Introduction

All animal species display aggression, an innate behaviour that is evolutionarily conserved. While natural aggressiveness is important for survival and reproduction, abnormal aggressiveness can cause the waste of energy, severe injuries, wars and destruction. Accumulated evidence supports that aggression is influenced by both environmental and genetic factors [1,2,3]. For instance, social experience has been shown to play an important role in modulating the levels of aggressiveness in humans as well as animal models [3,4,5,6]. Recent studies also begin to reveal genetic factors underlying heritable differences in aggressiveness [7,8,9].

Drosophila melanogaster is an excellent model system for studying neural and genetic basis of aggression. Aggressive behaviours in Drosophila were firstly reported by Alfred Sturtevant [10], and later studied in greater details by the groups of Jacobs [11], Hoffmann [12] and Kravitz [13]. Like that in mammals [14], manipulating the levels of neurotransmitters such as serotonin, dopamine and octopamine modulates aggressiveness in Drosophila [15,16,17,18]. Quantitative-trait linkage analyses by Mackay and coworkers suggest that a number of candidate genes may be associated with aggressive behaviours in Drosophila, many of which have homologs in mammals [19]. A recent study also shows that the fly homolog of the gene encoding for neuropeptide Tachykinin/Substance P associated with aggressive behaviors in mammals [20], is also required for aggression in Drosophila [21]. These studies support the evolutionarily conservation of certain genetic mechanisms underlying the control of aggression.

In a search for genetic factors involved in the control of fly aggression, we identify the peacefulness (pfs) gene as a novel and important player required for male-male aggression. Pfs encodes for molybdenum cofactor (MoCo) synthesis 1 protein, an evolutionarily conserved enzyme that catalyzes the first step in the MoCo biosynthesis pathway [22]. MoCo is absolutely required for the activity of molybdoenzymes such as sulphite oxidase, xanthine oxidase and aldehyde oxidase [23]. Interestingly, inhibition of MoCo-dependent xanthine oxidase has been shown to display anti-aggressive effects in humans [24,25,26,27].

In this report, we describe our study on the identification and characterization of pfs in the control of fly aggression. By taking a combination of behavioural analyses, transgene rescue, cell-type-specific knockdown and overexpression, we investigate the requirements and functions of Pfs in regulating fly aggressiveness.

Results

P-element insertion d03517 decreased intermale aggressiveness

In a search for novel genetic factors involved in the control of aggression, we found that mutants homozygous for P-element insertion P{XP}d03517 (d03517) showed a significant decrease in the levels of intermale aggressiveness (Fig. 1a and b). In each experiment, two male flies of same age (isolated for 5–7 days after eclosion) with similar size were introduced into a small chamber, and their behaviours were immediately videotaped for 15-min period. The movies were subsequently analyzed with an automated analysis system CADABRA [28]. Compared to Canton-S (CS) wild-type flies, d03517 mutant flies displayed much fewer lunges and wing threats (Fig. 1a and b, Additional file 1: Movie 1B). We did not observe tussle, a rare and more intense fighting behaviour, in wild-type (n = 27 pairs) or d03517 mutant flies (n = 27 pairs) within 15-min period.

Fig. 1
figure 1

The P-element insertion d03517 decreased male-male aggressiveness. a and b Intermale aggressive behaviors for 15-min period in wild-type and d03517 mutant flies. Prior to behavioral assays, d03517 flies were backcrossed with CS wild-type flies for four generations and thus were in w+ background. a Number of lunges. b Number of wing threats. Pairs of flies tested: wt, n = 27; d03517/ d03517, n = 27. Mann-Whitney U test, ***P < 0.001. c Aggressive behaviours for 15-min period were examined when a wild-type male fly was paired with a d03517 mutant male fly. Pairs of flies tested: n = 46. Mann-Whitney U test, **P < 0.01. d Dominance index for successful occupancy of food patch by a wild-type or a d03517 mutant male fly after 10-min period. Pairs of flies tested: n = 46. Student’s t-test, **P < 0.01. Error bars represent SEM

Full size image

Above phenotypes raise at least two possibilities. For instance, d03517 mutant flies were incapable of initiating aggression. Alternatively or additionally, they may be incapable of evoking aggression by other flies. To distinguish among these possibilities, we paired a d03517 mutant male fly with a CS wild-type male fly and examined their behaviours. We found that wild-type flies displayed much higher levels of aggressiveness than d03517 mutants (Fig. 1c). This result suggests that d03517 insertion interferes with the internal state required for aggression, but does not affect the ability to evoke aggression by wild-type flies (Fig. 1c).

