Analysis of Temperature-Sensitive Mutants Reveals New Genes Involved in the Courtship Song of Drosophila

Corresponding author: Jeffrey C. Hall, Department of Biology, 119 Bassine Bldg., Brandeis University, 415 South St., Waltham, MA 02254-9110, hall{at}binah.cc.brandeis.edu (E-mail).

Communicating editor: C.-I WU

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

cacophony (cac), a mutation affecting the courtship song in Drosophila melanogaster, is revealed to cause temperature-sensitive (TS) abnormalities. When exposed to high temperatures (37°), cac flies show frequent convulsions and pronounced locomotor defects. This TS phenotype seems consistent with the idea that cac is a mutation in a calcium-channel gene; it maps to the same X-chromosomal locus that encodes the polypeptide comprising the -1 subunit of this membrane protein. Analysis of the courtship song of some TS physiological mutants showed that slowpoke mutations, which affect a calcium-activated potassium channel, cause severe song abnormalities. Certain additional TS mutants, in particular para^ts1 and nap^ts1, exhibit subtler song defects. The results therefore suggest that genes involved in ion-channel function are a potential source of intraspecific genetic variation for song parameters, such as the number of cycles present in "pulses" of tone or the rate at which pulses are produced by the male's courtship wing vibrations. The implications of these findings from the perspective of interspecific lovesong variations in Drosophila are discussed.

DURING courtship, males of Drosophila melanogaster and of many other species vibrate their wings, producing a "lovesong" (BENNET-CLARK and EWING 1970 ). This has been implicated as a potential species recognition signal (e.g., KYRIACOU and HALL 1982 , KYRIACOU and HALL 1986 ). A few mutations affecting courtship song have been isolated (HALL 1994A ). Among these, cacophony (cac) is particularly interesting, for example, from an evolutionary point of view. The song of cac males is characterized by longer interpulse-intervals (IPIs) and pulses that contain more cycles than normal (SCHILCHER 1977 ; KULKARNI and HALL 1987 ). These are two features that commonly show differences among Drosophila species (e.g., COWLING and BURNET 1981 ; EWING and MIYAN 1986 ; HOIKKALA and LUMME 1987 ).

cac maps to a locus on the X-chromosome that is also the site of night-blind-A (nbA) visual and l(1)L13 lethal mutations. These genetic variants show a complex pattern of complementation. While the l(1)L13 mutations fail to complement the song and visual defects of cac and nbA, respectively, cac/nbA flies are apparently normal (KULKARNI and HALL 1987 ). Recently, we cloned a gene encoding a new -1 subunit of a voltage-sensitive calcium channel, named Dmca1A (SMITH et al. 1996 ; PEIXOTO et al. 1997 ), which maps to the cac locus. Previously, only one other voltage-sensitive calcium channel (Dmca1D) was known in Drosophila (ZHENG et al. 1995 ), but no behavioral defects have as yet been associated with variations at the autosomal locus encoding Dmca1D.

Ion channels are a diverse class of transmembrane proteins involved in a plethora of cellular phenomena (HILLE 1992 ). Voltage-gated calcium channels, for example, are usually divided into six different classes according to their electrophysiological characteristics, pharmacology, sequence similarities, and tissue distribution (STEA et al. 1995 ). They are involved in many important processes, such as neurotransmitter release and muscular contraction (MCCLESKEY 1994 ). Mutations affecting ion-channel function in D. melanogaster are often associated with temperature sensitivity, paralysis, and related phenotypes (WU and GANETZKY 1992 ). Here we report that cac flies show a temperature-sensitive (TS) abnormality that can best be termed a convulsion. This phenotype is consistent with cac being a mutation in a calcium-channel gene. We also analyzed the song of additional TS mutants, including other ion- channel ones. The results indicate that mutations in genes encoding or affecting ion-channel function are a source of intraspecific variation for the Drosophila's lovesong.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Basic fly handling and Drosophila strains used:
Flies were reared on a sucrose-cornmeal-yeast-Tegosept medium in glass vials (the last ingredient is a mold inhibitor). Stocks were maintained in 12hr:12hr, light:dark (LD) cycles at 25° and 70% relative humidity. Flies were collected and separated by sex as <1-day-old adults under CO₂ anesthesia.

The following stocks, involving genetic variations at the X-chromosomal cac locus, were used in tests of general locomotion and courtship song, except for certain heterozygous female types (see below): unmarked cac, night-blind-A^EE171, and nbA^H18. All three of these mutants had been separately backcrossed seven times to an attached-X stock: C(1)DX, y f. To generate hemizygous mutant females for testing of general locomotion, males from the above stocks were crossed to females from a In(1)FM7c/Df(1)RC29 g stock (FM7c is an X-chromosome balancer mutation, and g is an eye-color mutation). This cross yields heterozygotes carrying each one of the alleles over the deficiency RC29, which uncovers the mutant effects of cac and nbA alleles (KULKARNI and HALL 1987 ). Testing of cac/RC29 females for nonsexually dimorphic phenotypes controlled (as did the backcrossing noted above) for genetic background effects of any recessive factors remaining on the cac-bearing X-chromosome that might contribute to the impairments in question (e.g., Table 1). cac⁺ control flies, and those used to generate certain cac⁺-carrying heterozygotes, were derived from a Canton-S wild-type stock that had been backcrossed seven times to the attached-X stock indicated above.

