Proc. Nat. Acad. Sci. USA Vol. 73, No. 2, pp. 591-593, February 1976 Genetics

Multiple system controlling mating in (behavior/population structure/genetic variation/linkage) W. W. AVERHOFF AND R. H. RICHARDSON Department of Zoology, University of Texas at Austin, Austin, Texas 78712 Communicated by Charles D. Michener, November 14, 1975

ABSTRACT The signals essential to Drosophila melano- the most interesting genomes were lost when the inbred lines gaster courtship include emitted by the female failed to reproduce. which stimulate the male to court and pheromones emitted In new experiments, to avoid losing the most interesting by the courting male which stimulate the female to accept. genomes, we have crossed individuals to "balancer-crossover Genetic variation among these pheromones is a common (if a to not universal) requirement for stimulation of either . The suppressor" strains (4). We hypothesized that, for system signal from the courting male to the female involves both a maintain the genetic variation of a population, most seg- volatile and a nonvolatile component. The volatile compo- ments of the chromosomes would be involved. Therefore, nent is associated with loci on the second and/or third chro- stocks simultaneously containing balanced lethal inversions mosomes, while the nonvolatile component is associated for the X, second, and third chromosomes were used initial- with the X and/or fourth chromosomes. This widespread dis- ly. This approach failed because recombination increased tribution in the genome of loci controlling various compo- nents in the communication network inevitably results in due to multiple rearrangements ("Schultz-Redfield effect") linkage associations with other loci. The genetic array of (5), and having several balanced lethal loci had very was limited. When combined with the negative as- low viability. Instead, we have established a series of stocks sortitative mating pattern produced by the stimulation by balanced only for the second and third chromosomes (SM1- dissimilar pheromones, linkage disequilibrium creates a TM3) (6) and extracted nineteen inbred lines from a collec- strong counterforce to during population bottle- necks. tion of wild flies. Thirteen lines were successfully produced. Six lines produced the expected proportion of wild-type Drosophila communicate through a sensory-stimulus system flies, indicating an absence of deleterious or recessive lethal consisting of a number of modes, e.g., chemical, visual, audi- genes on either the second or third chromosome of these tory, and tactile. In the broadest , courtship behavior in lines. any Drosophila probably utilizes all of these modes. Mate preference tests of these six lines have revealed a Differences between species, or different responses of a sin- new behavioral component under pheromonal control in gle species under varying conditions, indicate emphasis of Drosophila melanogaster. The wild-type males and females different modes utilized for communication. Our research of one line courted and mated normally, both with males has centered primarily on the structure and function of the and females from other lines, and with samples from ran- chemical mode of communication in Drosophila (1-3). domly-mated populations. The males displayed normal Under certain conditions, we have manipulated male courtship with wild-type sibling females having the same courtship in Drosophila melanogaster (1).. We postulated second and third chromosome constitution, but the females that pheromonal variation is qualitative and genetically con- rejected the males' overtures by continually kicking and trolled in Drosophila melanogaster, and that individuals are moving away from the courting males as they attempted to less responsive to those pheromones constantly in higher mount. In our olfactometer tests, however, the females were concentration in that individual's environment. This could induced to accept sibling males when exposed to an airflow be due to accommodation (a spike is no longer carrying pheromones from other . In four such transmitted) to substances that constantly bombard the re- tests, involving a total of twenty males and twenty females ceptor cells. These substances include pheromones that the of this line (five pairs in each test), all of the females dis- individual itself produces, as well as those produced by simi- played immediate acceptance after having continually re- lar genotypes. In the absence of an effective pheromone jected their male sibs during the 30- to 60-min exposure to stimulus, males do not initiate courtship. Courtship ensues pure air. The females ceased kicking and remained in place, when they are supplied with appropriate pheromones that and seven copulations were immediately initiated. Although are qualitatively different from their own or those of their 100% of the females indicated acceptance, 65% of the males sibs. Males derived from some inbred lines neither courted appeared to be conditioned to rejection and continued to nor mated with their female sibs without supplemental stim- court for several minutes before attempting to mount. The ulation. Such induced courtship rarely culminated in copula- conditioned response of the males is analogous to avoidance tion, since females continued to reject the courting males. learning reported for Drosophila melanogaster (7). Both courted and mated normally with flies from un- Since our earlier experiments (1) and the new ones uti- related groups. In our olfactometer tests (1), we alternately lized different methods of extracting homozygous lines, the stimulated or arrested courtship by modulating the phero- two sets of results can be related to genetic differences aris- mnone supply. The modulated pheromones could be obtained ing from the two techniques. The pattern of chromosome directly from flies of a different or from an unpur- variability of a single line differs depending on the proce- ified sample of pheromone collected from a genetically vari- dure of extraction from the "outbred" population. Lines ex- able population. Since was usually not completed, tracted by the inbreeding technique have loci on all chromo- 591 Downloaded by guest on October 1, 2021 592 Genetics: Averhoff and Richardson Proc. Nat. Acad. Sci. USA 73 (1976) Table 1. Model of pheromone control of Drosophila melanogaster courtship and mating Principal Source of Behavior chromosomes Step Signal functional pheromone elicited involved

