Mating status-dependent dopaminergic modulation of auditory sensory neurons in Drosophila

Summary

Mating status often modulates responses to courtship sounds in animals. The neural mechanisms underlying this modulation, however, have not been well clarified. Here, we show that dopaminergic signals are involved in modulating the responses of auditory sensory neurons in Drosophila melanogaster females depending on their mating status. These neurons abundantly express three types of dopamine receptors, with some having direct synaptic connections with dopaminergic neurons. Of these receptors, suppressing the expression of Dop1R2 reduces sound responses of auditory sensory neurons in unmated but not mated females. Moreover, the expression of Dop1R2 in auditory sensory neurons enhances the song response behavior of unmated females, manifested by copulation receptivity when exposed to songs. Our research suggests that dopaminergic modulation via Dop1R2 is involved in mating the state-dependent regulation of auditory sensory processing.

Graphical abstract

Highlights

• •Auditory sensory neurons in female Drosophila express dopamine receptors
• •Reducing Dop1R2 expression lowers auditory neuron responses to sound stimuli
• •Dop1R2 effects on neural responses depend on mating drive in females
• •Dop1R2 enhances song response behavior in unmated females

Natural sciences; Biological sciences; Neuroscience; Behavioral neuroscience; Sensory neuroscience

Introduction

Mating states often modulate the auditory processing of animals. In green treefrogs (Hyla cinerea), neural responses in the auditory midbrain to mating calls are reduced in recently mated females compared to unmated females,¹ aligning with a report that post-reproductive females did not respond to mating calls.² In female toadfish (Porichthys notatus), both the sensitivity of auditory hair cells and the behavioral response to male courtship vocalizations increase during reproductive compared to non-reproductive states.^3,4 Flexible auditory processing should benefit from effectively utilizing limited brain resources by biasing sound representation in a state-dependent manner.^5,6 However, the mechanisms responsible for mating status-dependent auditory modulation remain poorly understood.

The acoustic communication system of the fruit fly Drosophila melanogaster has been used as a model for studying the mechanism of auditory processing (e.g., Baker et al.⁷; Deutsch et al.⁸; Kamikouchi et al.⁹; Wang et al.¹⁰). During mating rituals, Drosophila males produce courtship songs by wing vibrations to attract females. Amongst several song types, the pulse song is the primary component for mating success: It carries a species-specific temporal pattern¹¹ that selectively enhances female copulation receptivity.^12,13 After mating, females exhibit decreased copulation receptivity, even in the presence of courtship from conspecific males,¹⁴ suggesting a reduced response to courtship signals.

Fruit flies detect sound signals with their Johnston’s organs (JOs) located in the second antennal segment.¹⁵ The JO contains mechanosensory neurons, known as JO neurons, which project their axons to the antennal mechanosensory and motor center (AMMC) located at the ventrolateral side of the brain.¹⁶ Sound information from the courtship song is primarily transmitted to the AMMC and then relayed to higher-order centers for further processing. Higher-order neurons in the brain, such as AMMC-B1, pC1, pC2l, and vpoEN neurons, are involved in this processing, ultimately modulating the female’s mating decision.^8,9,10,17,18 While the neurons comprising the auditory pathway that processes song information are well characterized, it remains unclear whether the properties of this pathway undergo post-mating modulation. The first step in evaluating this is to investigate whether JO neurons, the initial component of the auditory pathway, receive mating status-dependent modulation.

JO neurons, comprising approximately 480 mechanosensory neurons in D. melanogaster, are subdivided into five groups, i.e., JO-A to E, each exhibiting distinct response properties and spatially segregated axonal projections in the brain.^9,16 Among them, JO-A and JO-B (JO-AB hereafter) neurons preferentially respond to antennal vibrations, serving as the major auditory sensory neurons.^9,19,20 We previously identified abundant postsynaptic sites in the axons of JO-AB neurons within the AMMC.²¹ This finding implied that JO-AB neurons receive synaptic inputs at their axons, which could modulate their response properties.

In this study, we aimed to elucidate cellular mechanisms underlying the mating-dependent auditory modulation in female flies. Using a single-nucleus RNA-seq database, we first examined which neurotransmitter receptors are expressed in JO neurons. Among the identified receptors, we focused on dopamine receptors, as dopamine is used by various animals to modulate sensory information processing in a state-dependent manner. Using T2A-Gal4 knock-in driver lines, we identified three types of dopamine receptors that are abundantly expressed in JO-AB neurons. Combining RNAi-mediated gene knockdown with calcium imaging in unmated and mated females revealed a mating state-dependent effect of Dop1R2 on JO-A responses. Finally, through behavioral assessment, we found that Dop1R2 expression enhances the song response behavior of unmated females. These results demonstrate the potential role of dopamine in linking the mating status to early-stage auditory processing in the female brain.

Results

Dopamine receptor expression in Johnston’s organs neurons

In fruit flies, major neurotransmitters include acetylcholine, glutamate, GABA, serotonin, dopamine, and octopamine.²² To identify which neurotransmitter receptors are expressed in JO neurons, we analyzed the JO neuron dataset derived from published single-nucleus RNA-seq data.²³ This analysis revealed that multiple types of receptors for each neurotransmitter are expressed, suggesting that JO neurons can receive signals from all six neurotransmitter types (Figure S1A).

Among these neurotransmitters, we focused on dopamine, which plays a role in the state-dependent modulation of sensory information processing across various animals, including mice, flies, and nematodes.^24,25,26 Dopamine is also involved in the control of mating drive in males based on their mating status (Zhang et al., 2016,²⁷ 2021²⁸ for flies and mice, respectively), as well as copulation receptivity in females,²⁹ suggesting that it may link mating status to neuronal modulation.

The Drosophila genome encodes four dopamine receptors: Dop1R1,³⁰ Dop1R2,^31,32 Dop2R,³³ and DopEcR.³⁴ Analysis of the JO neuron dataset identified the expression of all four types of dopamine receptors, with DopEcR being the most abundantly expressed (Figure 1A; Table 1). To validate these analyses, we used T2A-Gal4 knock-in driver lines that mimic dopamine receptor expressions.^35,36 In JO, Dop1R1-Gal4 and DopEcR-Gal4 labeled most cell bodies of JO neurons (about 85.7% and 94.1% respectively, N = 4 and 2 for each driver line; Figures 1B, 1E, S2I, and S2L; Table 2; see Figures S2D and S2H for another driver line) while Dop1R2-Gal4 and Dop2R-Gal4 labeled fewer (about 22.6% and 15.1% respectively, N = 4 and 3 for each driver line; Figures 1C, 1D, S2A–S2C, S2E–S2G, S2J, and S2K; Table 2).

Figure 1.

Dopamine receptor expression patterns in JO neurons

(A) Expression levels of dopamine receptors in JO neurons. Dots represent the normalized expression level of each gene in single JO neurons obtained from the single-nucleus RNA-seq data.²³

(B–I) Expression patterns of dopamine receptors in JO (B–E) and AMMC in the brain (F–I). DsRed marker was expressed by T2A-Gal4 knock-in lines to visualize putative dopamine-receptor-expressing cells (magenta). Dop1R1-Gal4 (B and F), Dop1R2-Gal4 (C and G), Dop2R-Gal4 (D and H), and DopEcR-Gal4 (E and I) (Kondo et al.³⁵) were used. Grayscale images of the top panels are shown in the lower panels. Elav (maker of neuronal cell bodies) immunolabeling was used to label JO neurons (green) within the second antennal segment (black square in the left diagram) (B–E). The rCD2::GFP marker expression was driven by the NV0114 strain, which labels most JO neurons (green) (F–I). JO neurons in each subgroup innervate into five subregions of the AMMC in the brain (yellow dotted lines and the left diagram in F–I). D, dorsal; L, lateral (the same in the following figures). JO, Johnston’s organ; AMMC, antennal mechanosensory and motor center. Scale bar = 10 μm. Genotypes of flies are listed in Table S1 (the same as in the following figures).