We then tested if higher aggressiveness in wild-type flies gives them competitive advantage over d03517 mutants to defending their territory. In such experiments, a wild-type male fly and a d03517 mutant male fly were placed into a small chamber and allowed to compete for food patch in the center. The frequency for successful occupancy of the food patch was quantitated (see Materials and methods). We found that wild-type flies were much more successful in occupying and defending the food patch than d03517 mutant flies (Fig. 1d. Additional file 2: Movie 2).

In summary, d03517 insertion caused a significant decrease in male fly aggressiveness, which may at least partially account for their disadvantage in defending territory when paired with wild-type male flies.

d03517 insertion did not affect locomotor activity

To test if the decrease in aggressiveness in d03517 mutants was due to some general defects in physical capabilities, we examined fly locomotion over 15-min period. A wild-type male fly was paired with a d03517 mutant male fly, and their movements were videotaped and analyzed. No significant difference in the patterns or total distance of movements was observed between wild-type and d03517 mutant male flies (Additional file 3: Figure S1A).

We also performed climbing test to examine potential effects of d03517 insertion on fly motor functions. No significant difference in climbing ability was observed between wild-type and d03517 mutant male flies (Additional file 3: Figure S1B). These results argue against that the observed decrease in aggressiveness was caused by defective physical capabilities.

d03517 insertion did not affect olfactory avoidance response

Olfactory sensation plays important roles in regulating fly behaviours, such as aggression and courtship [29,30,31]. This raises the possibility that the decrease in aggressiveness of d03517 mutants was caused by defective olfactory sensation. To test this, we performed experiments to examine the response of wild-type and d03517 mutant flies to benzaldehyde, a strong odorant repellent (Additional file 4: Figure S2A). We found that like wild-type flies, d03517 mutant flies could effectively detect and avoid the area with benzaldehyde (Additional file 4: Figure S2B).

d03517 insertion did not affect sexual behaviours

When encountering other flies, a male fly has to make certain mutually exclusive decisions, such as fighting or courtship. The observed decrease in intermale aggression of d03517 mutant male flies may reflect a specific failure of initiating and/or executing fighting, or reflect a switch in decision making due to altered sexual orientation.

To distinguish among these possibilities, we assessed the ability of d03517 mutant male flies to distinguish between males and females. A decapitated virgin female and a decapitated male were placed on different areas in a small chamber. A wild-type or a d03517 mutant male fly was then introduced into the chamber. Like wild-type male flies (Fig. 2a), d03517 mutant males selected the decapitated virgin female over the decapitated male for showing courtship behaviours (Fig. 2a, Additional file 5: Movie 3). This result indicates that d03517 mutant male flies were still able to recognize sexual identities of other flies, and their sexual preference was not altered.

Fig. 2
figure 2

d03517 insertion did not affect sexual behaviours. a Sexual discrimination assay. Both wild-type and d03517 mutant male flies preferred to show courtship behaviours towards decapitated virgin female flies than decapitated male flies. Number of male flies tested: wt, n = 20; d03517, n = 20. Mann-Whitney U test, **P < 0.01. b-d Male-male courtship behaviours for 15-min period. Wild-type and d03517 mutant male flies showed very similar male-male courtship indices (Mann-Whitney U test, P > 0.05), including one-wing extension frequency (b), circling frequency (c), latency to courtship (d). “ns”, not significant, P > 0.05. Pairs of flies tested: wt, n = 27; d03517, n = 27. Error bars represent SEM

Full size image

Above results, however, do not exclude the possibility that when only encountering a single male fly, a d03517 mutant male fly may choose courtship over aggression leading to a decrease in aggressiveness. To address this possibility, we examined male-male courtship behaviours. A male fly was paired with another male fly, and courtship indices (i.e. unilateral wing vibration, circling frequency, and courtship latency) were analyzed. Wild-type flies showed low-frequency male-male courtship behaviours (Fig. 2b-d). Compared to wild-type male flies, d03517 mutant male flies did not show an increase in male-male courtship behaviours (Fig. 2b-d).