L Among random wild types 1. Volatile 9 Emitting pheromones d Courtship At least X pheromone(s) different from 6 and/or fourth 2. (a) Volatile 6 Emitting pheromones pheromone(s) different from9 1 2 and 3 2 9 ~~Acceptance (b) Nonvolatile d Producing presumptive X and/or fourth signal(s) pheromones different from II. Within inbred lines (loci on all chromosomes homozygous) 1. Volatile Olfactometer source d Courtship pheromone(s) material 2. (a) Volatile Olfactometer source pheromone(s) material 1 Rejection (b) Nonvolatile Absent, not carried continues signal(s) across olfactometer III. Within 2nd and 3rd chromosome balancer extracted lines (loci on 2 and 3 homozygous, X and 4 segregating) 1. Volatile 9 Emitting pheromones 6 Courtship At least X pheromone(s) different from d 2. (a) Volatile Olfactometer source 2 and/or 3 pheromone(s) material 9 Acceptance (b) Nonvolatile d Producing presumptive X and/or fourth signal(s) pheromones different from

somes becoming homozygous as inbreeding progresses. Lines there is at least one other signal. Since this signal cannot be extracted by balancer techniques have only loci on the sec- supplied in olfactometer tests, it is of low volatility (or it is ond and third chromosomes homozygous when the "wild- not a pheromone). Its function is not lost in balancer-extract- type" is reconstituted. Many X and fourth chromosome loci ed lines, so essential loci are presumed to be on the X and/or are heterozygous, since they are segregating for both the fourth chromosome. The wing vibrations of the courting chromosomes of the balancer and those of the outbred popu- males (always the wing closest to the female being courted) lation. might function in transporting a slightly volatile substance to In Table 1, a summary of all of our findings is presented. the female. Our new data allow us to extend our earlier model to include female acceptance: In the first step, females stimulate the Postulated mode of pheromone action males to court via a'pheromone signal. Balancer extracted' In order to be selective, the females' pheromone receptors- lines consistently exhibited normal male courtship activity. controlling mate acceptance may have a limited number of From previous work, this process requires genetic differ- cells of two or more types, or cells responsive to two' or more ences between males and females. Thus, at least some of the compounds. For mating to occur, each cell type must be ac- components of the stimulus system for male courtship initia- tivated to a threshold level in order to produce an impulse tion are coded on the X and fourth chromosomes, and re- capable of removing mating inhibition or of stimulating mained fully functional. In contrast, inbred lines exhibited mating. Dual activation thresholds of this type are known no courtship behavior until males were stimulated by phero- for many receptors (8). Often, these receptors include mones from another source. Thus, we conclude that there cells that respond to compounds of widely different volatili- are loci associated with female signals on the X and/or ties. In Drosophila melanogaster, the complexity of the re- fourth chromosomes that induce males to court. ceptor system enables the females to be intraspecifically re- The second phase of the courtship sequence involves the ceptive and individually selective. Since this species, as well male's signals to the female, whereby she accepts or rejects as some other drosophilids, live in high density populations, his courtship overtures. This step involves at least two sepa- an effective mate selection system would require at least one rate signal components from the male, one of which can be component that functions only at a very short distance. Such supplied by olfactometer procedures, thus indicating that it a multi-component system appears to be mediating the phe- is a volatile pheromone. Because females of lines isolated by nomenon we have observed. The existence of both long- balancer technique fail to receive a functional volatile pher- range volatile pheromones and short-range slightly volatile omone and reject the males, the loci coding for the origin of or nonvolatile signals could explain some of the nonlinear this signal must be primarily on the second and/or third details of density and proportions which Ehrman (9) and chromosomes. Inability to induce inbred females to accept Spiess (10) have reported in the minority male advantage. sibling male courtship in the olfactometer (1) indicates that Furthermore, Ehrman's work with pheromone effects over Downloaded by guest on October 1, 2021 Genetics: Averhoff and Richardson Proc. Nat. Acad. Sci. USA 73 (1976) 593

