Sex- and clock-controlled expression of the neuropeptide F gene in Drosophila

Abstract

Drosophila neuropeptide F (NPF), a homolog of vertebrate neuropeptide Y, functions in feeding and coordination of behavioral changes in larvae and in modulation of alcohol sensitivity in adults, suggesting diverse roles for this peptide. To gain more insight into adult-specific NPF neuronal functions, we studied how npf expression is regulated in the adult brain. Here, we report that npf expression is regulated in both sex-nonspecific and male-specific manners. Our data show that male-specific npf (ms-npf) expression is controlled by the transformer (tra)-dependent sex-determination pathway. Furthermore, fruitless, one of the major genes functioning downstream of tra, is apparently an upstream regulator of ms-npf transcription. Males lacking ms-npf expression (through tra^F-mediated feminization) or npf-ablated male flies display significantly reduced male courtship activity, suggesting that one function of ms-npf neurons is to modulate fruitless-regulated sexual behavior. Interestingly, one of the ms-npf neuronal groups belongs to the previously defined clock-controlling dorsolateral neurons. Such ms-npf expression in the dorsolateral neurons is absent in arrhythmic Clock^Jrk and cycle⁰² mutants, suggesting that npf is under dual regulation by circadian and sex-determining factors. Based on these data, we propose that NPF also plays a role in clock-controlled sexual dimorphism in adult Drosophila.

Neuropeptides are one of the principal physiological regulators that control a wide array of biological events in both vertebrates and invertebrates (1). In Drosophila melanogaster, at least 30 genes encoding biologically active neuropeptides are predicted to be present in the fly genome (2, 3). Some of these Drosophila neuropeptides are structurally akin to those found in vertebrates and mediate similar physiological functions. For instance, the homeostatic maintenance of blood sugar levels is mediated by insulin-like peptides and glucagon-like peptide (adipokinetic hormone) in Drosophila (4–6). Such evolutionary conservation in both structure and function has also been shown for Drosophila neuropeptide F (NPF) and mammalian neuropeptide Y (NPY).

Mammalian NPY has received great attention because of its roles in the regulation of diverse aspects of behavior and physiology, including energy homeostasis, circadian rhythms, stimulation of food intake, reproduction, anxiety, seizure, learning and memory, and alcohol addiction (7–10). Members of the NPY family have also been identified in various invertebrate species, including mollusks, platyhelminths, and dipteran insects (11–14). These invertebrate peptides are referred to as NPF to reflect a consensus Phe residue at the C terminus instead of Tyr in vertebrate peptides (12, 14).

In Drosophila larvae, the NPF neuronal network is formed by three distinct pairs of neurons, two of which are located in the protocerebral lobe and one in the subesophageal ganglion (12, 15). Each pair of neurons displays distinct arborizations, implying specific roles played by each type. Consistent with this prediction, NPF-producing neurons in the subesophageal ganglion are likely to mediate gustatory response to sugars in the food (15), whereas protocerebral neurons are involved in developmentally controlled feeding behavior of larvae (16). The NPF signaling pathway was also shown to be involved in the modulation of sensitivity of adult flies to ethanol exposure (17). These various observations suggest that Drosophila NPF plays diverse roles, some of which are similar to those influenced by mammalian NPY. This finding indicates that the NPF signaling system in Drosophila can serve as a valuable genetic model for studying the molecular and neural mechanisms underlying regulation of behavior and physiology by neuropeptides.

Given the importance of NPF in regulating physiological and behavioral responses to distinct environmental and developmental cues (15–17), the identification of upstream regulators of npf is a key step toward understanding regulatory mechanisms of npf expression by these factors as well as its functions. Yet, previous studies have not addressed the question of what types of transcriptional factors control npf expression. In this article, we present evidence that npf transcription in the adult brain is controlled in remarkably diverse manners. npf transcription is regulated by both sex-nonspecific and sex-specific manners, and the sex-specific npf expression in a group of circadian clock-controlling neurons requires functions of sex-determining genes as well as of central clock regulators. Our data also support the roles that npf plays in certain aspects of male courtship behavior and circadian rhythms.

Results

Identification of npf Neurons in the Adult Brain.