We also analyzed male-female sexual behaviours. A male fly was paired with a virgin female fly, and courtship indices (i.e. unilateral wing vibration, circling frequency, courtship latency and copulation latency) were analyzed. No significant difference in male-female mating behaviours was observed between wild-type and d03517 mutant male flies (Additional file 6: Figure S3).

In summary, d03517 insertion decreased fly aggressiveness, but did not affect locomotor activity, olfactory avoidance response and sexual behaviours. We named the corresponding gene of this phenotype (i.e. decrease in aggressiveness) peacefulness (pfs), and d03517 insertion is hereinafter referred to as pfsd03517.

The pfs gene encodes for the fly ortholog of Mocs1

pfsd03517 is inserted into a genomic site within the first exon of the gene CG33048 located on the 3rd chromosome (Fig. 3a) [32, 33]. CG33048 encodes for an enzyme that is the fly ortholog of Molydenum Cofactor Protein 1 (Mocs1). In addition to CG33048, several other genes are also located close to the d03517 insertion site. Since d03517 is inserted into the first exon of CG33048, we performed complementation tests to examine if d03517 is allelic to available mutations affecting CG33048. We firstly tested mocs11, a partial loss-of-function mutation that decreases the enzymatic activity of Mocs1 [34, 35]. Sequencing analysis shows that mocs11 carries a C-to-T point mutation in the coding sequence that changes Thr304 into Ile (see Materials and methods). We found that both mocs11 homozygotes and d03517/mocs11 trans-heterozygotes showed a significant decrease in the levels of intermale aggressiveness (Fig. 3b). We then examined another P-element insertion line PBac{WH}f03019 (f03019) in which P-element is inserted into the 4th exon of CG33048 [32, 33]. Similarly, a significant decrease in intermale aggressiveness was observed in f03019 homozygotes and f03019/mocs11 transheterozygotes (Fig. 3b). These results suggest strongly that the pfs gene is CG33048.

Fig. 3
figure 3

The pfs gene is CG33048 that encodes for the fly ortholog of Mocs1. a The organization of genes near insertion sites of d03517 and PBac{WH}f03019 (f03019). d03517 is inserted into the 1st exon of CG33048, 126 bp downstream of the transcription start site. f03019 is inserted into the 4th exon of CG33048, 104 bp upstream of the transcription stop site. Exons and introns of CG33048 are indicated by blue rectangles and black lines, respectively. b Complementation tests and transgene rescue. Behaviours of a pair of male flies for 15-min period were recorded and analyzed by the CADABRA automated analysis system [28]. All mutant alleles were backcrossed with CS wild-type flies for four generations and thus were in w+ background. Pairs of flies tested: wt, n = 28; mocs11/ mocs11, n = 26; f03019/f03019, n = 25; f03019/mocs11, n = 21; d03517/mocs11, n = 22; genomic rescue construct/+; d03517/mocs11, n = 33. Kruskal-Wallis and post hoc Mann-Whitney U tests, ***P < 0.001. Error bars represent SEM

Full size image

To further confirm this, we performed transgene rescue experiments. We generated transgenic flies carrying a genomic rescue construct containing entire CG33048 sequence, and then crossed this rescue transgene into pfs mutant background. We found that the aggression phenotype in pfs mutants could be completely rescued by CG33048 (Fig. 3b).

Taken together, these results indicate that the corresponding gene of the aggression phenotype is CG33048 that encodes for the fly ortholog of Mocs1.