The following TS mutants were subjected to the song analyses (see below): seizure (sei^ts1), temperature-induced-paralysis-E (tipE), paralytic (para^ts1), no-action-potential (nap^ts1), slowpoke (alleles: slo¹ and slo²), and cysteine-string protein (csp^X1) (LINDSLEY and ZIMM 1992 ; ZINSMAIER et al. 1994 ). All of these mutations are autosomal except para^ts1. sei^ts1 is a mutation in a gene encoding a potassium channel of the eag subfamily, which is closely related to the HERG inwardly rectifying potassium channels (TITUS et al. 1997 ; WANG et al. 1997 ). sei^ts1 causes flies, when exposed to high temperatures, to exhibit uncontrolled flight activity followed by partial paralysis. para^ts1, tipE, and nap^ts1 are mutations affecting sodium-channel function that cause flies to paralyze at high temperatures (reviewed in WU and GANETZKY 1992 ); the paralytic locus encodes the -1 subunit of a voltage-dependent sodium channel (LOUGHNEY et al. 1989 ). The tipE locus codes for a novel membrane protein that enhances the function of para sodium channels (FENG et al. 1995 ). nap^ts1 is a gain-of-function allele of mle, a gene that is required for X-chromosome dosage compensation (KERNAN et al. 1991 ), and it probably affects para-gene expression (cf. WU and GANETZKY 1980 ). slo¹ and slo² are mutations in the slowpoke gene, which codes for a calcium-activated potassium channel (ATKINSON et al. 1991 ; N. S. ATKINSON, personal communication; B. GANETZKY, personal communication). Flies with mutations in this locus are uncoordinated and unable to climb at high temperatures. Even at lower temperatures they show abnormal locomotor behavior and diminished flight ability (ELKINS et al. 1986 ). slo² seems a more debilitating mutation in the sense that its homozygous stock was far more difficult to maintain at 25° than slo¹. The cysteine-string protein (csp) gene is involved in neurotransmitter release (ZINSMAIER et al. 1994 ) and is thought to potentially interact with calcium channels (MASTROGIACOMO et al. 1994 ). The csp^X1 mutant allele causes TS paralysis and early death (ZINSMAIER et al. 1994 ).

Some of these mutants were also analyzed in a cac background (see RESULTS). An attached-X stock was generated in which males were hemizygous for para^ts1 by backcrossing (seven times) flies from the available homozygous stock to the same C(1)DX, y f stock used as above. The majority of recordings were carried out using males from a given attached-X stock, but males from the homozygous para^ts1 stock were also recorded. The stocks containing nap^ts1 and csp^X1 were maintained, respectively, with In(2LR)O, Cy (CyO) and In(3LR)TM3, Sb (second- and third-chromosomal balancers). Homozygous stocks of the mutants sei^ts1, tipE (marked with se), slo¹ (marked with st), and slo² (se and st being eye-color mutants) led to males used in the song recordings. In the case of slo¹ and slo², males from balanced stocks (using In(3LR)TM2, Ubx, and the TM3 stock described above), produced using the original homozygous mutant strains, were also recorded. In Table 3 Table 4 Table 5, some of the flies heterozygous for the third chromosome mutations csp^X1, slo¹, and slo² also carried a balancer (TM2 or TM3).

General behavioral tests:
To quantify loss of body control in heated cac adults, the percentage of time that 20-day-old flies spent on their backs or sides, curling their abdomens, or spinning around was measured. The initial observations of these phenotypes were carried out by placing flies either over a hot plate brought to ~37° (see below) or inside an incubator (also at ~37°) with a glass door. To measure the percentages of time that the flies exhibited one of the three features of these convulsions, a plastic device with a circle of cylindrical chambers (each 10 mm diameter x 3 mm height) was prewarmed to ~37° (for the hot-plate application, the device was placed within a 25° temperature-constant room). One min after the animals were introduced into such chambers with the aid of an aspirator (hence no further anesthesia), they were observed and timed (one fly at a time) for 5 min. The timers involved a bank of electrical devices that accumulated time when switches attached to each timer were flicked on, then off again when the abdominal curling (say) ceased.

A genotypically more extensive version of testing for heat-induced convulsions also involved an adult-age component (Table 1). For this, one to three flies were observed at one time after placing them in a prewarmed mating wheel. The temperature on the surface of the hot-plate was raised to 38–39°, so that the top of the wheel of chambers was 35–36°. Both of these temperature values were measured with a thermometer whose probe lies flat on the surface in question. It was estimated that the temperature inside the chambers was ~37°. (As above, the hot-plate experiments leading to Table 1 were carried out inside a 25° room.) One min was allowed to elapse before the start of the observations, for which only the "on-their-backs-or-sides" phenotype was quantified as a percentage of the following 14 min after a 1-min interval, the subsequent 14 min of behavior was also quantified (Table 1).

Knock-out tests were performed in which eight flies (two from each of four genotypes) were aspirator-loaded into separate chambers of the wheel, which was kept initially at room temperature (25°). The wheel was then placed over the hot-plate at 46° (±0.5°) for 10-min trials to ascertain which flies were knocked out on the floor of the chamber within this period (Table 2).

Courtship song:
Recordings and song analysis was carried out as described by BERNSTEIN et al. 1992 and VILLELLA and HALL 1996 , using an INSECTAVOX (GORCZYCA and HALL 1987 ) and LifeSong software (BERNSTEIN et al. 1992 ). Courtships were recorded for ~5 min using a Sony Hi8 video camera (Sony, Parkridge, NJ). Only pulse song was examined in this report (courtship hums, or sine-song, being the other song, e.g., WHEELER et al. 1989 ; VILLELLA and HALL 1996 ). Usually, all the pulses of the song of a given fly are logged, that is, marked for storage in the relevant file using the computer as an event-recorder, while scanning the visual record of the song along with the video image of the flies' behavior. Logging of some songs extended for only 2 min, and more than 500 pulses were typically logged. Songs with less than 40 pulses were not included in the analysis. Four parameters of the flies' pulse song were measured: interpulse interval (IPI), Cycles-per-Pulse (CPP), amplitude, and intrapulse frequency (IPF) (cf. WHEELER et al. 1989 ; BERNSTEIN et al. 1992 ). CPP and IPF values can vary together among Drosophila types (see references in the INTRODUCTION), but there is no way to predict one value from knowledge of the other; thus, these were treated as separate song parameters. The pulse amplitude measurements were attempts to quantify a song's loudness. This is difficult to measure reliably, and the units specified are arbitrary (see RESULTS); however, we were careful to keep the gain settings during all recordings constant.

Because of the low amplitude of song produced at low temperatures and by some mutants, a background scaling factor (bsf) equal to 1 was used (see VILLELLA et al. 1997 , for an explanation of this feature of the method). Only trains with four or more pulses were logged, and the following IPI cutoffs (minimum to maximum in msec) were used, depending on the temperature: 15° (40–105 msec); 17.5° (30–95 msec); 20° (25–90 msec); 22.5° (20–85 msec); 25° (15–80 msec); 27.5° (10–75 msec); 30° (10–70 msec). These cutoffs were decided based on preliminary analysis and previous work (e.g., WHEELER et al. 1989 ; VILLELLA and HALL 1996 ) showing that, at some upper-limit, a so-called IPI would, in reality, be an interbout-interval.