As reported previously, six npf neurons are present in an entire CNS of the wandering third-instar larva: two pairs in the dorsal and medial protocerebrum (Fig. 1A) and a pair in the subesophageal ganglion (Fig. 1B) (see ref. 15). Such a simple spatial expression pattern changes significantly during metamorphosis, producing much more complicated patterns in adult brains (Fig. 1 C–E). There is an average of 25.6 ± 1.4 (SEM) npf neurons per brain lobe in males and 19.9 ± 1.1 in females (n = 7 for both sexes). These neurons are widely scattered in both anterior and posterior sides of the adult brain. The striking difference in the number of npf neurons between larval and adult stages suggests that expression of the npf gene is regulated in a stage-specific manner.

Fig. 1.

Spatial and developmental regulation of npf expression. (A-E) In situ localization of npf mRNA. (A and B) Two pairs of npf-neurons (arrowheads) in the protocerebrum (A) and one such pair in the subesophageal ganglion (B) in the wandering third-instar larval CNS. (C–E) Adult brains. npf neuronal clusters within the anterior region of the WT male (C) and female (D) brains and within a posterior region of the female brain (E). To reveal all anterior npf clusters, a composite image of three different optical planes is shown. (F–H) GFP-reported npf expression in the anterior brain of the male (F) and female (G) as well as in the posterior brain of the male (H). At least five specimens for each gender were examined, with similar results. Yellow arrowheads designate male-specific groups (D1, D2, and L1-s). Black arrowheads indicate groups found in common for both sexes (L1-l, L2, S, P1, and P2) (see also Table 1). A white arrowhead indicates a single L1-s cell, which is occasionally found in females (see also Table 2). (Scale bars, 100 μm.)

An npf-gal4 driver containing an ≈1-kb upstream sequence was demonstrated to recapitulate endogenous npf expression in larval CNS (16). Here, we show that npf-gal4-induced GFP-expression patterns are comparable with those detected by in situ hybridization in the adult CNS as well (Fig. 1 F–H; n = 20). Colocalization of the GFP signals in all NPF-immunoreactive neurons further demonstrated the fidelity of the npf-gal4 driver in marking all adult npf neurons (data not shown, n = 14 for both sexes). Because anti-NPF produces nonspecific signals (15), GFP-reported npf neurons together with in situ npf mRNA patterns were further used to examine their neuroanatomical characteristics. Based on the positions and sizes of perikarya and neurite arborization patterns, we classified adult npf neurons into eight different clusters: two posterior, three lateral, one subesophageal, and two dorsal groups (Table 1 and Fig. 1 C–H).

Table 1.

npf neuronal groups in adult brains

Group	Location	Male	Female
Anterior		n = 15	n = 9
L1-l	Dorsolateral	1.1 ± 0.1	1.1 ± 0.1
L1-s	Dorsolateral	2.6 ± 0.2	0.3 ± 0.2*
L2	Lateroventral	1.5 ± 0.2	1.3 ± 0.3
S	Subesophageal	1.5 ± 0.3	1.2 ± 0.3
D1	Dorsal	2.7 ± 0.2	0.0 ± 0.0*
D2	Dorsal	1.0 ± 0.0	0.0 ± 0.0*
Posterior		n = 7	n = 7
P1	Dorsomedian	1.0 ± 0.0	1.0 ± 0.0
P2	Dorsomedian	14.1 ± 0.9	15.1 ± 1.0

Values represent mean ± SEM per brain lobe.

*Male-specific neurons.

Sex-Specific Regulation of npf Expression in the Adult Brain.

Close inspection of npf expression patterns between male and female brains revealed three groups of npf neurons, which occur in a male-specific fashion (Table 2). Two of these neuronal groups are found exclusively in males; one such group (referred to as D1), consisting of 2.7 ± 0.2 neurons per brain lobe, is located in the dorsal area near the antennal lobe (Fig. 1C). Axonal projections stemming from the D1 neurons terminate at the anterior cortical surface above the central complex (Fig. 1F). The other male-specific group (D2) represented by a single neuron is situated dorsal to the D1 neurons and extends its projections into the dorsal brain region (Fig. 1 C and F). The third group (L1-s) located in the laterodorsal area displays sexual differences in cell number (2.6 ± 0.2 in male vs. 0.3 ± 0.3 in female) (Table 2 and Figs. 1 C and D and 2 A and B). Axonal projections from the L1-s neurons travel into the dorsal protocerebrum (Fig. 1F). Because these three male-specific npf neuronal groups are detectable consistently by both in situ hybridization and npf-gal4-driven GFP expression, the 1-kb npf upstream sequence is likely to contain necessary regulatory elements required for the sex specificity of npf transcription.