See also Figures S1–S3, and Tables 1, 2, and S1.

Table 1.

Dopamine receptor-positive JO neuron ratio (From single nucleus RNA-seq data)

Receptor	Proportion of JO neurons
Dop1R1	29.8%
Dop1R2	14.4%
Dop2R	4.11%
DopEcR	95.2%

Table 2.

Dopamine receptor-positive JO neuron ratio (From immunohistochemistry)

Receptor	AMMC zones	Proportion of JO neurons	Strain	Sample size
Dop1R1	A, B, C, D, E	85.7%	K-strain	4
Dop1R2	A, B	22.6% (Dop1R2-RA)	K-strain	4
	A, B	15.9% (Dop1R2-RA)	D-strain	4
	A, B	26.4% (Dop1R2-RC)	D-strain	3
Dop2R	No expression	15.1%	K-strain	3
	A	25.3%	D-strain	3
DopEcR	A, B, C, D, E	87.6%	D-strain	2
	A, B, C, D, E	94.1%	K-strain	2

Two Gal4 lines (Deng et al.,³⁶ D-strain; Kondo et al.,³⁵ K-strain) were used.

These patterns, together with the RNA-seq database analysis, strongly suggest that all four types of dopamine receptors are expressed in JO neurons but with distinct expression patterns (Table 2): Most JO neurons presumably express DopEcR and, less abundantly, Dop1R2 and Dop2R. Although there was a difference in the ratio of Dop1R1-positive cells between the RNA-seq data and T2A-Gal4 knock-in driver line (Figures 1A and 1B; Tables 1 and 2), both analyses indicate that many JO neurons express Dop1R1.

Next, we estimated which subgroups of JO neurons expressed these receptors. JO neurons in each subgroup project their axons to a distinct zone in the AMMC, specifically zones A to E, which correspond to the JO neuron subgroups A to E, respectively (but see Hampel et al.³⁷ for another subgroup JO-F; Kamikouchi et al.¹⁶). This anatomical segregation of axons facilitates the estimation of labeled JO neuron subgroups by examining the labeling patterns of each driver line in the AMMC zones. Dop1R1-Gal4 and DopEcR-Gal4 signals were detected broadly across AMMC zones A to E (Figures 1F, 1I, S3D, S3H, S3I, and S3L), suggesting that all JO neuron subgroups express these receptors. In contrast, Dop1R2-Gal4 signals were detected selectively in zones A and B, to which JO-A and JO-B neuron axons project, respectively (Figures 1G, S3A, S3B, S3E, S3F, and S3J). This observation suggests that Dop1R2 is expressed predominantly in JO-AB neurons. Dop2R-Gal4 signals were rarely detected (Figures 1H, S3C, S3G, and S3K), likely due to low expression levels. Given previous reports suggesting a lack of efferent innervation to JO, which houses the cell bodies of JO neurons,³⁸ neurotransmission via these dopamine receptors presumably occurs at the axons of JO neurons in the brain. Overall, we concluded that Dop1R1 and DopEcR are expressed across all subgroups, while Dop1R2 is expressed selectively in the auditory subset of JO neurons (i.e., JO-AB).

Dopaminergic neurons synapse with auditory sensory neurons

Dopamine-receptor expression patterns in JO neurons imply direct synaptic connections from dopaminergic neurons. To test this, we utilized a trans-synaptic labeling method, trans-Tango (Figure 2A),³⁹ driven by two independent Gal4 strains: One generated by a promoter-Gal4 insertion and the other using a Gal4 cassette knock-in into the tyrosine hydroxylase (TH) coding sequence,^36,40 both of which presumably label dopaminergic neurons (Figures S4A and S4B). Either strain induced trans-Tango signals in JO neuron cell bodies sparsely (Figures 2B and S4E), with approximately 70 JO neurons showing positive signals on average (Table 3). Given that the D. melanogaster JO contains ∼480 JO neurons,¹⁶ a fraction of JO neurons likely receive direct synaptic input from dopaminergic neurons, a proportion much smaller than the population of neurons expressing dopamine receptors (Figure 1; Tables 1 and 2; See also in Discussion). In the brain, trans-Tango signals were detected sparsely but broadly, spanning the AMMC zones A to E (Figures 2C, S4C, S4D, and S4F). These signals were stronger than those of their parental background flies (i.e., TH-Gal4/w¹¹¹⁸ and w¹¹¹⁸/trans-Tango), validating that the observed trans-Tango signals in the AMMC are above the autofluorescence level (Figure S4G). These findings suggest that a small subset of JO neurons, encompassing all subgroups, presumably receive synaptic input from dopaminergic neurons.

Figure 2.

JO neurons have synaptic connections with dopaminergic neurons

(A) trans-Tango analysis. trans-Tango signals are induced in postsynaptic neurons.³⁹

(B and C) TH-Gal4³⁶ -induced trans-Tango signals in JO (B) and AMMC (C). tdTomato was used to visualize trans-Tango signals, labeled with anti-DsRed antibodies (magenta). In the JO, cell bodies of JO neurons were labeled with anti-Elav antibodies (green; marker of neuronal cell bodies). In the AMMC, synapses were labeled with nc82 antibodies (green). Scale bar = 10 μm.

See also Figure S4 and Tables 3, 4, 5, and S1.

Table 3.

Number of trans-Tango positive cells in all JO neurons

Strain	trans-Tango positive cells (Mean value)	Sample size
F-strain	49.5	8
D-strain	95.1	7

Two Gal4 lines (Deng et al.,³⁶ D-strain; Friggi-Grelin et al.,⁴⁰ F-strain) were used.

We found that Drosophila neuron databases, including Flywire, Flycircuit, and Virtual Fly Brain,^{22,41,42,43,44} contain several dopaminergic neurons innervating the AMMC (Tables 4 and 5), further supporting our conclusion that dopaminergic neurons have synaptic connections to JO-AB neurons in the brain.

Table 4.

Putative dopaminergic neurons that innervate the AMMC in FlyWire

Name	FlyWire Root ID
VES.PVLP.5	720575940627146186
GNG.55	720575940621875821
AVLP.446	720575940613941618
AMMC.GNG.10	720575940636436916
VES.GNG.39	720575940614342038
AVLP.115	720575940618547641
GNG.55	720575940621875821
GNG.782	720575940617290043
AVLP.140	720575940622309726
IPS.SPS.24	720575940613966498
VES.PVLP.4	720575940610941906
IPS.113	720575940630678871
ME.2086	720575940627247802

Table 5.

