treated Seychelles males mass-crossed withC(I)RM females produced 650 F1 males which were individually crossed with C(I)RM females. Out of 295 lines inbred from the F2 generation, the strain 430 revealed significant hydrocarbon variations.

TABLE 2 Hydrocarbon productionin male flies resulting from crosses between the strain430 and the ywmf multimarked strain

Cross Q7-P Q7-T 427 Br Q29 Br CHc 7-T:7-P (female(ng) X male)(ng) (ng) N (ng) (ng) (log e) 7-T:27 Br

(FO) 73 283f 15 18f 2 159k6 85 f 12 650 f 26 2.85 f 0.08 1.74 f 0.06 strain 430 X strain 430 (12.9)(25.4)(2.9) (42.2) (Fd 33 951389f 128 f 9 233 f 15169 f 112485 f 1652.40 f 0.02 6.04f 0.22 yumf x yumf (9.6) (5.1) (55.8) (7.0) (FI) 20 1823 f 89 185f 10 271 f 11 168f 11 3067 f 125 2.30f0.056.77f 0.28 yumfX strain 430(5.5) (9.0) (6.1) (59.2) (FI) 22 732143k 187 f6 2782 12 193f93600f 120 2.44f0.037.88f0.32 strain 430 X yumf (5.4)(7.8) (5.2) (59.5)

Data shown are mean of absolute amounts (ng) 2 SE of 7-tricosene (7-T), 7-pentacosene (7-P), 2-methyl-hexacosane (27 Br) and 2- methyloctacosane (29 Br). The percentage of each variable relatively to the sum of all hydrocarbons (CHc) is shown in parentheses. The 7- T:7-P ratio, logarithmically transformed, and the 7-T:27 Br ratio are also shown.

Reproductive and behavioral effects of Rete: In TABLE 3 order to obtain homozygous kiti' females backcrosses Production of major hydrocarbons in recombinant males with were carried out. Hybrid females produced by cross- different morphological phenotypes ing strain 430 (kiti') with pmf(kiti+)flies were indi- vidually backcrossed to strain 430 (kiti') males over Morphological Q7-T 427 Br phenotype N (nn) (nn) 7-T:27 Br 14 subsequent generations. After 14 backcross gen- erations theprobability of a single female still contain- [YW"'fl 52 1851 f 78 298 f 10 6.26 f 0.20 [-wmfl 8 2005f 156 316f 15 6.37f0.42 ing the kite'+ allele should have been smaller than (1/ [ywm-] 33 1288f61 198f 11 6.69f0.23 2)14.However, we found that amongour samples both [-wm-] 5 1085f 156 168f 23 6.43 f0.31 the kite'+ and the kiti' alleles were still present: Q7-T [YW-fl 231417 f 111259 f 19 5.95 f 0.58 and 7-T:27 Br values were measured in virgin and in [-w--] 6 998 f 20982 f 23 5.17 f 0.84 mated females (Table 4b) and could be related to the [yw--1 100 794f 35 180f 5 4.67f0.22 production of viable offspring (Figure 2). There were ["fl 72 1339f77 301 f 12 4.68f0.28 [--m-] 30 928 f 76 202 f 11 4.69 f 0.36 notable differences between females producing viable [Y-nlfl 4 1056f 254 239f444.65f 1.20 offspring (fecund females) and sterile females (Table [Y---I 19 512f54 179f 11 2.82f0.21 4c). Fecund females could be clearly distinguished by [---fI 35 809f58 303f 17 2.71 f0.14 their 7-T:27 Br values which were always greater than LY-4 3 835f 239 296f 30 2.77 f 0.58 the previously established empirical cutoff point of 72 442+34 [+I 186f6 2.43f0.18 4.2 for recombinantmales (see Figure 1). The bimodal For parameters, referto Table2. pmffemales were crossed with distributions of 7-T:27 Br suggest that female flies strain 430 males; resulting F1 females were either crossed with F1 males or backcrossed to strain 430 males producing two progenies with 7-T:27 Br values (4.2 may be homozygous for (Fp and backcross). Males from these crosses showed no significant kiti' while other individuals are heterozygous with a difference in hydrocarbon production, were further pooled. Males were grouped according to their morphological phenotype: yellow k&+ phenotype. We still had a 1: 1 ratio (estimated (y), white (w), miniature (m) and forked (f). with a x* goodness of fit test) of the two alleles in the samples of five different generations, therefore there 564 J.-F. Ferveur and J.-M. Jallon