Table 2.

Numbers of male-specific npf neurons in various genetic backgrounds

L1-s neurons			D1 neurons	D2 neurons
Genotype	Male (n)	Female (n)	Male (n)	Male (n)
WT	2.6 ± 0.2 (15)*	0.3 ± 0.3 (9)*	2.7 ± 0.2 (15)	1.0 ± 0.0 (15)
XX tra1	ND	2.2 ± 0.2 (18)†	2.7 ± 0.2 (10)†	1.0 ± 0.0 (10)†
fru4	0.8 ± 0.2 (11)	0.6 ± 0.1 (13)	0.7 ± 0.3 (12)	0.1 ± 0.1 (12)
dsx1	2.6 ± 0.1 (15)*	0.2 ± 0.1 (10)*	3.3 ± 0.4 (11)	1.0 ± 0.0 (11)
ClkJrk	0.3 ± 0.1 (11)	ND	3.1 ± 0.2 (12)	1.0 ± 0.0 (12)
cyc02	0.1 ± 0.1 (9)	ND	2.9 ± 0.2 (17)	1.0 ± 0.0 (17)

Data represent mean numbers (± SEM) of npf transcript-expressing neurons per brain lobe. Data in parentheses indicate numbers of brain lobes examined. ND, not determined.

*Numbers in males are significantly greater than those in females (t test, P < 0.0001).

^†These counts were performed for XX tra¹ pseudomales.

Fig. 2.

Regulation of ms-npf expression by the tra-dependent sex determination pathway. (A–C) Expressions of ms-npf in D1 (Left) and in D2 (Center) neurons were examined by immunohistochemistry and in L1-s neurons (Right) by in situ hybridization (hyb). Male-specific D1, D2, and L1-s are indicated by blue arrowheads. (A) WT male. (B) Lack of D1 and D2 neurons in WT female. The single L1-s cell type (which was occasionally observed) is indicated by a green arrowhead. (C) Presence of D1, D2, and L1-s in XX tra¹. (D and E) Expression of lacZ in D1 and D2 neurons (arrowheads), in an npf-gal4/UAS-lacZ male (D), was abolished by tra^F-mediated feminization in the brain of a UAS-tra^F/+; npf-gal4/UAS-lacZ male (E). [Scale bars, 100 μm (A–C) and 50 μm (D and E).]

Male-Specific npf Expression Requires fru Function.

Because somatic sexual dimorphism is generated by sequential gene actions in the sex-determination hierarchy in Drosophila (18), we hypothesized that male-specific npf expression (ms-npf) is controlled by genes involved in sex determination. One of upstream sex-determining genes is transformer (tra). A loss-of-function tra¹mutation causes complete sex reversal from a chromosomal female to a pseudomale, which resembles natural male flies morphologically as well as behaviorally (19). Conversely, systemic expression of a female form of tra cDNA (tra^F) in a male during development gives rise to female characteristics (19, 20). Thus, we tested whether the tra-dependent sex determination pathway directs ms-npf expression: Masculinized brains of XX tra¹ flies produced a male-pattern of npf expression (Fig. 2C vs. A; Table 2). In a converse experiment, tra^F-mediated feminization of these neurons by using the UAS/gal4 transactivation system (21) abolished lacZ-reported ms-npf expression (n = 7, Fig. 2 D vs. E). These data support the hypothesis that sexually dimorphic npf expression is under the control of the tra-dependent sex-determination hierarchy.