Putative dopaminergic neurons that innervate zone A or B of the AMMC in FlyCircuit and Virtual Fly Brain

FlyCircuit ID	Virtual Fly Brain ID
TH-F-000025	VFB_00013621
TH-F-400015	VFB_00010484
TH-F-200114	VFB_00013689
TH-F-100058	VFB_00013576
TH-F-100005	VFB_00013322
TH-F-000020	VFB_00013597
TH-F-500006	VFB_00013699
TH-F-200125	VFB_00013772
TH-F-200067	VFB_00013516
TH-F-200029	VFB_00013465
TH-F-100051	VFB_00013561
TH-F-100044	VFB_00013541
TH-F-100031	VFB_00010473
TH-F-100024	VFB_00013427
TH-F-100019	VFB_00013411
TH-F-000100	VFB_00013766
TH-F-000097	VFB_00010495
TH-F-000073	VFB_00013721
TH-F-000051	VFB_00010483b

Dop1R2 expression enhances sound responses of Johnston’s organs-A neurons in unmated females

Next, we aimed to address whether dopamine receptors expressed in JO-AB neurons are involved in their neuronal responses to sound. The contribution of Dop1R1, Dop1R2, or DopEcR expression in JO neurons to the sound-evoked calcium responses was evaluated in unmated (i.e., mature virgin) females, using the following sound stimuli: a 100-Hz pure tone that activates JO-AB strongly,⁴⁵ a pulse song carrying a mean species-specific inter-pulse interval (IPI) (i.e., 35 ms), and a pulse song with a longer IPI (i.e., 75 ms) (Figure 3A). Calcium responses were monitored in the axons of JO neurons, allowing the visualization of the responses of each JO neuron subgroup due to the spatial segregation of their axons (Figure 3B).⁴⁶ Although most JO neurons express Dop1R1 and DopEcR, knockdown of these receptor genes in JO neurons had no significant effect on their calcium responses (Table 6; Figure S5). In contrast, knockdown of Dop1R2, expressed predominantly in JO-AB neurons, significantly reduced the calcium responses in the axons of JO-A neurons but not of JO-B neurons (Figures 3C and 3D; Table 6). These decreased responses were observed for all sound stimuli tested, indicating a general reduction in sound responses (p = 0.0107, 0.0107, and 0.00653 for 100 Hz, 35-ms IPI, and 75-ms IPI, respectively; Welch’s t-tests corrected with Benjamini and Hochberg method; Table 6). To evaluate the knockdown effects, we assessed RNAi effectiveness using quantitative RT-qPCR (Figure S6). Each RNAi construct, when driven by the actin-Gal4 driver, reduced the expression of the target gene by approximately 47–60% (median values of five repeats) compared to the corresponding controls. By fitting the calcium responses to a simple three-component model (Figure S7), we found that Dop1R2 knockdown effect in JO-A neurons can be explained by a reduced calcium increase rate during the stimulus ($r$) and accelerated calcium decrease rate after the stimulus ($d_2$) (Figures 3E–3G; Table 7). Dop1R2 expression thus likely accelerates the calcium increase during the stimulus and slows the decay afterward in JO-A neurons, thereby enhancing both the overall calcium response and its duration. Altogether, our findings suggest that dopamine signals via Dop1R2 increase the response amplitude of JO-A neurons to sound in unmated females.

Figure 3.

Dop1R2 expression increases JO-A neuron responses in unmated females

(A) Experimental setup of calcium imaging. Unmated females were stabilized ventral side up. Artificial courtship songs (35 ms and 75 ms IPIs) and 100-Hz pure tone were played through a loudspeaker.

(B) The region of interest (ROI, outlined in white) for the analysis. Two ROIs were set at the axon bundle of JO-A and JO-B neurons in the AMMC as indicated. A, anterior. Scale bar = 50 μm.

(C and D) Calcium responses of JO-A (C) and JO-B (D) in unmated females with Dop1R2 knockdown in JO neurons. Welch’s t-tests corrected with the Benjamini and Hochberg method; n = 10–11 per genotype. Left, Time traces of raw ΔF/F responses. Thin and bold lines show time traces of the response in each individual and the average of all individuals, respectively. Gray-shaded areas indicate the time window of sound playback. Right, peak calcium responses. Dots and bars show peak responses in each individual and the average of all individuals, respectively (the same in the following figures).

(E–G) Estimated increase and decrease rates in JO-A neurons. Welch’s t-tests corrected with the Benjamini and Hochberg method; n = 10–11 per genotype. A differential equation that incorporates three factors is fitted to the time series data of calcium responses. $r$, estimated increase rate during the stimulus (E); $d_1$, decrease rate during the stimulus (F); $d_2$, decrease rate after the stimulus (G). See also Figures S5–S7, S14 and Tables 6, S1. n.s., p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01.

Table 6.

Statistical summary of calcium imaging using Welch’s t-test corrected with Benjamini and Hochberg method

Figure	Strain	Female state	Sound	Neuron	t-value	p-value	Adjusted p-value	Cohen’s d
3C	F-Gal4, iav-Gal4>Dop1R2 RNAi, GCaMP6f	Unmated	100 Hz	JO-A	2.84	0.0107	0.0107	1.22
			35-ms IPI		3.01	0.00722	0.0107	1.30
			75-ms IPI		3.76	0.00218	0.00653	1.59
3D	F-Gal4, iav-Gal4>Dop1R2 RNAi, GCaMP6f	Unmated	100 Hz	JO-B	0.983	0.338	0.338	0.429
			35-ms IPI	JO-B	1.50	0.150	0.225	0.653
			75-ms IPI	JO-B	2.35	0.0305	0.0915	1.01
4B	F-Gal4, iav-Gal4>Dop1R2 RNAi, GCaMP6f	Mated	100 Hz	JO-A	0.214	0.833	0.956	0.0914
			35-ms IPI	JO-A	−0.169	0.868	0.956	−0.0720
			75-ms IPI	JO-A	0.0561	0.956	0.956	0.0239
S8	F-Gal4, iav-Gal4>Dop1R2 RNAi, GCaMP6f	Mated	100 Hz	JO-B	0.0904	0.929	0.935	0.0385
			35-ms IPI	JO-B	−0.278	0.784	0.935	−0.119
			75-ms IPI	JO-B	0.0828	0.935	0.935	0.0353
S5A	F-Gal4, iav-Gal4>Dop1R1 RNAi, GCaMP6f	Unmated	100 Hz	JO-A	−0.485	0.633	0.633	−0.206
			35-ms IPI	JO-A	−0.818	0.424	0.633	−0.346
			75-ms IPI	JO-A	−0.829	0.418	0.633	−0.352
S5B	F-Gal4, iav-Gal4>Dop1R1 RNAi, GCaMP6f	Unmated	100 Hz	JO-B	1.31	0.206	0.619	0.557
			35-ms IPI	JO-B	0.793	0.440	0.660	0.339
			75-ms IPI	JO-B	0.163	0.873	0.873	0.0690
S5C	F-Gal4, iav-Gal4> DopEcR RNAi, GCaMP6f	Unmated	100 Hz	JO-A	−0.780	0.446	0.669	−0.349
			35-ms IPI	JO-A	0.0783	0.938	0.938	0.0350
			75-ms IPI	JO-A	1.01	0.327	0.669	0.451
S5D	F-Gal4, iav-Gal4> DopEcR RNAi, GCaMP6f	Unmated	100 Hz	JO-B	1.78	0.0923	0.138	0.798
			35-ms IPI	JO-B	0.948	0.356	0.356	0.424
			75-ms IPI	JO-B	1.88	0.0767	0.138	0.840
S9	F-Gal4, iav-Gal4>GCaMP6f, TRiP background	Unmated vs. mated females	100 Hz	JO-A	−0.930	0.0363	0.922	−0.390
			35-ms IPI	JO-A	−0.511	0.615	0.922	−0.212
			75-ms IPI	JO-A	−0.0750	0.941	0.940	−0.900
S10	F-Gal4, iav-Gal4>Dop1R2 RNAi, GCaMP6f	Newly eclosed	100 Hz	JO-A	−0.132	0.897	0.953	−0.0589
			35-ms IPI	JO-A	−0.694	0.497	0.953	−0.310
			75-ms IPI	JO-A	−0.0592	0.953	0.953	−0.0265
S12	F-Gal4, iav-Gal4 GCaMP6f/+	Mated, L-DOPA fed	100 Hz (9.2 mm/s)	JO-A	−0.374	0.711	0.774	−0.142
			100 Hz (4.6 mm/s)	JO-A	−0.290	0.774	0.774	−0.110
			35-ms IPI (9.2 mm/s)	JO-A	−0.550	0.587	0.774	−0.208
			35-ms IPI (4.6 mm/s)	JO-A	−0.641	0.527	0.774	−1.02
4C	F-Gal4, iav-Gal4 GCaMP6f/+	Mated with sex peptide null males	100 Hz	JO-A	1.71	0.0988	0.140	0.604
			35-ms IPI	JO-A	1.52	0.140	0.140	0.538
			75-ms IPI	JO-A	1.73	0.0933	0.140	0.613

Table 7.