Pfs is broadly expressed in the brain

We then performed in situ hybridization to examine the expression pattern of Pfs/Mocs1. We found that pfs mRNAs were broadly expressed throughout the brain (Fig. 4a). The intensity of staining was significantly decreased in pfsd03517 homozygous mutants, supporting the specificity of the staining (Fig. 4b and c). Within the brain, pfs mRNA showed higher levels of expression in superior medial protocerebrum (SMP), antennal lobe (AL), and suboesophageal ganglion (SOG) (Fig. 4a).

Fig. 4
figure 4

pfs is broadly expressed in the adult male fly brain. a and b in situ hybridization detecting pfs mRNAs in an adult male fly brain. a Wild type. b pfsd03517 homozygous mutant. c Relative expression levels of pfs were quantified. Compared to that in wild type, the level of pfs mRNAs was significantly reduced in pfsd03517 homozygous mutants. Number of flies tested: wt, n = 7; pfsd03517, n = 7. Student’s t-test, *P < 0.05. Error bars represent SEM. d and e Confocal sectioning of fluorescence in male adult brains of w1118 control flies (d) and w1118 flies carrying GFP-tagged genomic pfs transgene (i.e. genomic-pfs-GFP) (e). d Autofluorescence background in w1118 control flies. e Pfs-GFP was broadly expressed in the adult brain of w1118 flies carrying genomic-pfs-GFP. Abbreviations: SMP, superior medial protocerebrum; AL, antennal lobe; SOG, suboesophageal ganglion. Scale bar: 50 μm

Full size image

To examine the distribution of Pfs protein, we tagged the Pfs protein by engineering the pfs genomic construct that rescued the aggression phenotype (see Fig. 3b). Consistent with the results from in situ hybridization (Fig. 4a), we found that the tagged Pfs under control of the endogenous regulatory sequence showed a broad distribution in the adult brain (Fig. 4e).

Neuronal-specific knockdown of pfs decreased intermale aggressiveness

To determine if pfs is required in neurons for the control of aggression, we performed neuronal-specific knockdown of pfs. A UAS-pfs-RNAi transgene (pfs GL01549) was expressed in all neurons under control of the neuronal-specific driver nSyb-Gal4. We found that male flies expressing this UAS-pfs-RNAi transgene displayed a significant decrease in aggressiveness (Fig. 5a and b). Similar results were obtained when pfs was knocked down by neuronal-specific expression of another independent UAS-pfs-RNAi transgene (pfs7858R1) (Fig. 5a and b). These results indicate an essential role for pfs in neurons for fly aggression.

Fig. 5
figure 5

Neuronal-specific knockdown of pfs decreased aggressiveness. Pfs was knocked down in flies carrying a pan-neuronal-specific driver nSyb-Gal4 and a UAS-pfs-RNAi transgene. Two independent UAS-pfs-RNAi transgenes pfsGL01549 and pfs7858R1 were used in the experiments. a Number of lunges. b Number of wing threats. Behaviours of a pair of male flies for 15-min period were examined. Pairs of flies tested: nSyb-Gal4/+, n = 25; pfsGL01549/+, n = 22; nSyb-Gal4/ pfsGL01549, n = 23; pfs7858R1/+, n = 28; nSyb-Gal4/+; pfs7858R1/+, n = 20. Kruskal-Wallis and post hoc Mann-Whitney U tests, **P < 0.01. ***P < 0.001. Error bars represent SEM

Full size image

Overexpression of pfs greatly increased intermale aggressiveness

Above results indicate a necessary role for pfs in the control of fly aggression. To determine if Pfs actively promotes aggressiveness, we overexpressed Pfs in flies and examined their intermale aggressive behaviours. The genomic rescue transgene containing the entire pfs gene was crossed into wild-type flies. Although one copy of this transgene did not significantly increase the number of lunges or wing threats (Fig. 6a and b), we found that with one copy of this transgene, there was a small but significant increase in tussling, an intense fighting behaviour that is rarely observed in wild-type flies (Fig. 6c). More strikingly, when two copies of this transgene were introduced into wild-type flies, all agonistic behaviours were greatly increased (Fig. 6a-c, compare Additional file 7: Movie S4B to S4A).