The recordings were carried out with the INSECTAVOX at the specified temperature, which, at the beginning of the recording, was usually 1° less than the nominal one indicated; and, at the end of the recording, was 0.5° higher. The room in which the recordings were done was adjusted to the desired temperature; however, the light inside the INSECTAVOX caused slight temperature increases during a recording session. For recordings performed at relatively high temperatures (27.5–30°), a water bath was used to keep the flies in the desired condition, and a heating fan was used to warm up the INSECTAVOX. Flies were acclimatized to the different temperatures for at least 30 min before recordings. Virgin females, 1 day old from the attached-X stock (indicated above), and 3–7-day-old males of the various genotypes were used for the recordings.

The number of pulses per minute shown in Table 5 was obtained from the results of the song analyses described above. The wing-extension time was measured with the aid of electric timers, and the logging of this behavior was performed observing the same videotape recordings used for song analysis (cf. VILLELLA and HALL 1996 ). The Wing-Extension Index (WEI) represents the durations of such behavioral bouts divided by the total recording time. The number of pulses per minute of wing extension was obtained by dividing the number of pulses per min by the Wing-Extension Index for each fly.

Statistics:
Statistical analyses were carried out using JMP software (Macintosh version 3.1; SAS Institute, Inc., Cary, NC) and according to SOKAL and ROHLF 1995 . Nonparametric tests (Wilcoxon/Kruskal-Wallis) were used for the statistical analysis of convulsion episodes (Table 1). The song parameters IPI and CPP shown in Figure 1 and the number of pulses per minute of wing extension shown in Table 5 were subjected to log transformation prior to analysis of variance to approximate normality in the distributions and homogeneity of variances. Comparisons to the wild type and cac controls in Table 3 and Table 4, respectively, were carried out using Dunnett's method, whereas the multiple comparisons of Table 5 used the Tukey-Kramer method.

View larger version (25K): In this window In a new window Download PPT slide   Figure 1. Temperature-dependence of courtship-song characteristics exhibited by wild-type, cacophony , and parats1 males. A minimum of 4 and a maximum of 16 flies were analyzed for each genotype and temperature. Four pulse-song parameters were measured: (A) amplitude of sound; (B) IPI (InterPulse-Interval); (C) CPP (Cycles-Per-Pulse); (D) IPF (Intrapulse-Frequency). Amplitudes were measured using an arbitrary scale, while IPIs and IPFs were measured, respectively, in milliseconds (msec) and Hertz (Hz). Means and their standard errors are plotted. Analysis of variance of all four song parameters was carried out, modeling temperature as a continuous variable. The analysis of the amplitude data indicated significant effects for genotype (F[2, 140] = 6.70, P = 0.0017) and temperature (F[1, 140] = 35.37, P < 0.0001), as well as for the genotype x temperature interaction (F[2, 140] = 6.03, P = 0.0031). The same is true in the case of the IPI data, for which there are significant effects of genotype, (F[2, 140] = 7.70, P = 0.0007), temperature (F[1, 140] = 1678.96, P < 0.0001), and their interaction (F[2, 140] = 3.70, P = 0.0272). Analysis of the CPP data showed significant effects for genotype (F[2, 140] = 22.13, P < 0.0001) and the genotype x temperature interaction (F[2, 140] = 5.09, P = 0.0073), but not for temperature (F[1, 140] = 0.86, P = 0.3543); whereas, for the IPF data, the analysis indicated significant effects for temperature (F[1, 140] = 33.32, P < 0.0001) and the genotype x temperature interaction (F[2, 140] = 4.50, P = 0.0127), but not for genotype (F[2, 140] = 3.01, P < 0.0524).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

cacophony is a temperature-sensitive mutant:
When ex-posed to high temperatures (~37°) cac flies show frequent convulsions and pronounced locomotor defects. This convulsion phenotype is characterized by flies turning upside-down or on their sides, shaking their legs for a few seconds, and then turning right-side up. The flies also curl their abdomen severely, either when on their backs or when walking, and twist their bodies at the same time. In addition, occasionally the cac adults will walk sideways, spin around on the same spot for a couple of seconds (apparently completely disoriented), leap across the chamber, or jump and tumble up and down out of control. There was no obvious sequence in the occurrence of these phenotypes. After long exposures to 37°, cac flies spend more and more time on their backs, shaking the legs until they seem to collapse. This typically requires more than 1 hr of heating for 1-day-old flies, but much less for older ones (cf. Table 1). As long as leg movement was still occurring, the mutant individuals usually recovered in a few minutes after transfer to room temperature (25°).

In tests involving exposure of 20-day-old cac-expressing flies to 37°, hemizygous mutant males and cac/Df(1)RC29 females were observed (MATERIALS AND METHODS explains why hemizygous mutant females were used, notwithstanding the male-limited song defects caused by cac). Over the course of 5 min at 37°, the former flies (n = 11) spent about half of this time period on their backs or sides, the latter (n = 10) ~40% of the time. The mutant males curled their abdomens during ~35% of the observation periods; the (hemizygous) mutant females, ~20%. Each mutant type spun around in the chamber for ~1% of the 5 min. These convulsions were not observed in similar tests of cac⁺-bearing male adults (n = 5). Such normal flies will occasionally lie on their backs after falling from the ceiling of the heated chamber, but they right themselves within a couple of seconds. Moreover, curling of the abdomen was rarely observed, and only as a fleeting action in cac⁺ males. When it occurred, it was less severe with no twisting of the body. The nbA^EE171 and nbA^H18 visual mutants (n = 5 males each), caused by mutations that map to the same locus as cac, behaved like cac⁺ in these assays.

Table 1 shows a documented comparison of heat-induced convulsions among genotypes and ages (albeit only enumerating the percentages of time flies spent on their backs or sides). It is obvious that females carrying the cac mutation heterozygous with the RC29 deletion spent far more time, at any age, on their backs or sides than females heterozygous for a wild-type derived X-chromosome and RC29, or that deletion and nbA^EE171 or nbA^H18. There was also a strong age effect in these experiments. Older flies (Table 1) showed convulsions more readily and collapsed sooner after the temperature was raised. This, however, might simply reflect the normal decrease in high temperature tolerance with age (ASHBURNER 1989 ), as suggested by the data from 20-day-old RC29/+ flies (Table 1).