Statistical summary of predicted increase ($r$) and decrease rates ($d_1$ and $d_2$)

Figure	parameter	Sound	Neuron	t-value	p-value	Adjusted p-value	Cohen’s d
3E	Increase rate	100 Hz	JO-A	2.67	0.0150	0.0154	1.17
		35-ms IPI	JO-A	3.81	0.00131	0.00219	1.63
		75-ms IPI	JO-A	4.06	0.00146	0.00219	1.70
3F	Decrease rate ($d_1$)	100 Hz	JO-A	−0.351	0.729	0.729	−0.152
		35-ms IPI	JO-A	2.26	0.0366	0.110	0.971
		75-ms IPI	JO-A	1.31	0.207	0.311	0.581
3G	Decrease rate ($d_2$)	100 Hz	JO-A	−3.46	0.00266	0.00290	−1.51
		35-ms IPI	JO-A	−3.42	0.00290	0.00290	−1.49
		75-ms IPI	JO-A	−3.69	0.00155	0.00290	−1.61

Dopamine reflects mating status to adjust Johnston’s organs-A responses

To investigate whether the Dop1R2-mediated modulation of JO-A neurons depends on the mating status of flies, we examined responses in mated females. Mated females were prepared by housing females with males for 12 to 24 h prior to calcium imaging (Figure 4A). This procedure resulted in the fertilization of over 98% of females, as confirmed by quantifying the population of females that produced larvae after housing with wild-type males. Although JO-A neurons in mated females exhibited robust calcium responses to sounds, they lost susceptibility to Dop1R2 expression observed in unmated females, as Dop1R2 knockdown in mated females did not affect the calcium responses of JO-A neurons (p = 0.956 for 100 Hz, 35-ms IPI and 75-ms IPI; Welch’s t-test corrected with Benjamini and Hochberg method; Figure 4B; Table 6; see Figure S8 for JO-B responses). A direct comparison of JO-A responses between unmated and mated females revealed no significant difference (p = 0.922 for 100 Hz, 35-ms IPI and p = 0.940 for 75-ms IPI; Welch’s t-test corrected with Benjamini and Hochberg method; Figure S9; Table 6). This finding suggests that JO-A neurons in unmated females require Dop1R2 expression to reach the response level observed in mated females.

Figure 4.

Dop1R2 expression does not affect JO-A responses in mated females

(A) Experimental scheme for calcium imaging in mated females. Females are maintained with males (blue) for 12 to 24 h before being subjected to the calcium imaging.

(B) Calcium responses in JO-A axons of mated females. Welch’s t-tests corrected with the Benjamini and Hochberg method; n = 11 per genotype.

(C) Calcium responses in JO-A axons of pseudo-unmated females (mated with sex peptide null males). Welch’s t-tests corrected with the Benjamini and Hochberg method; n = 16 per genotype.

See also Figures S8–S13, S14 and Tables 6, 7, 8, 9, S1. n.s., p > 0.05.

Mated females typically exhibit low copulation receptivity,¹⁴ suggesting a link between mating receptivity and Dop1R2-mediated modulation. Since newly eclosed females also exhibit low copulation receptivity,⁴⁷ we tested their JO-A responses with Dop1R2 knockdown. Interestingly, Dop1R2 knockdown did not significantly affect the calcium responses of JO-A neurons (p = 0.953 for 100 Hz, 35-ms IPI and 75-ms IPI; Welch’s t-test corrected with Benjamini and Hochberg method; Figure S10; Table 6). Therefore, together with the results in Figure 3C, these findings suggest that dopamine supports the responses of JO-A neurons via Dop1R2 specifically in females with high mating drive.

The molecular mechanism underlying this status-dependent modulation might be attributed to several steps in dopamine signaling, including receptor amount, dopamine release, and subsequent intracellular signaling pathways. Among these steps, we first examined the amount of receptors located in JO-A axons. We utilized the split-GFP method, where endogenous Dop1R2 is tagged with the C-terminal fragment of the GFP protein (GFP₁₁), allowing GFP to be reconstituted when the N-terminal fragment (GFP_1–10) is expressed in the target neurons.³⁵ The reconstituted GFP signals in JO-A axons showed no significant changes between unmated and mated females (p = 0.974, Welch’s t-test; Figure S11; Tables 8 and 9), suggesting that receptor abundance in JO-A axons is similar between unmated and mated females. We next examined whether the decrease in dopamine release in mated females underlies this status-dependent modulation. Feeding mated females L-DOPA, a dopamine precursor, did not, however, alter the calcium responses of JO-A neurons (p = 0.774 for 100 Hz, 35-ms IPI and 75-ms IPI, Welch’s t-test corrected with Benjamini and Hochberg method; Figure S12; Table 6). These results propose a model in which the modulation of the Dop1R2 signaling pathway, rather than the expression level of Dop1R2 or the amount of dopamine release, influences the mating-status dependent function of Dop1R2 in JO-A neurons (Figure S13A).

Table 8.

Statistical summary of split-GFP analysis using Welch’s t-test

Figure	Condition	t-value	p-value	Cohen’s d
S11B	Unmated vs. mated females	−0.0337	0.974	−0.0138

Table 9.

Statistical summary of split-GFP analysis

Figure	Condition	Presence of GFP11	Mean	Median	Standard error
S11B	Unmated	Yes	526	529	11.6
S11B	Mated	Yes	504	510	22.8
–	Unmated	No (only GFP1-10)	477	484	16.8
–	Mated	No (only GFP1-10)	452	449	14.3

We also quantified the fluorescence of GFP1-10 as GFP1-10 itself has fluorescence.³⁵ Fluorescence without GFP11 was weaker than fluorescence with GFP11.

Finally, we attempted to identify the factors underlying the mating status-dependent, the Dop1R2-mediated regulation of JO-A neurons. One promising candidate is sex peptide (SP), which is transferred from males to females during mating and is known to trigger post-mating responses in females.^48,49,50,51 Analysis of the single-nucleus RNA-seq dataset of fruit flies revealed that the receptor for SP is likely expressed in putative JO-AB neurons (Figure S1B). To test whether SP contributes to changes in Dop1R2 function in mated females, we measured calcium responses in pseudo-unmated females that had mated with SP-null males SP0/Df(3L)Δ130.¹⁴ These pseudo-unmated females exhibited a trend toward reduced JO-A calcium responses when Dop1R2 was knocked down, although the effect did not reach statistical significance (p = 0.140 for 100 Hz, 35-ms IPI and p = 0.940 for 75-ms IPI; Welch’s t-test corrected with Benjamini and Hochberg method; Figure 4C; Table 6). These results suggest that SP may have a partial, albeit minor, role in the loss of Dop1R2 dependency in JO-A responses following mating (Figure 4B).