must exist a very strong selection against homozygos- ity for kiti’. These data strongly suggest that kiti’ homozygous females are sterile. If the sterility mapped to a different locus on the X chromosome, recombi- nation between the “sterile” locus and kiti should have occurred in the backcrosses. However, in more than w+m+ 200 female flies tested a sterility phenotype and low 7-T:27 Br ratiowere correlated without exception (see below). It therefore appears that kite‘’ homozy- nnfL 0 2 gous females are sterile. Female flies, subsequentlycharacterized by their hydrocarbon phenotypes as kiti’ homo- or heterozy- gotes, were individually crossed for six hours with kite” males. No significant differences were found for the number of courtships induced, courtship latency or copulation levels in a two hour period. The pro- portion of females laying eggs and the numberof eggs laid per female also showed no significant differences between the two genotypes. Females did, however, 10 12 14 2 4 8 i ln“- differ for the proportion of eggs hatching: 87% of z eggs laid by heterozygous females hatched, while no 16 eggs laid by homozygotes were observed to hatch. Hydrocarbon levels were compared in kiti’lkiti’ and kitt?/kite” females of different ages and reproduc- tive states (Figure3). Between 4 and 10 days old, heterozygous females showed a generalized increase in Q7-T (from 1716 k 78 ng to 2532 f 86 ng; t = 8.74, d.f.= 207, P < 0.001) and in Q27 Br (from 187 k 7 to 299 f 11 ng; t = 12.47, d.f. = 207, P 0.001) 1 6. with a stabilization for both Q7-T and Q27 Br be- !L2 4 14 6 128 10 tween 10and 14 days old. Homozygous females showed a similar increase of (227 Br but their Q7-T

1 4. remained always low (729 f 34 ng and 645f 137 ng

12. at 4 and 14 days old, respectively).

1 0. Moreover,a significant effect of mating was ob-

8. served on Q7-T but not on Q27 Br (see Table 4b).

0 The postmating decrease of Q7-T was only observed

4 in heterozygous females of age 4 and 10 days.

2

0 DISCUSSION 0 2 4 14 6 128 10 7-T : 27 Br The cuticle of male and female D.simulans flies contains long-chain hydrocarbons(23-29 carbons) FIGURE1 .-Frequency distribution of 7-T:27 Br values for re- branched (mainly with methyl on carbon 2) or linear combinant male flies showing combinations of the white (w) and miniature (m) markers of their X chromosome. pmf(X chromosome (saturated or with one double bondmainly in position multimarker) females were crossed with strain 430 males showing 7; JALLON 1984; PECHINEet al. 1985). There is no a mutant hydrocarbon phenotype. F, females were crossed with F1 qualitative sexual dimorphism but females of most males (Fpprogeny) and with strain 430 males (backcross progeny). strains show higher levelsof 7-tricosene, a species Both male progenies, which didnot significantly differfor this contact pheromone, than males. Our previous studies hydrocarbon parameter, were subsequently pooled.Dotted lines indicate the empirical cutoff point (4.2)between mutant and wild- have suggested that variations in 7-T levelin this type hydrocarbon phenotypes(left and right, respectively). This species are under polygenic X-linked and autosomal value produces a misclassification probability of 0.056 (calculated control (FERVEUR,COBB and JALLON 1989). Using from both groups w+m+ and w-m-). Out of 129 w”m+ flies, 67 were natural variants FERVEUR (1991)showed that an au- kiti’ (P = 0.48). Out of 106 w+m- flies, 49 were kiti’ (P = 0.54). tosomal locus Ngbo (11, 65.3) controls the main poly- Based on the position of w (1.5) and m (36.1), kiti was mapped at 18.5 f 1.8 (the standard error is derived from the misclassification morphic variation between the two major hydrocar- probability). bons 7-T and 7-P, in D. simulans. In this study, we report the discovery of an EMS- Genetic Control of Pheromones 565 TABLE 4 Inheritance and matingeffects of keti on hydrocarbon productions in 4-day-old femaleflies