The convulsion phenotype of cac and its associated locomotor problems might be exploited in the future, for example, to isolate new cac alleles or suppressers using simple knock-out assays. The results in Table 2 shows how this might be done. Here, almost all of the cac/RC29 flies were knocked out within 10 min when placed over a 46° hot-plate, whereas the same happened only to 8 out of 150 (5.3 %) flies of the other three genotypes: RC29/+, nbA^EE171/RC29, and nbA^H18/RC29.

The fact that cac is a TS convulsion mutant raises two questions. The first is whether temperature variation might have an effect on the song produced by cac that is different from its effects on wild-type flies (SHOREY 1962 ; RITCHIE and KYRIACOU 1994 ). The second question is whether other D. melanogaster mutants causing TS paralysis and related phenomena show any alterations in the courtship song. These two questions are addressed below.

Temperature effects on the cacophony courtship song:
To examine the effects that temperature variation might have on the pulse song produced by cac, a song analysis of cac and wild-type flies was carried out at temperatures ranging from 15–30° in steps of 2.5°. (cac⁺ males are here called wild-type ones, although in reality the relevant Canton-S stock had been outcrossed to an attached-X one.) Also included in this analysis was the mutant para^ts1, because a preliminary analysis had found it to have an effect on song at 25° (see below). The results are shown in Figure 1, A–D. Four pulse-song parameters were examined: amplitude of sound, IPI, CPP, and IPF. It is evident in these plots that temperature had a major effect on amplitude and IPI of all three genotypes, while it is far less clear in the case of CPP and IPF, even though the temperature effect is significant for the latter (see legend for Figure 1). Significant genotype differences were observed for amplitude, IPI, and CPP but not for IPF. The results also show the basic differences between cac and normal songs (cf. KULKARNI and HALL 1987 ), that is, higher amplitude and CPP, as well as longer IPIs in the former compared to the latter. IPFs were also found in a previous study to be similar between these two genotypes (WHEELER et al. 1989 ).

Although the overall trend observed for amplitude and IPI is similar for the three genotypes (as the temperature rises, there is an increase in the former and a decrease in the latter), differences were revealed in the way the various types of males reacted to temperature. These differences are responsible for the significant genotype x temperature interactions observed (see legend to Figure 1). This is further illustrated in Figure 2, where the differences in IPI (Figure 2A) and amplitude (Figure 2B) between cac and wild type, and between para^ts1 and wild type at each temperature, are plotted.

View larger version (15K): In this window In a new window Download PPT slide   Figure 2. Temperature-dependence of song IPI and amplitude. The courtship songs of the cacophony and parats1 mutants were compared to those of (outcrossed) wild-type males. Significant correlations with recording temperature were found for the IPI differences (A) cacophony (r = -0.7569, P = 0.0488) and parats1 (r = 0.7761, P = 0.0402), and for the amplitude differences (B) of parats1 (r = -0.8854, P = 0.0080), but not for the amplitude differences of cacophony (r = -0.1725, P = 0.7116). The linear regression slopes are shown for the significant correlations.

The difference in IPI between cac and wild type shows a significant negative correlation with temperature (see legend to Figure 2). The difference is actually larger at lower temperatures, a result that is somewhat counterintuitive if one considers that the convulsion phenotype of this mutant occurs at elevated temperatures. It is possible that this reflects in part the nonlinear nature of the IPI change with temperature. No significant correlation with temperature was observed for the amplitude differences.

The difference in IPI between para^ts1 and wild type shows the opposite trend observed for cac. There is a significant positive correlation with temperature (see legend to Figure 2) with the larger IPI difference at 30°. In the case of amplitude, however, the differences between para^ts1 and wild type show a significant negative correlation.

Song analysis of TS mutants:
The fact that cac exhibits TS locomotor phenotypes prompted us to analyze the courtship song of certain TS mutants, most of which were isolated in D. melanogaster as general-locomotor mutants: seizure (sei^ts1), temperature-induced-paralysis-E (tipE), paralytic (para^ts1), no-action-potential (nap^ts1), slowpoke (alleles; slo¹ and slo²), and cysteine-string-protein (csp^X1) (see MATERIALS AND METHODS for references, including one that reports the creation of the csp^X1 mutant by reverse genetics).

The results of the song analysis of flies carrying the above mutations in homozygous and/or heterozygous condition are shown in Table 3. An extension of this analysis, an examination of a subset of these mutations in a cac background, is presented in Table 4. All recordings were done at 25°, and the same four pulse song parameters were measured: IPI, CPP, amplitude, and IPF.

As can be seen from the results of Table 3, para^ts1 significantly increases the IPI at this temperature compared to the wild-type controls. This result led us to use this mutant in the analysis at different temperatures of the kind introduced above. Significant effects on singing were also observed for the two other mutations affecting sodium-channel function: lower amplitude and IPF in the case of tipE, and longer IPI for nap^ts1. However, the most interesting nap^ts1 effect occurred when this mutation was in a cac background (see Table 4). While nap^ts1 seems to enhance the IPI defects of cac, it had the opposite effect on cycles per pulse (CPP) and amplitude. For those parameters, nap^ts1 seems to suppress some of the mutant nature of cac's song. This interaction between nap^ts1 and cac is illustrated in Figure 3, which shows some examples of song traces of males carrying the relevant genotypes. Males expressing (homozygous) nap^ts1 alone generated trains of song sounds whose individual pulses not only looked liked those of wild type (not shown in Figure 3), but were also normal by analysis of intrapulse parameters (Table 3). The only numerical difference between nap^ts1 songs and wild-type ones was a longer-than-normal interpulse interval (Table 3).

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. Examples of song traces of wild-type, cacophony , and doubly mutant cac napts1/napts1 males. How the napts1 mutation suppresses aspects of the polycyclic and high-amplitude nature of cacophony's pulse-song, while elongating its IPI, is revealed in these pulse trains. The time base (abscissa) is in sec; these numbers vary considerably among the three lines of traces because they depict the singing behavior that occurred at different moments from the beginning of a given recording session.

Attempts were made to record males homozygous for the csp^X1 mutation of the cysteine-string-protein gene. However, these flies seemed too feeble to show any sign of courtship behavior. The same was true for the few cac ; csp^X1/csp^X1 males that were obtainable. In this respect, flies expressing either of these genotypes flies died within less than a week after adult-emergence. The heterozygous csp^X1/+ type gave increases in CPP and amplitude, as well as a decrease in IPF (Table 3). These effects were not significant in a cac background (cac ; csp^X1/+ in Table 4), although the song-parameter changes were in the same direction as in csp^X1/+.