Dop1R2 expression in Johnston’s organs-AB neurons affects song response behavior

Dopaminergic modulation of JO neuron responses in unmated females might affect their behavioral responses to sounds. We examined whether dopamine receptors expressed in JO neurons influence the song response behavior of unmated females, which can be measured as copulation receptivity to mute males under artificial song exposure (Figure 5A).¹⁸ An F-Gal4 driver line that selectively labels most JO neurons was used for knocking down dopamine receptor expression, although it also labels a few neurons projecting from the legs to the ventral nerve cord (VNC) (Figures 5B and S14A).^9,52

Figure 5.

Dop1R2 knockdown in JO neurons decreases the song response behavior of unmated females

(A) Female copulation assay. An unmated female was paired with a wing-clipped wild-type male in a chamber. An artificial pulse song was played back through a loudspeaker. The experiment was also conducted under a no-sound condition (See Figures S15 and S16).

(B) F-Gal4 driver strain was used to express mCD8::GFP markers (green). nc82 antibodies were used for counter-labeling (magenta). Scale bar = 50 μm.

(C–E) Cumulative copulation rates during song playback. Females with Dop1R1 (C), Dop1R2 (D), or DopEcR (E) knockdown in JO neurons were tested. RMST tests corrected with the Benjamini and Hochberg method; n = 40–48 per genotype.

(F) Evaluation of female song response behavior with Restricted mean time lost (RMTL). F-Gal4 driver strain was used to express each RNAi construct. Dot and vertical line represent the estimated value and standard error, respectively. RMST tests corrected with the Benjamini and Hochberg method; n = 40–48 per genotype.

See also Figures S14–S17 and Tables 10, 11, 12, and S1. n.s., p > 0.05; ∗∗∗, p < 0.001.

Among the three receptor types that are abundantly expressed in JO-AB neurons, only knockdown of Dop1R2 in JO neurons decreased the song response behavior (Figure 5D), while the knockdown of neither Dop1R1 nor DopEcR affected it (Figures 5C and 5E). Comparison of the overall amount of behavioral responses, represented by restricted mean time lost (RMTL),^13,53 also detected a significant reduction in song response behavior in the Dop1R2 knockdown group but not in others, further validating the reduced behavioral responses in Dop1R2 knockdown females (p = $8.54\times {10}^{-2}$, $2.04\times {10}^{-5}$, and 0.312 for Dop1R1, Dop1R2, and DopEcR, respectively; RMST tests corrected with Benjamini and Hochberg method; Figure 5F; Table 10; See Figures S15 and S16 for no-sound controls). Quantifying the walking speeds of Dop1R2 knockdown and control flies revealed similar activity levels (Figure S17; Table 11), ruling out the possibility that the reduced copulation phenotype observed in the Dop1R2 knockdown group was due to aberrant locomotor activity in female flies. These analyses, therefore, suggest that the modulation via Dop1R2 in JO neurons, but not Dop1R1 and DopEcR, affects song response behavior in unmated females.

Table 10.

Statistical summary of female receptivity using RMTL corrected with Benjamini and Hochberg method

Figure	Strain	Sound	RMTL (min)	SE	p-value	Adjusted p-value
5C, S16A	F-Gal4>Dop1R1 RNAi	35-ms IPI	18.2	1.32	0.0650	0.0854
	F-Gal4>TRiP background		21.3	1.02
S15A, S16A	F-Gal4>Dop1R1 RNAi	No sound	0.00532	0.00526	0.0854	0.0854
	F-Gal4>TRiP background		1.01	0.583
5D, S16B	F-Gal4>Dop1R2 RNAi	35-ms IPI	12.6	1.70	1.02E-05	2.04E-05
	F-Gal4>TRiP background		21.2	0.934
S15B, S16B	F-Gal4>Dop1R2 RNAi	No sound	0	0	0.311	0.311
	F-Gal4>TRiP background		0.423	0.418
5E, S16C	F-Gal4>DopEcR RNAi	35-ms IPI	13.3	1.43	0.296	0.312
	F-Gal4>w1118		15.5	1.48
S15C, S16C	F-Gal4>DopEcR RNAi	No sound	0	0	0.312	0.312
	F-Gal4>w1118		0.180	0.178
6A, S16D	JO15>Dop1R2 RNAi	35-ms IPI	15.0	1.76	3.82E-06	6.57E-06
	JO15>TRiP background		24.1	0.892
S15D, S16D	JO15>Dop1R2 RNAi	No sound	0.168	0.166	0.575	0.575
	JO15>TRiP background		0.417	0.413
6B, S16E	R74C10>Dop1R2 RNAi	35-ms IPI	20.0	1.39	0.0192	0.0384
	R74C10>TRiP background		23.9	0.890
S15E, S16E	R74C10>Dop1R2 RNAi	No sound	0.242	0.239	0.212	0.212
	R74C10>TRiP background		0.909	0.478
6C, S16F	R91G04>Dop1R2 RNAi	35-ms IPI	16.3	1.70	0.00817	0.0163
	R91G04>TRiP background		22.2	1.48
S15F, S16F	R91G04>Dop1R2 RNAi	No sound	1.10	0.655	0.107	0.107
	R91G04>TRiP background		3.07	1.03

Table 11.

Statistical summary of locomotor activity using Welch’s t-test

Figure	Condition	t-value	p-value	Cohen’s d
S17	F-Gal4 >Dop1R2 RNAi, TRiP background	0.246	0.813	0.155

To validate the contribution of Dop1R2 expression in JO-AB neurons, we next restricted the knockdown population to these neurons using the JO15 strain (Figure S14B).^16,54 Dop1R2 knockdown in JO-AB neurons reduced song response behavior (Figure 6A), manifested by a significant decrease of RMTL (p = $6.57\times {10}^{-6}$; RMST tests corrected with Benjamini and Hochberg method; Figure 6D; Table 10; See also Figures S15 and S16 for no-sound controls). To compare the changes in receptivity levels between the knocked-down populations, we calculated ΔRMTL as the reduction in RMTL value from the control group resulting from Dop1R2 knockdown.^55,56 ΔRMTL in JO-AB neurons was similar to that in all subgroups of JO neurons (mean ΔRMTL = 8.55 and 9.16, respectively), suggesting that JO-AB neurons primarily contribute to the effect of Dop1R2 expression in JO neurons for this behavioral phenotype (Figures 5F, 6D, and 6E; Table 12).

Figure 6.

Dop1R2 knockdown in JO-AB decreases the song response behavior of unmated females

(A–C) Cumulative copulation rates during song playback. Unmated females with Dop1R2 knockdown were tested. RNAi-mediated knockdown was induced in JO-AB (A), JO-A (B), and JO-B neurons (C). Gal4 driver strains that selectively label JO-AB neurons, JO-A neurons, and JO-B neurons (JO15, R74C10, and R91G04 strains, respectively) were used to express the Dop1R2 RNAi construct. RMST tests corrected with the Benjamini and Hochberg method; n = 40–52 per genotype.

(D) Comparison of female song response behavior using RMTL. RMST tests corrected with the Benjamini and Hochberg method; n = 40–52 per genotype. Dot and vertical line represent the estimated value and standard error, respectively.

(E) Dop1R2 knockdown effect. ΔRMTL represents the difference in RMTL between experimental and control groups in each condition. ΔRMTLs of Dop1R2 knockdown in all subgroups of JO neurons, JO-AB neurons, JO-A neurons, and JO-B neurons are shown. Dot and vertical line represent the estimated value and 95% Confidence Interval, respectively.