Two major genes functioning downstream of tra within this hierarchy are doublesex (dsx) and fruitless (fru); dsx controls almost every aspect of sexually dimorphic differentiation of nonneural somatic tissues (22, 23), whereas fru primarily regulates neural functions that are essential for execution of sequential courtship actions displayed by a male (24–27). It is, therefore, possible that products of these downstream genes are actual regulators of ms-npf. We tested this possibility by examining npf expression in various dsx and fru mutant brains. As a result, WT patterns of ms-npf remained unchanged in the male brains of dsx-null (dsx²³/Df(3R)dsx¹⁵, n = 14) and dsx¹ mutants (Fig. 3A and B and Table 2). By comparison, ms-npf is down-regulated in fru mutant brains, because no ms-npf was detectable in a fru-null mutant type (P14/Cha^M5, n = 5), or it was severely reduced in a homozygous fru⁴ mutant (Fig. 3 C and D and Table 2). However, npf expression in sex-nonspecific neurons was unaffected by these fru mutations (e.g., L1-l in Fig. 3D).

Fig. 3.

Expression of ms-npf is regulated by fru. (A and B) Normal expression of ms-npf in dsx mutants. (A) Three NPF-immunoreactive D1-neurons in a dsx-null mutant. (B) npf transcripts in three L1-s and one D2 neurons in an XY dsx¹ fly. (C) Lack of NPF-immunoreactive D1-neurons in a fru-null mutant male. (D) Lack of npf transcripts in L1-s and D2 neurons of a fru⁴ homozygous male. (E–G) Colocalization of FRU^M in ms-npf neurons. GFP-labeled ms-npf neurons (green) are shown (Left), FRU^M-immunoreactivity (red) (Center), and a merger of the two signals (Right). FRU^M immunoreactivity is seen in the nuclei of D1 (E), D2 (F), and L1-s (G) neurons but not in sex-nonspecific L1-l neurons. [Scale bars, 100 μm (A–D) and 25 μm (E–G).]

To further verify cell-autonomous regulatory roles of fru in ms-npf expression, we performed FRU immunohistochemistry on the adult brains containing GFP-marked npf neurons. The results showed FRU^M immunoreactivity in all three types of ms-npf neurons (Fig. 3 E–G; n = 7) but not in npf-neurons common to both sexes (data not shown). These results further support FRU^M as an upstream activator of the male-biased npf expression.

By comparisons with previously classified FRU^M neuronal clusters (28), D1 and D2 neurons are likely to be a subset of fru-AL (#6) and fru-aSP3 (#3) group, respectively, whereas L1-s neurons could be a member of either fru-Lv (#4) or fru-Ld (#10) clusters.

Because FRU^M function is essential for normal male courtship (26, 27), neuronal functions of the ms-npf could be associated with overt male courtship performance. To address this question, we examined courtship of males lacking ms-npf (npf-gal4/UAS-tra^F) as well as males devoid of most npf- neurons as a result of targeted expression of a cell-death gene: reaper (rpr), (npf-gal4/UAS-rpr). NPF immunoreactivity verified that most npf neurons were eliminated by rpr expression, except for a few neurons with projections in the central complex (see Supporting Text and Fig. 6, which are published as supporting information on the PNAS web site).

In a single-pair courtship assay (compare with ref. 29), npf-ablated males displayed subnormal courtship activity toward virgin females: Courtship indices (CIs) measured for npf-gal4/UAS-rpr males were substantially lower than those of appropriate genetic controls (Fig. 4Right). The reduction in such CI values is partly because of prolonged courtship-initiation latency before commencement of active courtship (Fig. 4 Left). Importantly, a comparable deficit in courtship activity was also observed for npf-gal4/UAS-tra^F males lacking only ms-npf (Fig. 4). Thus, these data support our hypothesis that the male mode of specific npf neuronal functions is required for the normal male’s courtship activity.

Fig. 4.

Function of ms-npf neurons in male courtship. Male flies of the genotypes npf-gal4/UAS-tra^F and npf-gal4/UAS-rpr gave subnormal courtship initiation latency (Left) and CI values (Right). These data represent means ± SEM for a number of males tested per genotype inserted within each bar. Significant statistical differences between test groups and a pooled control (bracket) are designated by asterisks (P < 0.01).

Cell-Specific Regulation of ms-npf by Circadian Factors.