Mutations in two potassium-channel genes were examined. sei^ts1 males showed no significant defects in their songs. The mutant alleles, slo¹ and slo², however, define slowpoke as a new courtship-song gene. The sounds produced by these two mutants were clearly aberrant in the pulse songs produced, and they were in fact often difficult to log due to the low-amplitude or polycyclic nature of pulses (at a given moment of singing: see below). Using the same criteria and IPI cutoffs used with the other mutants, all four song parameters examined are affected by these two slo alleles, which cause somewhat distinct song abnormalities. Males homozygous for the slo¹ mutation produce very low-amplitude songs with long IPIs, and low CPP and IPF values. Isolated putative pulses, usually monocyclic signals, often occurred in slo¹ song records; however, they were not logged because they did not occur in pulse trains (see MATERIALS AND METHODS). In the case of the slo² allele, the IPIs of homozygous mutant males were not as long, and the sound amplitude not as low, as in the case of slo¹. A train of pulses in the song produced by flies homozygous for slo² often ends with a highly polycyclic pulse. In fact, the mean number of cycles per pulse of slo²/slo² flies is higher than the wild-type control (see Table 4). Isolated pulses were also often observed, but in this case (cf. slo¹) they are usually highly polycyclic.

Examples of song traces of these two mutants are shown in Figure 4. Heterozygous flies slo¹/slo² show effects intermediate between the two homozygotes. The differences in the phenotypes between the two mutants obviously suggest differences in the molecular nature of the lesions that are unknown. slo¹ is a chemically induced mutation, while slo² was generated using gamma rays. Neither shows any gross chromosomal rearrangements (ATKINSON et al. 1991 ; N. S. ATKINSON, personal communication).

View larger version (16K): In this window In a new window Download PPT slide   Figure 4. Examples of pulse song recorded from males homozygous for the slo1 and slo2 mutations. Note the low amplitude and long InterPulse Intervals in the slo1/slo1 traces. The types of pulse trains exemplified by the two slo2/slo2 ones shown, which frequently included a polycyclic pulse at the end, were rarely seen in slo1 recordings. Thus, slo2 songs (notably, the bottom trace shown here) include pulse trains similar to those generated by dissonance mutant males (cf. RENDAHL et al. 1996 ; STANEWSKY et al. 1996 ).

In a cac genetic background, the songs of the slo¹/slo¹ flies also exhibited longer IPIs and lower amplitudes compared to the control (Table 4). Interestingly, the isolated pulses are polycyclic in this case, and the pulse trains resemble the ones produced by slo²/slo² flies. Examples of song traces of this double mutant are shown in Figure 5. The aberrant nature of the sounds produced by flies carrying both mutations made it even more difficult to log their songs than for the single slowpoke mutants. In fact, trains with polycyclic pulses and IPIs longer than our standard cutoffs used were occasionally observed (see Figure 5). Flies homozygous for the slo² allele rarely emerged as adults when the genetic background included cac (and no cac⁺ allele). This could suggest some sort of interaction between the two genes, or just reflect the fact that slo²/slo² flies are quite sick (see MATERIALS AND METHODS). The song of a single cac ; slo²/slo² male was recorded and analyzed (IPI = 46 Hz, CPP = 5.4 Hz, amplitude = 11.0 Hz, and IPF = 275 Hz). These song-parameter values parallel the effects of this mutation in a cac⁺ background, and the overall pattern resembles a more polycyclic version of cac ; slo¹/slo¹ songs. Note that in Table 5 the number of cycles-per-pulse in cac ; slo²/+ is significantly higher than the cac control, suggesting that, in this background, the slo² mutation is not completely recessive.

View larger version (19K): In this window In a new window Download PPT slide   Figure 5. Examples of pulse song from slo1/slo1 males in a cacophony mutant background. The song trace on the top resembles the song of slo2/slo2, whereas the other two traces show polycyclic pulses; very long IPIs were also observed in these doubly mutant flies.

Finally, one effect that both slo¹ and slo² alleles shared is that flies carrying these mutations exhibited many courtship wing extensions without actually producing audible sound. To quantify this phenotype, the proportion of time flies extended their wings was logged and compared to the number of pulses per minute produced. As can be seen from the data in Table 5, the number of pulses per minute of wing extension is far lower in slo¹/slo¹, slo¹/slo², and slo²/slo² flies than in the case of slo¹/+ and slo²/+ controls.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Much recent progress has been made in the genetic dissection of Drosophila's sexual behavior, as more and more genes are being discovered and characterized, including at the molecular level (HALL 1994A ; YAMAMOTO et al. 1997 ). The identification and isolation of mutants that affect the courtship song production in D. melanogaster have two main goals. One is the dissection of the genetic and molecular components of the sound production machinery, as well as its neural control. The other is the identification and characterization of genes that might diverge in their structure and function over evolutionary time and be involved in the courtship-song differences that are so salient among species of this genus (e.g., EWING 1989 ).

Mutations in Drosophila ion channels are often associated with gross locomotor defects, including heat-induced paralysis (WU and GANETZKY 1992 ), but some of these mutants exhibit subtle behavioral impairments, such as learning and olfactory defects (COWAN and SIEGEL 1986 ; LILLY et al. 1994 ). The finding that a song mutation cacophony, also shows a TS convulsion phenotype is therefore not surprising and is consistent with the likelihood that cac has suffered a mutation in a calcium channel gene (SMITH et al. 1996 ; PEIXOTO et al. 1997 ). This kind of pleiotropy, which is further exemplified by the other visual and lethal mutations at the locus (KULKARNI and HALL 1987 ; SMITH et al. 1996 ), is a common feature of behavioral genes (HALL 1994B ). For example, the clock gene period (per) influences at least three different temporal aspects of Drosophila behavior and development (e.g., KONOPKA and BENZER 1971 ; KYRIACOU and HALL 1980 ; KYRIACOU et al. 1990 ). Another example among song genes is dissonance, mutated at the nonA locus; the nonA^diss mutant has both singing and visual problems (KULKARNI et al. 1988 ; RENDAHL et al. 1992 , RENDAHL et al. 1996 ; STANEWSKY et al. 1996 ). These pleiotropies have important implications for the evolution of such behavioral genes and the phenotypes they control (see below).