Fig. 6
figure 6

pfs overexpression greatly increased intermale aggressiveness. Behaviours of a pair of male flies for 15-min period were examined. a Number of lunges. b Number of wing threats. c Number of tussles. Introduction of one copy of the pfs genomic rescue construct into wild-type flies did not increase the number of lunges or wing threats, but led to a small but significant increase in tussling, an intense fighting behaviour rarely observed in wild type. When two copies of the pfs genomic rescue construct were introduced into wild-type flies, all agonistic behaviours were significantly increased. Pairs of flies tested: wt, n = 22; 1× genomic construct, n = 27; 2× genomic construct, n = 25. Kruskal-Wallis and post hoc Mann-Whitney U tests, *P < 0.05, ***P < 0.001. “ns”, not significant, P > 0.05. Error bars represent SEM

Full size image

Knocking down another component of the MoCo synthesis pathway also decreased intermale aggressiveness

Pfs may regulate intermale aggression through its function in the MoCo synthesis pathway. Alternatively, Pfs may function in a different pathway that is required for fly aggression. To distinguish among these possibilities, we tested if knocking down cinnamon (cin), encoding for another enzyme catalyzing the last step in the MoCo synthesis pathway (Fig. 7a) [22], causes a pfs-like aggression phenotype.

Fig. 7
figure 7

Knocking down cin decreased aggressiveness. a Schematic illustration of the MoCo biosynthesis pathway in Drosophila and humans. Mocs1 catalyzes the conversion of GTP to cPMP, which is converted to molybdopterin (MPT) by Mocs2 in Drosophila and by Mocs2A and 2B in humans. Cin or its human ortholog Gephyrin catalyzes the conversion of MPT to MoCo. b and c cin was knocked down in flies carrying a pan-neuronal-specific driver nSyb-Gal4 and a UAS-cin-RNAi transgene. Two independent UAS-cin-RNAi transgenes cinKK102795 and cinHMS00420 were used in the experiments. b Number of lunges. c Number of wing threats. Behaviours of a pair of male flies for 15-min period were examined. Pairs of flies tested: nSyb-Gal4/+, n = 21; cinKK102795/+, n = 24; nSyb-Gal4/cinKK102795, n = 25; cinHMS00420/+, n = 22; nSyb-Gal4/+;cinHMS00420/+, n = 23. Kruskal-Wallis and post hoc Mann-Whitney U tests, *P < 0.05. **P < 0.01. ***P < 0.001. Error bars represent SEM

Full size image

The expression of cin was knocked down in flies by expressing a UAS-cin-RNAi transgene (i.e. cinKK102795) under control of the neuronal-specific driver nSyb-Gal4. Compared to control flies carrying either the driver or the UAS-cin-RNAi alone, cin knockdown flies showed a significant decrease in aggressiveness (Fig. 7b and c). We also performed knockdown by using a different UAS-cin-RNAi transgene (i.e. cinHMS00420). A similar decrease in aggressiveness was observed (Fig. 7b and c).

To determine potential effects of reducing cin on physical capabilities of knockdown flies, we examined fly climbing ability and locomotor activity. No significant difference in climbing ability was observed when knockdown flies carrying both nSyb-Gal4 driver and UAS-cin-RNAi (i.e. cinKK102795 or cinHMS00420) were compared to control flies carrying nSyb-Gal4 driver or UAS-cin-RNAi only (Additional file 8: Figure S4A).

We also examined fly locomotion over 15-min period. For knockdown using UAS-cinHMS00420, no significant difference in total distance of movements was observed when comparing knockdown flies carrying both nSyb-Gal4 driver and UAS-cinHMS00420 to control flies carrying nSyb-Gal4 driver or UAS-cinHMS00420 only (Additional file 8: Figure S4B). For knockdown using UAS-cinKK102795, the travel distance of knockdown flies carrying both nSyb-Gal4 driver and UAS-cinKK102795 is lower than that of control flies carrying nSyb-Gal4 driver only, but is not significantly different from that of control flies carrying UAS-cinKK102795 only (Additional file 8: Figure S4B). Those results argue against that the decrease in the levels of aggressiveness in cin knockdown flies was caused by impaired physical capabilities.