We noticed that neuroanatomical features of male-specific L1-s neurons are remarkably similar to those of a well defined neuronal group, called dorsolateral neurons (LN_ds), which produce circadian-clock regulators (30, 31). To confirm whether the L1-s neurons belong to the LN_d cell group, male brains expressing npf-gal4-driven GFP were immunologically processed for the detection of one of the major clock factors, TIMELESS (TIM), as a marker for LN_d cells. Because intracellular levels of TIM protein oscillate with a peak at around Zeitgeber time (ZT)20 (4 h before light-on) and a trough at ZT8 (4 h before light-off) under 12-h light/12-h dark (LD) cycling conditions (e.g., ref. 32), the double-labeling experiment was performed with brains dissected at these two time points, after flies were entrained to 3–5 LD cycles. As expected, strong nuclear TIM immunoreactivity was evident in all known clock neurons, including LN_ds at ZT20 (Fig. 5), whereas such signals were undetectable at ZT8 (data not shown). Approximately six TIM-positive LN_ds were observed, without any noticeable sexual difference (Fig. 5 A vs. B; n = 6 for each sex). Importantly, TIM immunosignals were observed in the nuclei of all NPF-containing (GFP-marked) L1-s neurons, thus verifying those brain cells as a member of the LN_d group (Fig. 5C). Consistent with previous in situ hybridization results (Fig. 2B), GFP-marked L1-s expression is lacking in females (Fig. 5D).

Fig. 5.

NPF-containing L1-s neurons are a subset of TIM-containing LN_d clock neurons. (A and B) Presence of six TIM-immunoreactive LN_ds at ZT20 (arrowheads) in both male (A) and female (B) brains. (C and D) TIM immunoreactivity in L1-s neurons. Nuclear TIM expression (red) is colocalized in all three GFP-labeled L1-s neurons (yellow arrowheads) in the male brain. (C) Because of the lack of GFP-reported npf expression within the L1-s neurons, only TIM expression is seen (red) in the female brain. White arrowheads point to sex-nonspecific L1-l neurons (D). (E and F) Absence of npf mRNA in the L1-s neurons of Clk^Jrk (E), and cyc⁰² mutant (F) males. Expression of this transcript was found to be intact in D2 and sex-nonspecific L1-l neurons. [Scale bars, 25 μm (A–D) and 50 μm (E–F).]

Limited expression of fru-regulated ms-npf in a subset (2–3 cells) of the LN_ds suggests that these “clock” neurons are heterogeneous with respect to their sexual identities and neurochemical phenotypes. Furthermore, the identical numbers of LN_ds present in both sexes (Fig. 5 A and B) demonstrates that sexual dimorphism in the number of L1-s neurons is not due to a gender-associated structural difference but, rather, to fru-directed npf expression in males.

Previous studies showed that expression of a neuropeptide gene encoding pigment-dispersing factor is controlled by circadian-clock regulators, Clock (Clk) and cycle (cyc) in a group of clock neurons called small ventrolateral neurons (s-LN_vs) (33). These data prompted us to investigate whether the LN_ds’ npf expression is also influenced by these clock factors. Remarkably, LN_d-specific npf mRNA was undetectable in the arrhythmic Clk^Jrk or cyc⁰² mutants (Fig. 5 E and F and Table 2), whereas expression of npf in other male-specific D1 (data not shown) and D2 neurons (Fig. 5 E and F), as well as in sex-nonspecific neurons, was unaffected by the same mutations (e.g., L1-l neuron in Fig. 5 E and F). These observations strongly support the argument that npf in the LN_ds is a clock-controlled gene and likely to participate in chronobiological functions.

One possible function of npf within LN_ds is to modulate a gradual increase in locomotor activity levels before the light-to-dark transition, i.e., evening anticipation (34, 35). In our behavioral assays (performed at 20°C), npf-gal4/UAS-rpr male flies show subtle but consistently subnormal evening anticipation; after the quiescent midday period, activity levels were elevated slightly earlier than in controls, between 7 and 11 h after lights-on (Fig. 7, which is published as supporting information on the PNAS web site). This type of behavioral deficit was not noticeable in npf-ablated females. From these results, we speculate that NPF may play a neuromodulatory role for LN_d-controlled behavioral performance of males at dusk.

Discussion

Regulation of ms-npf Expression by a Male Behavior-Determining Factor fru.