The song analysis of temperature-sensitive mutants identified at least one new song gene, slowpoke. Flies carrying mutant alleles at this locus present a number of problems in their singing. The song of one of the two alleles examined slo¹ is characterized by very low-amplitude pulses and long IPIs. In the case of slo², the trains of pulses often end with a highly polycyclic pulse. This makes some of the singing bouts for slo² males resemble those often produced by dissonance (KULKARNI et al. 1988 ; RENDAHL et al. 1992 , RENDAHL et al. 1996 ; STANEWSKY et al. 1996 ). Both slowpoke alleles performed many wing extensions that resulted in little or no song.

As introduced above, cacophony is one of the most interesting song mutations from an evolutionary point of view, at least in part because its abnormal pulses are nicely patterned, as in the case of wild-type males from various Drosophila species, and do not appear to be pathologically defective. A similar statement is possible about the songs of slowpoke males, although perhaps some of these mutant song bouts are more in the category of an erratic mess. Nevertheless, it is hard to believe that the song produced by double mutants cac ; slo¹/slo¹ comes from D. melanogaster males, so striking are the differences from the wild-type patterns.

The behavioral analysis revealed some additional candidates for song genes. Although some of these defects were subtle, the results obtained with para^ts1 and nap^ts1 are potentially interesting. For example, the changes in the IPI x temperature and amplitude x temperature slopes obtained with para^ts1 (Figure 1) are connected with a possible evolutionary variation in Drosophila courtship that has not often been examined (however, see RITCHIE and GLEASON 1995 ). The phenotypic interaction observed between nap^ts1 and cac not only points to the former as another song gene, but also suggests a possible interaction at the molecular level. It is possible that nap^ts1 affects expression of the X-chromosomal cacophony gene in a manner that is analogous to nap's (mle's) interaction with the similarly located para gene (WU and GANETZKY 1980 ; GANETZKY 1984 ; KERNAN et al. 1991 ).

Another issue concerning the evolution of song genes is their molecular nature and the pleiotropy of neuronal-excitability mutants, which is one theme of this article. Might any neurological or behavioral mutant be so pleiotropic (cf. HALL 1994B ) that this would include song defects? We think not because KULKARNI and HALL 1987 performed a survey of several such mutants and none was found to be song-defective. tipE males were deemed normal in the older study. The slight anomaly of this mutant's song indicated in Table 5 had to be teased out by a more detailed numerical analysis than was performed previously (KULKARNI and HALL 1987 ).

A fair fraction of the song mutants resulting from changes in genes that have been characterized at the molecular level involve membrane excitability. Not surprisingly, these basic functions, when mutated, lead to grossly appreciable defects in behavior. Only some of these mutants are song-defective as well (KULKARNI and HALL 1987 , and the current article). cacophony now finds itself in this category, that is, the courtship variant mutant that started out as a song mutant but is now known to have other phenotypic defects, such as heat-induced convulsions. This kind of general impairment could be at least as detrimental to fitness as cac's song abnormality.

Other pleiotropic song mutants with molecular correlates involve the regulation of gene expression (considered in general terms: transcription or RNA processing). In addition to the period and dissonance mutants in this category (as reviewed by HALL 1994A , HALL 1994B ), consider the fruitless gene and its mutants. These courtship mutations defined a locus encoding a transcription factor (RYNER et al. 1996 ; ITO et al. 1996 ). fru mutations affect courtship song, as well as other aspects of the fly's reproductive behavior (VILLELLA et al. 1997 ), including fertility. Pleiotropies of these sorts place important constraints on the evolution of these behavioral genes.

Genetic variation for features of the Drosophila courtship song have been reported from natural populations (e.g., IKEDA et al. 1980 ; KAWANISHI and WATANABE 1980 ; RITCHIE and KYRIACOU 1994 ; RITCHIE et al. 1994 ). It is possible that the level of genetic variability observed is influenced not only by sexual selection acting on the song parameters themselves, but also by selection on the pleiotropic effects of these putative song genes, which are likely to have broader biological significance than that. These pleiotropic effects could even include other aspects of the mate recognition system (KANESHIRO 1987 ). For example, there are smellblind mutations at the para locus (LILLY et al. 1994 ), which affect the response of males to female pheromones (TOMPKINS et al. 1980 ; GAILEY et al. 1986 ). It is also conceivable that directional selection acting on some of these pleiotropic effects, for example, selection for temperature tolerance and ion-channel genes, could drive changes in the song repertoire that could eventually lead to reproductive isolation between different populations.

While the constraints associated with pleiotropy certainly do not prevent the rapid evolution of Drosophila courtship songs (e.g., RITCHIE and GLEASON 1995 ), it might explain why there is little evidence for genes with major effects in the song in crosses between closely related species (e.g., PUGH and RITCHIE 1996 ). It is likely that the lovesong differences between most such species are based on the cumulative effect of very mild and subtle changes in several genes, at least a handful of them involving, for example, interspecific variations at the cac, slo, and mle (nap) loci.

The major innovations in song production in the genus Drosophila seem to have occurred among Hawaiian flies (HOY et al. 1988 ; HOIKKALA et al. 1994 ) for which founder-effect models of speciation have been proposed (CARSON and TEMPLETON 1984 ). These include, for example, the idea of fixation of a mutation in a major locus, via genetic drift, followed by selection for modifiers on its deleterious effects. Pleiotropy and epistasis have major roles in these models. For instance, PALOPOLI and WU 1994 revealed from detailed analysis of hybrid male sterility between two sibling species of Drosophila that epistasis between conspecific genes is a key component of this sexually related phenotype. Epistatic interactions among song genes, such as the one found between cac and nap^ts1 within D. melanogaster, could also have important implications for sexual selection on the phenotypes they control and on their potential role in speciation (WRIGHT 1982 ).

Because of the role acoustic signals such as the Drosophila's lovesong play in female receptivity, mating preferences, and sexual isolation between species (BENNET-CLARK and EWING 1970 ; SCHILCHER 1976 ; KYRIACOU and HALL 1982 , KYRIACOU and HALL 1986 ; GREENACRE et al. 1993 ; TOMARU et al. 1995 ), song factors are among the best candidates for the so-called "speciation genes" (COYNE 1992 ). The behavioral analysis we present here reveals that mutations in loci affecting ion-channel function might be a source of genetic variation in the fly's lovesong. Because of their enormous diversity (HILLE 1992 ), channel genes might turn out to be among the most common classes of song genes.