~~ ~ ~ a. Fo pmfx rwmf Virgin 9 1325f 191 159f 18 2478f219 8.28f 0.93 F1 pmfx strain 430 Virgin 18 1374f 75 159+5 2117 f 107 8.68f0.37 FI strain 430 X pmf Virgin 15535 f 151 118 f 7 2319 f 172 13.302 1.47

b. BackcrossF4 Virgin 1334k8980 221 f8 2198f 125 6.35f0.41 Backcross F4 Mated 981 105 f43 231 f9 17162634.85 f 0.29 t test 0.753.86***d.f. 183 NS 3.14**3.70***

C. Backcross Fq Mated and fecund 42 1329f52 190f8 2179f76 7.29 f 0.31 Backcross F4 Mated and nonfecund 760f41 63 258 f 13 1408f69 3.22 f 0.26 t test 10.09***7.41***3.93***8.78*** d.f. 103 Data given are mean absolute amount (~sE)and ratio of 7-T:27 Br; for symbols, refer to Tables 1 and 2. pmfflies were reciprocally crossed with strain 430 flies (see Table 1). F1 females resulting from pmffemales X strain 430 males cross were individually backcrossed to strain 430 males; resulting females (backcross F1) and those of subsequent backcross generations were individually backcrossed to strain 430 males. Females analyzed here belong to the fourth backcross generation. Females were mated when 3 days old. Hydrocarbon productions were compared with a Student's t-test; Levels of significance: NS = not significant, * P < 0.05, ** p < 0.01, *** p < 0.001.

I In

Q 7-T 7-T : 27 Br FIGURE2.-Frequency distribution of Q7-T, 7-T:27 Br for virgin (a) or mated (b) female flies. Fecund flies producing viable progeny are shaded. Females were collected at backcross F4 generation (see Table 4). Kiti genotypes were defined according to their 7-T:27 Br values (homozygote < 4.2 < heterozygote). induced mutation on theX chromosome which affects gous kite" backcross females as compared to hetero- the productionof linear hydrocarbons.This mutation, zygous females. kiti', mapped at 18.5 f 1.8, produces a similar de- kiti seems to affect not only the production of Q7- crease of the amount of the pheromone 7-T in flies T but also that of Q7-P (-72% from Seychelles to of both sexes and of different genetic backgrounds: strain 430 males), without significantly affecting their strain 430 males show a 69% reduction as compared ratio. The production of all otherdetected linear to Seychelles males; kiti' hemizygotes show a 59% hydrocarbons is also reduced. Except in4-day-old reduction in F2 recombinant males as compared to females, no significant effect of kiti was found on the kiti+ males, and there is a 58% reduction in homozy- production of branched hydrocarbons (27 Br). 566 J.-F. Ferveur and J.-M. Jallon expressed with two known alleles. The production of 7-tricosene is dependent on Ngbo and on other auto- Q) E somal and X chromosome loci (FERVEUR,COBB and Q) VI JALLON 1989). We have previously suggested that 0 .: .: 2000 different alleles of Ngbo might changethe relationship L r:c between the speed of either elongation or decarbox- Y ylation during hydrocarbon biosynthesis (FERVEUR -1L I // 199 1). In most , including Drosophila, linear and branched hydrocarbons seem to be derived by decar- boxylation from long-chain fatty acids resulting from