Sexual dimorphism of brain structure and function generates differential neural circuitries, ultimately leading to the production of gender-specific behaviors (36). In Drosophila, fru is an essential neural sex determinant responsible for male courtship behavior. Because fru-encoded FRU^M protein is a BTB-Zn-finger transcription factor (37), FRU^M likely regulates expression of an array of genes to establish neural substrates controlling male behavior. However, such downstream targets of the FRU^M are poorly known, hampering our understanding of the molecular mechanisms underlying fru-controlled establishment of male-specific neural circuitry. Intriguingly, our studies identified npf as an at least indirect target of the FRU^M, suggesting that NPF is a neurochemical factor mediating FRU^M functions.

Recent studies on the expression of sex-specific fru transcripts and of reporter expression driven by fru-gal4 suggest that fru acts in establishing sexually dimorphic anatomical differences and in rendering male-specific functions to neurons that are commonly present in both sexes (25–28). Thus, one important question is whether ms-npf is due to the lack of corresponding neurons in the female brains or to cell-specific transcriptional activation of npf by FRU^M in males. Our data support the argument that the latter is the case, at least for L1-s neurons, because a comparable number of such neurons independently marked by anti-TIM was observed in both males and females (Fig. 5), and because FRU^M is persistently present in well differentiated ms-npf neurons (Fig. 3). Therefore, we suggest that one way of masculizing neurons directed by fru is to establish sex-specific production of neurosignaling molecules, which are likely to deliver male-specific neuronal functions. In line with this suggestion, it is notable that male-specific serotonin production in a group of eight neurons in the abdominal ganglion is also controlled by fru (38), and such serotonergic neurons innervate male reproductive organs to control appropriate male mating activities (29). Our data indicate that ms-NPF is another neurosignaling molecule for fru-controlled male courtship. Although the neuronal targets of ms-npf are unknown, prolonged courtship-initiation latencies and general attenuation of courtship activities caused by the absence of ms-npf (Fig. 4) suggest that ms-NPF is involved in the central processing of courtship-activating stimuli or in an “output pathway” that mediates courtship actions.

Sexually dimorphic NPY expression has been described in the rat hypothalamus. Large populations of hypothalamic NPY mRNA-producing cells are localized within the arcuate nucleus. Interestingly, the caudal region of the arcuate nucleus contains significantly more NPY cells in males than in females (39). Further studies suggested that the male gonadal hormone testosterone is a positive regulator of male-biased NPY expression (39–41). Similar sexually dimorphic NPY/NPF expression in the brains of distantly related species suggests that these neuropeptides play conserved roles associated with male-specific CNS functions that underlie sexual behavior.

Clock-Controlled Sexual Dimorphism.

Dual regulation of npf by sex and clock factors within a subset of male LN_d neurons suggests that NPF is associated with clock-controlled sexually dimorphic behavioral performance. In light/dark cycles, the circadian timing system directs bimodal daily peaks of locomotion, occurring at lights-on (morning) and at lights-off (evening) transitions in WT flies (34, 35). Interestingly, a distinct sexual difference was observed in the peak of the morning locomotion, which occurs ≈1 h earlier in males than in females (42). However, this male-specific phase of the morning activity was unaffected by npf-ablation (Fig. 7), suggesting that npf is not associated with this aspect of sexual dimorphism.

The brain-behavioral system is also capable of anticipating photic transitions, as demonstrated by the gradual increase in activity levels before lights-on or lights-off. Among six groups of clock neurons defined in the Drosophila adult brain (30, 31), clock-relevant functions are relatively well studied for the s-LN_vs and the LN_ds. In the former cell type, clock-controlled pigment-dispersing factor production is essential for circadian locomotor activity rhythms as well as lights-on anticipation (33, 43), whereas the LN_ds are required for anticipation of lights-off (34, 35). Our data implicate NPF as a neuromodulatory substance within a subset of the LN_ds, involved in this “late-day” component of the locomotor cycle in males (Fig. 7).

Although the biological meaning of LN_d-regulated lights-off anticipation remains unknown, it is notable that the flies’ increasing locomotion at dusk is temporally coincident with especially vigorous mating activities of fruit flies (44, 45). This finding, then, raises the possibility that anticipatory activity at dusk is causally connected with maximum mating propensity. In this respect, that there are sex- and clock-controlled npf expressions in the LN_ds provides the first glimpse of this neuropeptide as a putative output factor, which would participate in certain aspects of the clock-controlled reproductive behavior (45, 46). These actions are intimately connected to evolutionary fitness, which may be one reason for the circadian system of Drosophila to have evolved and been refined.