Taken together, these results support that Pfs/Mocs1 regulates aggression through its action in the MoCo biosynthesis pathway.

Discussion

In this study, we identify Pfs as a novel and important player in the control of intermale aggression in Drosophila. Mutations in pfs decreased intermale fly aggressiveness, but did not affect locomotor activity, climbing ability, olfactory avoidance response and sexual behaviours. Like pfs mutations, knocking down another component (i.e. Cin) of the MoCo synthesis pathway also decreased intermale aggressiveness, supporting a necessary role for Pfs/Mocs1 in the MoCo biosynthesis pathway for fly aggressiveness. That overexpression of Pfs caused a dramatic increase in intermale aggressiveness suggests strongly that Pfs/Mocs1 and the MoCo synthesis pathway actively promote intermale aggression in Drosophila.

We propose that Pfs/Mocs1 controls fly aggression by regulating the synthesis of MoCo, which in turn modulates the activity of MoCo-dependent molybdoenzymes. The MoCo biosynthesis pathway is conserved throughout evolution [22]. MoCo synthesis involves multiple steps that convert guanosine triphosphate (GTP) to MoCo. Mocs1 catalyzes the first step that is the conversion of GTP to cyclic pyranopterin monophosphate (cPMP). cPMP is then converted to molybdopterin (MPT) dithiolate by MPT synthase, which consists of two subunits Mocs2A and Mocs2B. The final step is catalyzed by Gephyrin, leading to the conversion of MPT to MoCo. MoCo forms the active site of all eukaryotic molybdenum-dependent molybdoenzymes such as sulphite oxidase, xanthine oxidase/dehydrogenase and aldehyde oxidase [23]. Interestingly, it is reported that allopurinol, an inhibitor of xanthine oxidase, displays anti-aggressive effects, and could effectively treat dementia and schizophrenia patients associated with escalated aggression [24,25,26,27]. Thus, Pfs/Mocs1-dependent MoCo pathways may control aggression across phylogeny.

Pfs/Mocs1 may control aggression by regulating metabolic activities in the brain. MoCo-dependent molybdoenzymes are involved in the regulation of a number of metabolic activities [23]. Sulphite oxidase is required for the degradation of sulphur-containing amino acids and lipids [36]. Xanthine oxidase mediates the catabolism of purines by converting hypoxanthine to uric acid [37]. And aldehyde oxidase is involved in the catabolism of bioamines such as serotonin and dopamine [38]. Together, these molybdoenzymes may modulate the metabolic state within the brain for promoting aggression. A link between glucose metabolism and aggressiveness has been reported recently [39]. By manipulating oxidation phosphorylation in honeybee and Drosophila, Robinson and coworkers show that aerobic glycolysis increases aggressiveness. Similarly, we speculate that Pfs/Mocs1 regulates MoCo-dependent molybdoenzymes through MoCo synthesis, which in turn modulate metabolic plasticity in the brain for the control of aggression.

Pfs/Mocs1-dependent MoCo pathways may also promote aggressiveness by increasing oxidative stress. Both xanthine oxidase and aldehyde oxidase catalyze the reactions leading to the generation of reactive oxygen species (ROS), such as hydrogen peroxide and superoxide ion [40, 41]. Oxidative stress caused by the accumulation of ROS, has been linked to anxiety and aggression in animal models [42, 43]. For instance, mouse defective in superoxide dismutase 1 (Sod1), an enzyme with antioxidant activity, displays a dramatic increase in aggressiveness [43].