oly...... , . . . . i repeated elongations of medium size fatty acids (C 12- Cl8) (BLOMQUIST,DILLWITH and ADAMS1987; PEN- NANEC’H et al. 1991). The enzymes involved in ho- mologous elongations and decarboxylations might be more orless specific for linear or branched substrates. Moreover, the methyl-branched group seems to be introduced early in chain synthesis. Several fatty acid synthetases (FAS) display various substrates specifici- ties (especially for linearand branchedchain initiators) and tissue specificities (BLOMQUIST,DILLWITH and ADAMS1987; JUAREZ,CHASE and BLOMQUIST1992). mL In Drosophila a soluble FAS has been characterized PC rv and semipurified. It is also able to produce shorter medium size fatty acids (C12-C14) as well as the U 8) 0 4 10 14 days longer ones(C 16-C 1 but cannotuse methylmalonyl CoA, the main chain initiatorleading tobranched FIGURE3.-Effect of kiti on hydrocarbon ontogeny in females. (DE Females of different backcross generations were pooled according fatty acids, alone as a substrate RENOBALES,WO- to age and mating status. Homozygous (circles) and heterozygous PODIN and BLOMQUIST1986). We suggest that kiti females (squares) were analyzed separately depending on their might act on theearly biosynthetic steps especially the mating status (virgin = empty; mated = filled) for 7-tricosene fatty acid synthesis of linear compounds. production. Virgin and mated females were pooled for the analysis De novo hydrocarbon biosynthesis is greatest be- of 2-methylhexacosane. tween 2 and 4 days in D. melanogaster flies of both In order to map kiti in males, flies carrying this sexes (CHAN-YONGand JALLON 1986). It seems that mutation were crossed withflies carrying the ywmf this active biosynthesis continues over 10 days in D. marked X chromosome. The 7-T:27 Br ratio seems simulans females. Depending onthe kiti genotype, controlled mainly by kiti. This parameter identified Q7-T seems to reflect changes in reproductive physi- females of different genotypes: there was astrong ology ofheterozygous females with increasing age and correlation between the distribution of this hydrocar- following mating. Homozygous kite’ females produce bon ratio and fecundity. Other minor loci on X chro- low amounts of Q7-T which remain constant through- mosome affected Q7-T or/and Q27 Br levels. Those out the life of the , irrespective of mating status. loci are probably responsible for differences detected The 7-T production may thus depend on two genetic either between Seychelles and pmf strains (see RE- systems: one including kiti, which is physiologically SULTS) or between strain 430 (kiti’) males and F* regulated, and another one which is independent of recombinant kiti’ males where Q7-T seems different this regulation and includes loci likeNgbo. Preliminary (283 & 15 ng and 606 & 21 ng, respectively). The data show that kiti’/Y; NgboCam/NgboCamflies have de- polygenic control of Q7-T onthe X chromosome fects in both viability and fertility. Presumably the could explain the continuous variation observed be- double mutant has very little if any 7-T. This reduc- tween differentrecombinants for this chromosome tion of fitness therefore supports the hypothesis that (Table 3). The higher levelsof (227 Br seemed to a minimal synthesis of 7-tricosene, or its precursors, segregate with forked. No significant correlation was is required for survival in D. simulans. found between Q7-T and Q27Br variations. Although no major behavioral anomaly was found We have previously described alocus in this species, to be linked to the kktk’ hydrocarbon phenotype, a Ngbo (11, 65.3), which controls the ratio of Q7-T and more detailed analysis of male and female courtship Q7-P (FERVEUR 1991).The production of 7-P is di- behaviors would be necessary in order to definitely rectly related to theNgbo genotype, which is additively exclude any behavioral effect of the mutation. It is Genetic Control of Pheromones 567 clear that the amountof 7-T synthesized by kkti' flies technical assistance of TERUYOIWATSUBO. MATTHEW COBB, RALPH is sufficient to elicit a response in sexual partners. GREENSPANand the reviewers are thanked for their comments on the manuscript. Although homozygous females copulateand lay eggs, the mutation appears to exert a maternal effect LITERATURECITED which prevents egg hatching. Despite 14 generations of backcrosses in isofemale lines, it was not possible to BENAMAR,O., and J.