	FOOTNOTES

¹ Present address: Fundação Oswaldo Cruz, Departamento de Bioquímica e Biologia Molecular, Rio de Janeiro 21045-900, Brazil.

	ACKNOWLEDGMENTS

We thank B. GANETZKY and N. S. ATKINSON for discussions, and B. GANETZKY for supplying many of the excitability mutants used in this study. We also thank A. VILLELLA for advice about courtship-song analysis. This work was supported by National Institutes of Health grant GM-21473 to J.C.H.

Manuscript received July 10, 1997; Accepted for publication November 7, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ASHBURNER, M., 1989 Drosophila: A Laboratory Handbook. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ATKINSON, N. S., G. A. ROBERTSON, and B. GANETZKY, 1991  A component of calcium-activated potassium channels encoded by the Drosophila slo locus. Science 253:551-555[Abstract/Free Full Text].

BENNET-CLARK, H. C., and A. W. EWING, 1970 The love song of the fruit fly. Sci. Amer. 223 (No. 1): 84–92.

BERNSTEIN, A. S., E. K. NEUMANN, and J. C. HALL, 1992  Temporal analysis of tone pulses within the courtship songs of two sibling Drosophila species, their inter-specific hybrid, and behavioral mutants of D. melanogaster.. J. Insect Behav. 5:15-36.

CARSON, H. L. and H. L. TEMPLETON, 1984  Genetic revolutions in relation to speciation phenomena: the founding of new populations. Annu. Rev. Ecol. Syst. 15:97-131.

COWAN, T. M. and R. W. SIEGEL, 1986  Drosophila mutations that alter ion conduction disrupt acquisition and retention of a conditioned odor avoidance response. J. Neurogenet. 3:187-201[Medline].

COWLING, D. E. and B. BURNET, 1981  Courtship songs and genetic control of their acoustic characteristics in sibling species of the Drosophila melanogaster subgroup. Anim. Behav. 29:924-935.

COYNE, J. A., 1992  Genetics and speciation. Nature 355:511-515.

ELKINS, T., B. GANETZKY, and C.-F. WU, 1986  A Drosophila mutation that eliminates a calcium-dependent potassium current. Proc. Natl. Acad. USA 83:8415-8419[Abstract/Free Full Text].

EWING, A. W., 1989 Arthropod Bioacoustics. Cornell University Press, Ithaca, NY.

EWING, A. W. and J. A. MIYAN, 1986  Sexual selection, sexual isolation and the evolution of song in the Drosophila repleta group of species. Anim. Behav. 34:421-429.

FENG, G., P. DEAK, M. CHOPRA, and L. M. HALL, 1995  Cloning and functional analysis of tipE, a novel membrane protein that enhances Drosophila para sodium-channel function. Cell 82:1001-1011[Medline].

GAILEY, D. A., R. C. LACAILLADE, and J. C. HALL, 1986  Chemosen-sory elements of courtship in normal and mutant olfaction-deficient Drosophila melanogaster.. Behav. Genet. 16:375-405[Medline].

GANETZKY, B., 1984  Genetic studies of membrane excitability in Drosophila: lethal interaction between two temperature-sensitive paralytic mutations. Genetics 108:897-911[Abstract/Free Full Text].

GORCZYCA, M. and J. C. HALL, 1987  The INSECTAVOX, an integrated device for recording and amplifying courtship songs of Drosophila. Dros. Inf. Serv. 66:157-160.

GREENACRE, M., M. G. RITCHIE, C. B. BYRNE, and C. P. KYRIACOU, 1993  Female song preference and the period gene of Drosophila melanogaster.. Behav. Genet. 23:85-90[Medline].

HALL, J. C., 1994a  The mating of a fly. Science 264:1702-1714[Abstract/Free Full Text].

HALL, J. C., 1994b Pleiotropy of behavioral genes, pp. 15–27 in Flexibility and Constraint in Behavioral Systems, edited by R. J. GREENSPAN and C. P. KYRIACOU. John Wiley & Sons, Chichester, UK.

HILLE, B., 1992 Ionic Channels of Excitable Membranes. Sinauer Associates, Sunderland, MA.

HOIKKALA, A. and J. LUMME, 1987  The genetic basis of evolution of the male courtship sounds in the Drosophila virilis group. Evolution 41:827-845.

HOIKKALA, A., K. KANESHIRO, and R. R. HOY, 1994  Courtship songs of the picture-winged Drosophila planitibia subgroup species. Anim. Behav. 47:1363-1374.

HOY, R. R., A. HOIKKALA, and K. KANESHIRO, 1988  Hawaiian courtship songs: evolutionary innovation in communication signals of Drosophila.. Science 240:217-219[Abstract/Free Full Text].

IKEDA, H., I. TAKABATAKE, and N. SAWADA, 1980  Variation in courtship sounds among three geographical strains of Drosophila mercatorum.. Behav. Genet. 10:361-375[Medline].

ITO, H., K. FUJITANI, K. USUI, K. SHIMIZU-NISHIKAWA, and S. TANAKA et al., 1996  Sexual orientation in Drosophila is altered by the satori mutation in the sex-determination gene fruitless that encodes a zinc finger protein with a BTB domain. Proc. Natl. Acad. Sci. USA 93:9687-9692[Abstract/Free Full Text].

KANESHIRO, K. Y., 1987  The dynamics of sexual selection and its pleiotropic effects. Behav. Genet. 17:559-569[Medline].

KAWANISHI, M. and T. K. WATANABE, 1980  Genetic variations of courtship song of Drosophila melanogaster and D. simulans.. Jpn. J. Genet. 55:235-240.

KERNAN, M. J., M. I. KURODA, R. KREBER, B. S. BAKER, and B. GANETZKY, 1991  nap^ts, a mutation affecting sodium channel activity in Drosophila, is an allele of mle, a regulator of X-chromosome transcription. Cell 66:949-959[Medline].

KONOPKA, R. J. and S. BENZER, 1971  Clock mutants of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 68:2112-2116[Abstract/Free Full Text].

KULKARNI, S. J. and J. C. HALL, 1987  Behavioral and cytogenetic analysis of the cacophony courtship song mutant and interacting genetic variants in Drosophila melanogaster.. Genetics 115:461-475[Abstract/Free Full Text].