See also Figures S14–S16 and Tables 10, 11, 12, and S1. ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Table 12.

ΔRMTL values

Figure	Strain	Sound	ΔRMTL (min)
6E	F-Gal4>Dop1R2 RNAi, TRiP background	35-ms IPI	8.55
6E	JO15>Dop1R2 RNAi, TRiP background	35-ms IPI	9.16
6E	R74C10>Dop1R2 RNAi, TRiP background	35-ms IPI	3.87
6E	R91G04>Dop1R2 RNAi, TRiP background	35-ms IPI	5.94

Finally, we estimated the contribution of Dop1R2 in each JO subgroup using two GAL4 drivers (Figures S14C and S14D),^19,21 which selectively label JO-A and JO-B neurons, respectively. Dop1R2 knockdown in either JO-A or JO-B significantly decreased song response behavior (p = $3.84\times {10}^{-2}$ and $1.63\times {10}^{-2}$ for JO-A and JO-B, respectively; RMST tests corrected with Benjamini and Hochberg method; Figures 6B and 6C; Table 10; See Figures S15 and S16 for no-sound controls). Notably, knockdown in both subgroups of JO neurons induced a greater reduction in behavior (Figure 6A) than knockdown in single subgroups, as depicted in the ΔRMTL (ΔRMTL = 3.87 min for JO-A and 5.94 min for JO-B; Figures 6D and 6E; Table 12). The summation of ΔRMTLs from these two single-subgroup knockdowns seems close to the calculated Dop1R2 knockdown effect in JO-AB neurons (Figure 6E; Table 12), suggesting that Dop1R2 expressed in JO-A and JO-B neurons collaborate to enhance the song response behavior of unmated female flies.

Discussion

Here, we demonstrate a mating state-dependent modulation of auditory processing via the dopaminergic system in the female fruit fly brain. Three types of dopamine receptors are abundantly expressed in JO-AB neurons, some of which presumably mediate direct synaptic transmission from dopaminergic neurons to JO-AB neuron axons. Among these receptor types, Dop1R2 expression significantly contributed to sound-evoked neuronal responses by enhancing sound responses of JO-A neurons in unmated females. The enhanced JO-A responses resulting from Dop1R2 expression were not observed in mated females, suggesting a mating-state-dependent dopaminergic effect on neuronal responses. Behavioral analyses further suggested that Dop1R2 expression in JO-A and JO-B neurons additively enhances the song response behavior of unmated female flies, highlighting the contribution of Dop1R2 in both JO neuron subgroups to maximize song response behaviors. Interestingly, JO-A responses in newly eclosed females were not affected by Dop1R2 knockdown, highlighting mating drive as a key factor for the state-dependent Dop1R2 effect. These findings suggest that female mating drive affects the properties of auditory sensory neurons via dopamine signals, thereby modulating female behavioral responses to the courtship sounds emitted by males for accepting copulation. Mated females exhibited JO-A responses comparable to those of unmated females, regardless of Dop1R2 expression. This suggests that Dop1R2-independent mechanisms may contribute to maintaining the high level of JO-A neuronal activity observed in mated females (Figure S13).

Dopamine signals to Johnston’s organs-AB neurons

Dopamine is involved in state-dependent sensory processing in various animals. In mice, the decrease of dopamine synthesis in olfactory mucosa during starvation contributes to olfactory enhancement.²⁴ In flies, dopamine regulates the sensitivity of gustatory sensory neurons in a hunger-state-dependent manner.²⁵ This study suggests that dopamine modulates auditory sensory neurons depending on females’ mating drive in Drosophila. Although many dopaminergic neuron clusters have been identified in the Drosophila brain (e.g., PAM, PAL, T1, Sb, PPL1, PPL2ab, PPL2c, PPM1, PPM2, and PPM3 clusters),⁵⁷ it is unclear whether any of these clusters affect auditory processing. Several dopaminergic neurons innervating the AMMC, identified in Drosophila neuron databases (Tables 4 and 5), serve as promising candidates for providing dopaminergic modulations to JO-AB neurons.

Most JO neurons express dopamine receptors, but a limited number receive direct synaptic connections from dopaminergic neurons (Figures 1 and 2). In mammals, volume transmission is suggested to be the primary mode in dopamine circuits.⁵⁸ In Drosophila, some studies suggest that dopamine diffuses outside of the postsynaptic region, implying that the volume transmission of dopamine takes place in addition to direct synaptic transmission.^59,60 These two modes of dopaminergic transmission potentially coordinate to modulate the properties of JO neurons. Alternatively, it is also possible that dominant effects occur via volume transmission. Further validation is necessary to understand the synaptic mechanism that mediates this modulation.

Dop1R2 function in Johnston’s organs-AB neurons

We found that the sound-response of JO-A neurons in unmated females is enhanced by Dop1R2, which is classified as a metabotropic D1-like receptor in fruit flies.⁶¹ In the vertebrate hearing system, the function of D1 receptors has been analyzed pharmacologically, revealing its effect in enhancing the neural response in the cochlea: Application of a D1 receptor agonist SKF38393 increased the cochlear compound action potential in guinea pigs,⁶² and microphonic potentials and intracellular calcium levels in zebrafish hair cells.⁶³ Our findings align with these reports, together highlighting the excitatory role of D1 receptors at the first stage of auditory processing. Notably, Dop1R2 expression levels in JO-A axons were stable before and after mating (Figure S11). Given that L-DOPA feeding did not increase calcium responses in mated females (Figure S12), changes in the downstream pathway of Dop1R2 signaling might mediate the specific dopamine effect in unmated female flies (Figure S13A). On the other hand, the robust, Dop1R2-independent calcium responses in mated females may be mediated by alternative mechanisms, which warrant further investigation in future studies.

Previous studies have shown that Drosophila Dop1R2 activates intracellular cyclic AMP (cAMP) and calcium signaling pathways upon dopamine binding.^31,32,64 These pathways may increase the response amplitude of JO-A neurons by modulating various cellular processes.^65,66,67 Our model-based estimation of the calcium dynamics further suggests that Dop1R2 expression accelerates the intracellular calcium increase ($r$) and delays its decrease ($d_2$) in JO-A neurons (Figure 3), together possibly resulting in a net increase in neurotransmitter releases during the stimulus and facilitation of calcium responses to a subsequent stimulus. Identifying the cellular processes responsible for these kinetic changes would enhance our mechanistic understanding of Dop1R2-mediated modulation in JO-A neurons.

Although Dop1R2 knockdown in either JO-A or JO-B neurons reduced the song response behavior of females (Figure 6), it did not decrease the sound-evoked calcium responses in JO-B neurons (Figure 3). A possible explanation is that dopamine signals through Dop1R2 affect the molecular machinery involved in neurotransmission in JO-B neurons without affecting the stimulus-induced calcium responses. Monitoring neurotransmitter release directly from JO-B neurons using genetically encoded neurotransmitter sensors, e.g., GRAB sensor for acetylcholine,⁶⁸ would clarify this possibility, further highlighting the distinct modulatory mechanisms regulating JO-A and JO-B auditory functions.