Materials and Methods

Fly Strains.

Flies were reared on a cornmeal–yeast–agar medium. Canton-S or white¹¹¹⁸ (w¹¹¹⁸) flies were used as WT or otherwise normal controls. The mutants used in this study are: dsx¹ and dsx²³/Df(3R)dsx¹⁵ (47), tra¹ (37), and cyc⁰² and Clk^Jrk (33). For fru mutants, a FRU^M-null transheterozygote (Df-P14/Df-Cha^M5) and a FRU^M-hypomorphic mutant (fru⁴) were used (38). Two transgenic strains, npf-gal4^II (inserted into the 2nd chromosome) and npf-gal4^III (inserted into the 3rd chromosome), were used in this study (16). For visualization of npf neurons, the npf-gal4 driver was crossed to a UAS-lacZ (48) or UAS-mCD8-GFP (49). To induce npf neuron-specific feminization, a UAS-tra^F (21) was crossed to the UAS-lacZ; npf-gal4^III double-transgenic line. The progeny were assayed for lacZ expression. npf-ablated flies were obtained by crossing npf-gal4 to a UAS-rpr (e.g., ref. 43).

Histochemistry.

In situ detection of npf mRNA in whole-mounted CNS was performed by using digoxigenin-labeled antisense npf cRNA probe as described in ref. 28. To detect npf promoter-driven lacZ expression, brains were dissected, fixed, and stained with X-gal (5-bromo-4-chloro-3-indoyl β-d-galactoside), as described in ref. 33. After maximal levels of staining intensity were elicited, the tissues were rinsed in PBS, dehydrated, and mounted on a slide glass with glycerol.

For NPF immunohistochemistry, brains were incubated with rabbit anti-NPF (16) at a dilution of 1:300 and, subsequently, with FITC- or TRITC-tagged secondary antibodies at 1:200 dilution (Jackson ImmunoResearch, West Grove, PA). To assess colocalization of FRU^M or TIM proteins in npf neurons, brains dissected from UAS-mCD8-GFP/npf-gal4 flies were incubated with rabbit anti-TIM (30) or rat anti-FRU^M (28), both at a 1:300 dilution. The respective primary antibodies were detected by applications of TRITC-tagged secondary antibodies. Colocalization of the two fluorescence signals was assessed by merging two relevant channels. Other details of the immunohistochemistry followed procedures described in ref. 28. All histologically processed brains were viewed under an Olympus (Center Valley, PA) BX61 microscope equipped with a CC12 digital camera, and images were captured by using the program analySIS (v.3, Soft Imaging System Solutions, Lakewood, CO).

Behavioral Assays.

All males and females tested for the courtship assay were in a w¹¹¹⁸ genetic background regardless of other genotypes. Both males and females were collected within 6 h of eclosion and aged in a group of 10–20 flies per food vial for 4–6 days at room temperature. After a single male–female pair was placed in a chamber, as described in ref. 29, the time lapse between initial pairing and commencement of male–female interactions was recorded as a measurement of “initiation latency.” A given male’s behavior was observed over the course of 10-min period or until he initiated mating, if any, before the end of this period. The CI was calculated as a proportion of the observation period during which any courtship actions occurred (29).

Statistical Analyses.

These analyses were performed by using the program Instat (v.3, GraphPad, San Diego, CA). The Student t test was used to detect statistical difference in numbers of L1-s neurons between sexes. One-way ANOVA showed no significant differences in ILs (F = 2.531, P = 0.073) and CIs (F = 2.827, P = 0.057) among genetic control groups (UAS-rpr/+, npf-gal4^III/+, npf-gal4^II/+, and UAS-tra^F/+). Thus, these groups were pooled to generate one control set (n = 39) for normal npf expression (compare with ref. 29). Dunnett’s multiple comparisons after ANOVA were performed to detect significant difference between the control and test group.

Glossary

Abbreviations

CI

courtship index

LN_d

dorsolateral neuron

ms-npf

male-specific npf

NPF

neuropeptide F

ZT

Zeitgeber time.