KULKARNI, S. J., A. F. STEINLAUF, and J. C. HALL, 1988  The dissonance mutant of Drosophila melanogaster: isolation, behavior and cytogenetics. Genetics 118:267-285[Abstract/Free Full Text].

KYRIACOU, C. P. and J. C. HALL, 1980  Circadian rhythm mutations in Drosophila affect short-term fluctuations in the male's courtship song. Proc. Natl. Acad. Sci. USA 77:6929-6933.

KYRIACOU, C. P. and J. C. HALL, 1982  The function of courtship song rhythms in Drosophila.. Anim. Behav. 30:784-801.

KYRIACOU, C. P. and J. C. HALL, 1986  Inter-specific genetic control of courtship song production and reception in Drosophila.. Science 232:494-497[Abstract/Free Full Text].

KYRIACOU, C. P., M. OLDROYD, J. WOOD, M. SHARP, and M. HILL, 1990  Clock mutations alter developmental timing in Drosophila.. Heredity 64:395-401.

LILLY, M., R. KREBER, B. GANETZKY, and J. R. CARLSON, 1994  Evidence that the Drosophila olfactory mutant smellblind defines a novel class of sodium-channel mutation. Genetics 136:1087-1096[Abstract].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego.

LOUGHNEY, K., R. KREBER, and B. GANETZKY, 1989  Molecular analysis of the para locus, a sodium-channel gene in Drosophila. Cell 58:1143-1154[Medline].

MASTROGIACOMO, A., S. M. PARSONS, G. A. ZAMPIGHI, D. J. JENDEN, and J. A. UMBACH et al., 1994  Cystein string proteins: a potential link between synaptic vesicles and presynaptic Ca²⁺ channels. Science 263:981-982[Abstract/Free Full Text].

MCCLESKEY, E. W., 1994  Calcium channels: cellular roles and molecular mechanisms. Curr. Opin. Neurobiol. 4:304-312[Medline].

PALOPOLI, M.F. and C.-I. WU, 1994  Genetics of hybrid male sterility between Drosophila sibling species: a complex web of epistasis is revealed in inter-specific studies. Genetics 138:329-341[Abstract].

PEIXOTO, A. A., L. A. SMITH, and J. C. HALL, 1997  Genomic organization and evolution of alternative exons in a Drosophila calcium-channel gene. Genetics 145:1003-1013[Abstract].

PUGH, A. R. G. and M. G. RITCHIE, 1996  Polygenic control of a mating signal in Drosophila.. Heredity 77:378-382.

RENDAHL, K. G., K. R. JONES, S. J. KULKARNI, S. H. BAGULLY, and J. C. HALL, 1992  The dissonance mutation at the no-on-transient-A locus of D. melanogaster : genetic control of courtship and visual behaviors by a protein with putative RNA-binding motifs. J. Neurosci. 12:390-407[Abstract].

RENDAHL, K. G., N. GAUKHSHTEYN, D. WHEELER, T. A. FRY, and J. C. HALL, 1996  Defects in courtship and vision caused by amino acid substitutions in a putative RNA-binding protein encoded by the no-on-transient A (nonA) gene of Drosophila.. J. Neurosci. 16:1511-1522[Abstract/Free Full Text].

RITCHIE, M. G. and C. P. KYRIACOU, 1994  Genetic variability of courtship song in a population of Drosophila melanogaster.. Anim. Behav. 48:425-434.

RITCHIE, M. G. and J. M. GLEASON, 1995  Rapid evolution of courtship song pattern in Drosophila willistoni sibling species. J. Evol. Biol. 8:463-479.

RITCHIE, M. G., V. H. YATE, and C. P. KYRIACOU, 1994  Genetic variability of the inter-pulse interval of courtship song among some European populations of Drosophila melanogaster.. Heredity 72:459-464.

RYNER, L. C., S. F. GOODWIN, D. H. CASTRILLON, A. ANAND, and A. VILLELLA et al., 1996  Control of male sexual behavior and sexual orientation in Drosophila by the fruitless gene. Cell 87:1079-1089[Medline].

SCHILCHER, F. V., 1976  The role of auditory stimuli in the courtship of Drosophila melanogaster.. Anim. Behav. 24:18-26.

SCHILCHER, F. V., 1977  A mutant which changes courtship song in Drosophila melanogaster.. Behav. Genet. 7:251-259[Medline].

SHOREY, H. H., 1962  Nature of the sound produced by Drosophila melanogaster during courtship. Nature 137:677-678.

SMITH, L. A., X. J. WANG, A. A. PEIXOTO, E. K. NEUMANN, and L. M. HALL et al., 1996  A Drosophila calcium channel -1 subunit gene maps to a genetic locus associated with behavioral and visual defects. J. Neurosci. 16:7868-7879[Abstract/Free Full Text].

SOKAL, R. R., and F. J. ROHLF, 1995 Biometry , Ed. 3, W. H. Freeman & Co., New York.

STANEWSKY, R., T. A. FRY, I. REIM, H. SAUMWEBER, and J. C. HALL, 1996  Bioassaying putative RNA-binding motifs in a protein encoded by a gene that influences courtship and visually mediated behavior in Drosophila: in vitro mutagenesis of nonA.. Genetics 143:259-275[Abstract].

STEA, A., T. W. SOONG and T. P. SNUTCH, 1995 Voltage-gated calcium channels, pp 114–151 in Ligand- and Voltage-gated Ion Channels, edited by R. A. NORTH. CRC Press, Inc., Boca Raton, FL.

TITUS, S. A., J. W. WARMKE, and B. GANETZKY, 1997  The Drosophila erg K⁺ channel polypeptide is encoded by the seizure locus. J. Neurosci. 17:875-881[Abstract/Free Full Text].

TOMARU, M., H. MATSUBAYASHI, and Y. OGUMA, 1995  Heterospecific inter-pulse intervals of courtship song elicit female rejection in Drosophila biauraria.. Anim. Behav. 50:905-914.

TOMPKINS, L., J. C. HALL, and L. M. HALL, 1980  Courtship-stimulating volatile compounds from normal and mutant Drosophila.. J. Insect Physiol. 26:689-697.

VILLELLA, A. and J. C. HALL, 1996  Courtship anomalies caused by doublesex mutations in Drosophila melanogaster.. Genetics 143:331-344[Abstract].

VILLELLA, A., D. A. GAILEY, B. BERWALD, S. OHSHIMA