The role of Johnston’s organs-A neurons during song response behavior

Dop1R2 knockdown in JO-A neurons reduced female song response behaviors (Figure 6), suggesting that JO-A neurons play a crucial role in this behavior. In male Drosophila, the tetanus-toxin-based suppression of JO-A neurons reduced the song-response behavior,¹⁹ together suggesting their involvement in processing song information in both males and females. However, JO-A neurons are not strongly connected with AMMC-B1 neurons, the main secondary auditory neurons that process song information in the fly brain.²¹ Thus, the neural circuit mechanisms underlying their function in song response behavior remain poorly understood, unlike JO-B neurons, which have strong synaptic connections with AMMC-B1 neurons.²¹ Notably, AMMC-A2 neurons, one of the downstream neurons of JO-A neurons,^69,70 have been identified as pulse-song-preferring neurons.⁷ AMMC-A2 neurons further connect to pC2l and vpoIN neurons in two synaptic steps (i.e., hops),⁷ which are also tuned to the pulse song and involved in regulating mating acceptance/rejection in females.^10,71 These findings support the idea that AMMC-A2 neurons serve as another type of song-relaying secondary auditory neuron, alongside AMMC-B1 neurons, in song response behavior. Exploring the interaction between the downstream neural circuits of JO-A neurons, possibly including AMMC-A2 neurons, and the song-relay circuit leading to the mating decision of females will elucidate how the two major types of auditory sensory neurons, JO-A and JO-B, coordinate to affect song response behaviors in flies. The different Dop1R2 knockdown phenotypes in neuronal responses observed between JO-A and JO-B, along with their additive role for Dop1R2-mediated song response behavior of females, are likely related to their distinct roles in song information processing; further study is needed to understand this relationship.

Possible modulations across multiple circuit levels

In the Drosophila olfactory pathway, the mating status of females modulates the neuronal responses to polyamines in both sensory and higher-order neurons.^72,73 While the modulation in olfactory sensory neurons occurs only within 24 h after mating, the effect on higher-order neurons lasts for 3 to 5 days. This suggests that different mechanisms underlie these modulations in the olfactory pathway, orchestrating the state-dependent behaviors.^72,73,74 It is thus intriguing to test whether responses of higher-order auditory neurons in the song-processing pathway, such as AMMC-B1 neurons and pC2 neurons,^8,9 are also modulated according to the mating status of female flies. Such modulation could potentially adjust the properties of the entire song processing circuit, thereby influencing female mating decisions in response to courtship songs.

In the green treefrog, changes in midbrain responses to mating calls after mating specifically occur in the low-frequency range, which covers one of the two key frequency ranges in mating calls that attract females.^1,2 In contrast, Dop1R2-mediated modulation of JO-A neuronal responses in flies occurs for all sound stimuli tested in this study, suggesting a non-specific dopaminergic effect on JO-A neurons. Previous studies in flies have reported that many higher-order auditory neurons in the song-processing pathway are tuned to the sound stimuli carrying species-specific temporal features (e.g., Wang et al.¹⁰; Yamada et al,¹⁸), whereas JO-A neurons respond to a broad range of sounds.¹⁹ Testing whether modulation occurs in higher-order auditory neurons, and if so, whether it happens in a sound-specific manner, may help us better understand the state-dependent modulation of females’ response to courtship songs. Moreover, in addition to the dopaminergic receptors, JO neurons express various types of neurotransmitter receptors (Figure S1A), suggesting that auditory sensory neurons receive multiple neurotransmitters, likely at their axons. How these neurotransmitters modulate JO neuron function is an interesting question for future research.

Mechanisms underlying mating status-dependent modulations

Neural mechanisms underlying mating status-dependent modulations have been studied in several model animals. In mice, signaling factors in seminal fluid that influence female immune response are transferred from males to females during copulation.⁷⁵ While factors equivalent to the Drosophila SP, which significantly affect female behaviors, have not yet been identified,⁷⁶ sex hormones such as estrogen and progesterone, which increase during pregnancy,^77,78 are suggested to contribute to changes in food intake behaviors in mammals.⁷⁹ In female Drosophila, the SP transferred from the male during copulation triggers the post-mating behavioral switch via SP receptors.^48,49,50,51 Neurons modulated by SP alter the gustatory response behaviors of female flies to salt and yeast, suggesting that SP modulates gustatory processing neurons based on mating status.⁸⁰ Additionally, the post-mating increase in polyamine response behaviors is accompanied by an enhanced neuronal response to polyamines, induced by the upregulation of SP receptor expression in chemosensory neurons.⁷² The involvement of a juvenile hormone (JH) in mating status-dependent modulations in olfactory sensory neurons has also been reported,⁸¹ suggesting that multiple cellular mechanisms coordinate to regulate their sensory behaviors.

This study suggests that SP has a minor effect on the mating status-dependent modulation of auditory sensory neurons in female flies, where the Dop1R2 knockdown effect on JO-A neuronal responses was absent in both mated and pseudo-unmated females. Other factors such as sperm, seminal fluid, courtship signals from males, and/or mechanical stimulation during mating potentially link the mating status to the suppression of the Dop1R2 signaling pathway in mated females. Interestingly, our analysis of the single-nucleus RNA-seq dataset suggests the expression of JH receptors in putative JO-AB neurons (Figure S1B), implying that the JH pathway functions in these neurons. Understanding whether the JH systems influence Dop1R2 signaling in JO-AB neurons requires further investigation.

Although mating-status-dependent auditory modulations have been observed in several animal species, the underlying neural mechanisms have yet to be well studied compared to the post-mating modulation of olfactory and gustatory processing. A possible reason may be the paucity of identified phenomena in model animals where molecular genetic tools to monitor and manipulate neuronal functions are well-developed. This study finds that auditory processing is modulated based on the mating state of Drosophila, opening avenues to explore the underlying molecular mechanisms.

Modulations across different sensory modalities

Mating status-dependent modulation of sensory processing in females has been observed in various sensory modalities (Drummond-Barbosa & Spradling⁸²; Hussain et al.⁷²; Walker et al.⁸⁰ for flies; Choo et al.⁸³ for mice; Duffy et al⁸⁴; Kölble et al.⁸⁵ for humans in olfaction and gustation). Although numerous studies have focused on chemosensory systems, such as olfaction and gustation, a few studies have reported modulations in the visual and auditory systems. One such example is found in cichlids (Astatotilapia burtoni), where reproductive females exhibit enhanced visual sensitivity and stronger behavioral responses to male courtship displays compared to non-reproductive females.⁸⁶ Another example is crickets (Teleogryllus oceanicus), where the behavioral latency of females to the male’s calling song becomes longer after mating.⁸⁷ These modulations likely enhance fitness by improving the females’ ability to detect males during periods when they can produce offspring. Given that animal behaviors heavily depend on multimodal sensory integration, coordinated modulation across different sensory modalities is likely essential to support behavioral changes associated with mating states. The dopaminergic system likely plays a pivotal role in this cross-modal regulation, facilitating adaptive behaviors.

Limitations of the study

Our research revealed that Dop1R2 is involved in mating-status dependent modulation in auditory sensory neurons in Drosophila. However, this study has the following limitations: (1) The ratio of Dop1R1-positive cells differs between the RNA-seq data and the expression patterns of T2A-Gal4 knock-in driver lines. This discrepancy may stem from differences between transcriptional (RNA-seq) and translational (marker expression) levels. It may also be influenced by experimental conditions (e.g., sex, rearing environment), as the RNA-seq data were obtained from a mixed population of 5-day-old males and females,²³ while the marker-expression patterns in this study were observed exclusively in 5–10 days old females. The expression level of Dop1R1 may be dynamically regulated by these factors, but further investigation is needed to clarify this. (2) It is reported that the trans-Tango system has caveats, leading to false positives and negatives.^39,88 Although we followed the protocol to suppress the false negatives by raising flies at 18°C, we cannot exclude the possible detection of false signals. (3) DopEcR is highly expressed in JO neurons; however, RNAi-mediated knockdown did not result in any detectable phenotypic changes. This suggests the possibility of synergistic or antagonistic interactions between DopEcR and Dop1R2. To explore these potential interactions, further studies using double-knockout experiments are warranted. (4) Theoretically, GFP_1-10 is expressed excessively due to the amplification effect of the Gal4/UAS system. The limiting factor for reconstituted GFP expression is the level of the GFP₁₁ fragment, which reflects the endogenous expression of Dop1R2. Therefore, Dop1R2 expression can be quantified by measuring GFP fluorescence. However, to our knowledge, there is no prior publication validating the sensitivity of the split-GFP method in detecting expression level differences.