-M. JALLON,1983 The production of 7- tricosene, an aphrodisiac of Drosophila simulans, is under the separate the hydrocarbon phenotype associated with control of one gene on the X chromosome (abstr.) Behav. kiti from the egglethality. This strongly suggests that Genet. 13: 533. the various phenotypes eitherdepend on a single BLOMQUIST,G. J., J. W. DILLWITH and T. S. ADAMS, pleiotropic locus or on several very tightly linked loci. 1987 Biosynthesis and endocrine regulation ofsex phero- kiti might belong to the B1 class of X-linked female mone production in Diptera, pp. 217-243 in Pheromone Bio- chemistry, edited by G. PRESTWICHand G. J. BLOMQUIST. Aca- sterilemutations as defined in D. melanogaster by demic Press, San Diego, Calif. PERRIMONet al. (1 986). The egg sterility is due to a CHAN-YONG,T. P., and J.-M. JALLON,1986 Synthese de novo de maternal defect certainly related to oogenesis while hydrocarbures potentiellement aphrodisiaques chez les Droso- the second phenotype is zygotically expressed by a philes. C. R. Acad. Sci. 303: 197-202. COBB,M. J., and J.-M. JALLON,1990 Pheromones, mate recogni- reduction in Q7-T. Under this assumption, the kiti' tion and courtship stimulation in the mutation might be an amorphicor hypomorphic allele species sub-group. Anim. Behav. 39 1058-1067. of a nonvital gene. DE RENOBALES,M., T. S. WOPODINand G. J. BLOMQUIST, As yet we have no data about the morphogenetic 1986 Drosophilamelanogaster fatty acid synthetase. defects responsible for egg lethality in kiti' mutants. Biochem. 16: 887-894. A link between the two mutant phenotypes might be FERVEUR,J.-F., 1991 Genetic control of pheromones in Drosophila simulans. I. Ngbo, a locus on the second chromosome. Genetics the implication of common biochemical intermediates 128: 293-301. in both hydrocarbon biosynthesis and thelipid part of FERVEUR,J.-F., M. J. COBBand J.-M. JALLON, 1989 Complex the egg yolk and/or chorion lipoproteins. chemical messages in Drosophila, pp. 397-409 in Neurobiology If kiti is not itself a sex-determinationgene, it might of Sensory Systems, edited by R. N. SINGHand N. J. STRAUSFELD. be triggered by sex-determination genes in order to Plenum Press, New York. JALLON,J.-M., 1984 A few chemical words exchanged by Drosoph- controlthe structural genes of pheromone biosyn- ila during courtship. Behav. Genet. 14: 441-478. thesis enzymes in D. simulans. The gene kiti appears JUAREZ, P., J. CHASEand G. J. BLOMQUIST,1992 A microsomal to control at least two different sex-specific characters fatty acid synthetase from the integument of Blattella germanica in D. simulans. Although 7-T production shows no synthesizes methyl-branched fatty acids, precursors to hydro- qualitative differences between males and females, it carbon and contact sex pheromone. Arch. Biochem. Biophys. 293: 333-341. clearly acts to induce male wing vibration (JALLON LEWIS,E. B., and F. BACHER,1968 Methods of feeding EMS to 1984). It is also possible that female-specific behaviors Drosophila males. Drosophila Inform. Serv. 43: 193. are induced by this substance. In D.melanogaster 7-T LUYTEN,I., 1982 Variation intraspkcifiques et interspkcifiques des is a'male-specific product; females primarily produce hydrocarbures cuticulaires chez Drosophilasimulans. C.R. dienes with 27 and 29 carbons. Given the phenotypic Acad. Sci. 295: 733-736. PECHINE,J. M., F. PEREZ,C. ANTONYand J.". JALLON,1985 A homology in the production of 7-monoenes between further characterization of Drosophila cuticular monoenes these two sibling species, and given the quantitatively using a mass spectrometry method to localize double bonds in important effect of kiti, we suggest that this locus complex mixtures. Anal. Biochem. 145: 177-182. might control an ancestralhydrocarbon character. PENNANECH, M.,J.-F. FERVEUR,D.B. PHO and J.-M. JALLON, However, hydrocarbon phenotypes do not follow the 1991 Insect fatty acid related pheromones. Ann. SOC.Ento- mol. Fr. 27: 245-263. evolutionary relationship between closely related spe- PERRIMON,N., D. MOHLER,L. ENGSTROMand A. P. MAHOWALD, cies in the melanogaster subgroup (COBBand JALLON 1986 X-linked female-sterile lociin Drosophilamelanogaster. 1990) and characterswhich may have been subject to Genetics 113: 695-712. reinforcing selection are generallynot suitable for SCOTT,D., and R. C. RICHMOND,1988 A genetic analysis of male- establishing phylogenetic relationships. predominant pheromones of Drosophila melanogaster. Genetics 119 639-646. This work would not have been possible without the invaluable Communicating editor: T. SCHUPBACH