MoCo deficiency (MOCOD) is a rare and severe disease in humans [44]. Patients with MOCOD display severe neurological symptoms, such as intellectual disability, autism, seizures, feeding difficulties, and neurodevelopmental abnormalities. It is suggested that neural damages are mainly due to sulfite oxidase deficiency and the accumulation of toxic levels of sulphite [45]. Over 50% of MoCo deficiency in humans are due to mutations in the MOCS1A open reading frame [46], which mostly result in early death of children [47]. By contrast, we did not observe any developmental defects in fly pfs mutants. One likely explanation is that pfs alleles are not null, and thus do not completely eliminate the activity of MoCo-dependent molybdoenzymes. Alternatively or additionally, flies may be more resistant to the accumulation of toxic metabolic intermediates due to the decrease in the activities of MoCo-dependent molybdoenzymes.

Our results showing the aggression phenotype caused by manipulating the MoCo biosynthesis pathway in Drosophila, together with observed anti-aggressive effects by the inhibition of MoCo-dependent enzymes in human patients, support the existence of a novel and evolutionarily conserved MoCo-dependent mechanism for the control of aggression. A number of neurological and psychiatric disorders, such as schizophrenia, dementia and Alzheimer’s disease, show a substantial association with abnormal aggressiveness [14, 48]. It would be interesting to determine if patients with these disorders show elevated levels of MoCo and/or MoCo-dependent molybdoenzymes. Targeting Pfs/Mocs1 and molybdoenzymes may thus allow the development of novel therapeutic strategies to treat diseases associated with escalated aggression.

Materials and methods

Genetics and rearing conditions

P-element insertion lines P{XP}d03517 and PBac{WH}f03019 were obtained from the Exelixis collection at Harvard. mocs11, UAS-pfs-RNAi-GL01549, and UAS-cin-RNAi-HMS00420 lines were obtained from Bloomington Stock Center. The UAS-pfs-RNAi-7858R1 line was obtained from National Institute of Genetics Fly Stock Center in Japan. The UAS-cin-RNAi-KK102795 line was obtained from the Vienna Drosophila Resource Center. To eliminate the effects of different genetic backgrounds on fly behaviours, all pfs mutant alleles were backcrossed with Canton-S (CS) wild-type flies for 4 generations and were in w+ background. CS flies were used as wild-type controls in the experiments. For knockdown experiments, female flies carrying the Gal4 driver were crossed with male flies carrying the UAS-RNAi transgene. The progeny male flies carrying both Gal4 and UAS-RNAi transgenes were then compared to male flies carrying Gal4 driver or UAS-RNAi transgene only. Flies were reared on standard corn meal at 25 °C and 50–60% humidity with 12 h light-dark cycle.

For rescue experiments, the genomic fragment containing the entire coding sequence (2823 bp) of the CG33048 gene, the 1019 bp sequence upstream of CG33048 and the 755 bp sequence downstream of CG33048, was amplified by PCR with two primers 5’ CTCCGAGCGGAGACTCTAGCGCTAGCCTCTGTGTACTGCACCGTGTA 3′ and 5’ CTCACCATGGATCCAGATCCACTAGTGGGCCCAAAGATGGATGACA 3′. The resulting PCR fragment and the pJFRC-MUH-eGFP vector linearized with Nhe1 and Spe1, were ligated together using In-Fusion HD Cloning Plus Kits (Takara Bio UAS Inc.) to generate genomic rescue construct for making transgenic flies. Genetic crosses were then performed to introduce the CG33048 genomic rescue construct into mocs1 trans-heterozygous mutants (i.e. pfsd03517/mocs11). To overexpress Pfs, wild-type flies carrying one or two copies of the CG33048 genomic rescue construct were generated.

To tag the pfs coding sequence in above genomic rescue construct with GFP, the genomic rescue construct was used as templates for PCR amplification. Two primers 5’ CTCCGAGCGGAGACTCTAGCGCTAGCCTCTGTGTACTGCACCGTGTA 3′ and 5’ CTCACCATGGATCCAGATCCACTAGTTTCGACTTCTGTAACTATCC 3′ were used for amplifying the entire upstream regulatory sequence and the pfs coding sequence before the stop codon. The resulting PCR fragment and the pJFRC-MUH-eGFP vector linearized with Nhe1 and Spe1, were ligated together using In-Fusion HD Cloning Plus Kits (Takara Bio UAS Inc.) to generate the genomic-pfs-GFP construct. The genomic pfs-GFP construct differs from the rescue construct only in the sequence downstream from the coding sequence; while the genomic rescue construct contains 755 bp genomic sequence downstream of the pfs coding sequence, the genomic pfs-GFP construct contains an in-frame fusion between the last codon of the pfs coding sequence and the GFP sequence.