short distances (11) may involve the final step in our postu- losophy by W.W.A., and was supported by N.S.F. Grant BMS 72- lated communication sequence involving a slightly volatile 02425. pheromone. 1. Averhoff, W. W. & Richardson, R. H. (1974) "Pheromonal Although we assume that the second signal is a slightly control of mating patterns in Drosophila melanogaster," volatile pheromone, our evidence is indirect. We have been Behav. Genet. 4,207-225. unable to detect audio cues functioning in this way. Females 2. Averhoff, W. W. & Richardson, R. H. (1976) "Pheromones of noncourting, nonmating inbred lines are rarely induced to versus wings in Drosophila melanogaster courtship," Behav. mate in olfactometer tests, even though male sibs are stimu- Genet. 6, in press. lated to court and appear to vibrate their wings normally. If 3. Richardson, R. H. (1974) "Effects of dispersal, habitat selec- cues are tion, and competition on a speciation pattern of Drosophila the frequencies of wing-produced audio strictly endemic to Hawaii," in Genetic Mechanisms of Speciation in species-specific, as reported by Bennet-Clark and Ewing , ed. White, M. J. D. (Australia and New Zealand Book (12), the ability to function in intraspecific discrimination Co., Sydney), pp. 140-164. would be limited. However, some evidence suggests in- 4. Sturtevant, A. H. & Beadle, G. W. (1939, reprinted 1962) in traspecific variations exist (13) and may be involved in some, An Introduction to Genetics (Dover Publications, New York), as yet unknown, way. pp. 207-209. Any population has a finite amount of genetic variation 5. Schultz, J. & Redfield, H. (1950) "Interchromosomal effects for pheromones, from which genetically different homozy- on crossing over in Drosophila," Cold Spring Harbor Symp. gous lines can be extracted (14). The limited number of pos- Quant. Biol. 16, 175-197. sible lines is further reduced with 6. Lindsley, D. L. & Grell, E. H. (1967) "Genetic variations of by linkage disequilibrium Drosophila melanogaster," Carnegie Inst. Washington Publ. deleterious and recessive lethal genes. Our experience with no. 627, 472 pp. both inbred and balancer extracted lines has illustrated the 7. Quinn, W. G., Harris, W. A. & Benzer, S. (1974) "Conditioned case. With each technique, we lost certain lines on the one behavior in Drosophila melanogaster," Proc. Nat. Acad. Sci. hand, while on the other hand, we have produced two or USA 71, 708-712. more lines that yield essentially identical results in phero- 8. Schneider, D. & Steinbrecht, R. A. (1968) "Checklist of insect mone tests. Therefore, a population entering a bottleneck olfactory sensilla," in Invertebrate Receptors, eds. McCarthy, contains a limited array of gametes. Linkage disequilibrium J. D. & Newell, G. E. (Academic Press, New York), pp. 279- between pheromone loci and other loci, coupled with the 297. negative assortitative mating (1), extends the effects of out- 9. Ehrman, L. (1967) "Further studies on mating success and ge- serves to notype frequency in Drosophila," Am. Nat. 101, 415-424. breeding to many loci. This process thereby in- 10. Spiess, E. B. (1968) "Low frequency advantage in mating of crease effective population size through bottlenecks or until Drosophila pseudoobscura karyotypes," Am. Nat. 102, 363- recombination has reduced the linkage disequilibrium. 379. Thus, a complex communication network such as the one we 11. Ehrman, L. (1970) "Simulation of the mating advantage of have postulated can function as an outbreeding system anal- rare Drosophila males," Science 167, 905-906. ogous to self-sterility alleles in (1). 12. Bennet-Clark, H. C. & Ewing, A. W. (1967) "Stimuli provided by courtship of male Drosophila melanogaster," 215, 669-671. The technical assistance of Ms. S. Twombly is gratefully ac- 13. Miller, D. D., Goldstein, R. B. & Patty, R. A. (1975) "Semi- knowledged. Colleagues who have been very helpful in discussing species of Drosophila athabaska distinguishable by male these results include Drs. M. A. Rankin, A. R. Templeton, J. L. Lar- courtship sounds," Evolution, in press. imer, F. M. Bronson, and H. S. Forrest. We thank Dr. L. L. Wheel- 14. Barrows, E. M., Bell, W. J. & Michener, C. D. (1975) "Individ- er for reviewing the manuscript. This study was conducted as par- ual differences and their social functions in insects," tial fulfillment of the requirements for the degree of Doctor of Phi- Proc. Nat. Acad. Sci. USA 72,2824-2828. Downloaded by guest on October 1, 2021