To identify mutation sites in the mocs11 allele, genomic DNA was isolated from homozygous mocs11 mutant flies and used as templates for PCR amplification. The resulting fragment containing the entire sequence (2823 bp) of pfs was sequenced completely, which was then compared to that of wild-type pfs genomic sequence. A C-to-T missense mutation in the coding sequence that changes Thr304 to Ile in Pfs/Mocs1 protein was identified in the genomic DNA of the mocs11 allele.

Aggression assays

Newly emerged male flies were collected and isolated in 2 ml Eppendorf tubes containing 1 ml fly food for 5 to 7 days before behavioural experiments. Male flies of same age with similar body size were selected for behavioural assays. Aggressive behaviours were examined similarly as described previously [28]. Briefly, two male flies were gently aspirated into a fighting chamber at 25 °C with 50–60% humidity. Their behaviours were immediately recorded with a CCD camera for 15-min period. The movies were then analyzed by using the CADABRA (Caltech Automated Drosophila Aggression-Courtship Behavioral Repertoire Analysis) automated analysis system [28]. For examining aggressive behaviours between a wild-type and a mutant male fly, wild-type flies were labeled with color paint.

To determine the potential dominance, wild-type and mutant male flies were anesthetised by CO2, and marked on thorax with yellow and white acrylic paints, respectively. Flies were allowed for recovery at least 24 h before behavioural assays. A wild-type male fly was paired with a mutant male fly and introduced into a small chamber containing a food patch in the center similarly as described previously [49]. A score of 1 was given if a fly (e.g. wild type) successfully occupied the central food patch after 10-min period. A score of 0 indicates that the fly did not occupy the central food patch after 10-min period. Successful occupancy of the central food patch is considered as an indication of dominance.

Analysis of pfs expression in male adult brains

For fluorescence in situ hybridization, custom Stellaris® FISH probes were designed to detect pfs/mocs1 mRNAs by utilizing the Stellaris® FISH Probe Designer (Biosearch Technologies, Inc., Petaluma, CA) available online at www.biosearchtech.com/stellarisdesigner. The probes were conjugated to the Quasar670 dye and used in FISH assays as described previously [50]. Confocal microscopy was performed by using Olympus laser scanning microscope FV1000. For comparing the relative levels of pfs mRNAs in wild-type and pfs mutant flies, fluorescent intensities in the central brain region were measured.

To examine the expression pattern of genomic-pfs tagged with GFP, adult heads were dissected in phosphate buffer. The brains were immediately mounted on glass slides in phosphate buffer and imaged with confocal microscopy using Olympus laser scanning microscope FV1000.

Male-male courtship assay

To examine male-male courtship behaviours, a pair of male flies with same genotype were introduced into a rectangular chamber. Their behaviours were recorded for 15 min. Unilateral wing extensions and circling numbers were quantified by using the CADABRA automated analysis system. Courtship latency was quantified manually.

Sexual discrimination assay

For each experiment, a CS wild-type virgin female fly and a CS wild-type male fly were decapitated. The decapitated flies were placed to different areas in a rectangular chamber. A test male fly (wild-type or mutant) was then introduced into the chamber. The time that the test male fly showed courtship behaviours towards the decapitated female or the decapitated male fly was quantified.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 7 software. Before data analysis, their normality distributions were examined. Nonparametric tests were performed for data not normally distributed. For comparing more than two genotypes, a Kruskal-Wallis test was performed. If the null hypothesis (i.e. means of all genotypes were the same) was rejected (P < 0.05), we performed multiple Mann-Whitney U tests between a pair of interest to assess whether the means of the two genotypes were significantly different. For comparing two independent groups, an unpaired Mann-Whitney U test was performed. Student’s t-test was performed for data normally distributed.