Resource availability

Lead contact

Further information and requests for resources and regents can be directed to the lead contact, Azusa Kamikouchi (kamikouchi@bio.nagoya-u.ac.jp).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Original data and scripts have been deposited at Mendeley Data: (https://data.mendeley.com/datasets/29hpr3nrrv/1).
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We thank Dr. Yoichi Oda, Dr. Matthew P. Su, Dr. Hiroshi Ishimoto, Dr. Ryo Hoshino, Tai-Ting Lee, and Dr. Toshiharu Ichinose for discussions; Dr. Takuro S. Ohashi, Daisuke Takaichi, and Dr. Ryosuke F. Takeuchi for Python and MATLAB scripts; Dr. Shu Kondo, Dr. Hiromu Tanimoto, Dr. Shun Hiramatsu, Dr. Kuniaki Saito, Dr. Ryusuke Niwa, Drosophila Stock Center, and Vienna Drosophila Resource Center for providing fly strains; Developmental Studies Hybridoma Bank for antibodies. This study was supported by MEXT KAKENHI Grants-in-Aid for Scientific Research (B) (Grant JP20H03355 to A.K.), Grant-in-Aid for Transformative Research Areas (A) “iPlasticity” (Grant JP23H04228 to A.K.), Hierarchical Bio-Navigation (Grant JP22H05650 to R.T.), Materia-Mind (Grant JP24H02200 to A.K.), and Dynamic Brain (Grant JP25H02496 to A.K.), Grant-in-Aid for JSPS Fellows (Grant JP24KJ1282 to H.Y.), The Graduate Program of Transformative Chem-Bio Research, Nagoya University, Japan (to H.Y.), and JST FOREST (Grant JPMJFR2147 to A.K.), Japan.

Author contributions

Conceptualization, H.Y, M.H., and A.K.; methodology, H.Y., M.H., S.Y., S. Iwanami, S. Iwami, R.T., Y.I., and A.K.; investigation, H.Y., M.H., and A.K.; resources, A.K.; Writing, H.Y. and A.K.; funding acquisition, H.Y., R.T., and A.K.; supervision, R.T., Y.I., and A.K.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT in order to improve the language of the article. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-elav (Mouse monoclonal)	Developmental Studies Hybridoma bank	Cat#Elav-9F8A9; RRID: AB_528217 (1:250)
anti-GFP (Rat monoclonal)	NACARAI TESQUE, INC	Cat#GF090R; RRID: AB_2314545 1:1000)
anti-DsRed (Rabbit polyclonal)	Takara Bio	Cat#632496; RRID: AB_10013483 (1:1000)
nc82-s (Mouse monoclonal; supernatant)	Developmental Studies Hybridoma Bank	RRID: AB_2314866 (1:20)
anti-rat-Alexa 488 (Goat polyclonal)	Thermo Fisher Scientific	Cat#A-11006; RRID: AB_2534074 (1:300)
anti-rabbit-Alexa 555 (Goat polyclonal)	Thermo Fisher Scientific	Cat#A-21429; RRID: AB_2535850 (1:300)
anti-mouse-Alexa 647 (Goat polyclonal)	Thermo Fisher Scientific	Cat#A-21236; RRID: AB_2535805 (1:300)
Chemicals, peptides, and recombinant proteins
L-DOPA	Tokyo Chemical Industry Co., Ltd	#D0600
HCl	NACALAI TESQUE, INC.	#37314-15
glucose	Sanei Toka Co., Ltd	Glu Final
agarose	NIPPON GENE CO., LTD.	#312-01193
ethanol	KANTO CHEMICAL CO., INC.	#14033-81 or #14032-08
PBS	Takara Bio Inc.	#T9181
4% paraformaldehyde/PBS	Wako Pure Chemical Industries, Ltd	#163-20145
formaldehyde	Thermo Fisher Scientific	#28908
glyoxal	FUJIFILM Wako Pure Chemical Corporation	#077-05291
methanol	Sigma-Aldrich	#34885
Triton X-100	Sigma-Aldrich	#X100-500 ML
Glycerol	Wako Pure Chemical Industries, Ltd	#075-04751
RNA later	Sigma-Aldrich	#R0901
chloroform	KANTO CHEMICAL CO., INC.	#07278-00
2-propanol	KANTO CHEMICAL CO., INC.	#32435-00
OMNIPAQUE 350 INJECTION	GE Healthcare Pharma	#7219415H4062
Critical commercial assays
RNAiso	Takara Bio Inc.	#9109
ReverTra AceTMqPCR RT Master Mix with gDNA Remover	TOYOBO	#FSQ-301
THUNDERBIRD SYBRTM qPCR Mix	TOYOBO	#QPS-201
Deposited data
Raw and analyzed data	This paper; Mendeley Dataset	Mendeley Data: https://data.mendeley.com/datasets/29hpr3nrrv/1, https://doi.org/10.17632/29hpr3nrrv.1
Experimental models: Organisms/strains
Drosophila: Canton-S	Gift from Dr. K Ito	Hotta-lab strain
Drosophila: F-GAL4	Gift from Dr. Changsoo Kim	RRID: BDSC_24903 (Backcrossed by Canton-S 11 times for female copulation assay)
Drosophila: iav-GAL4	Bloomington Drosophila Stock Center	RRID: BDSC_52273
Drosophila: DopEcR-Gal4	Gift from Dr. Deng	RRID: BDSC_84629
Drosophila: Dop1R2-GAL4-RA	Gift from Dr. Deng	RRID: BDSC_84710
Drosophila: Dop1R2-GAL4-RC	Gift from Dr. Deng	RRID: BDSC_84711
Drosophila: Dop2R-GAL4	Gift from Dr. Deng	RRID: BDSC_84628
Drosophila: DopEcR-Gal4	Gift from Dr. Kondo	NA
Drosophila: Dop1R1-GAL4	Gift from Dr. Kondo	NA
Drosophila: Dop1R2-GAL4	Gift from Dr. Kondo	NA
Drosophila: Dop2R-GAL4	Gift from Dr. Kondo	NA
Drosophila: QUAS-mtdTomato, trans-Tango	Bloomington Drosophila Stock Center	RRID: BDSC_77124
Drosophila: UAS-IVS-mCD8::GFP	Bloomington Drosophila Stock Center	RRID: BDSC_32187
Drosophila: NV0116/SM1; UAS-DsRed (C5), lexAop-rCD2::GFP/TM6b	Gift from Dr. Tzumin Lee and Dr. Kei Ito	NA
Drosophila: TH-GAL4	Bloomington Drosophila Stock Center	RRID: BDSC_8848
Drosophila: TH-GAL4	Bloomington Drosophila Stock Center	RRID: BDSC_86289
Drosophila: JO15-Gal4	Bloomington Drosophila Stock Center	RRID: BDSC_6753
Drosophila: R74C10-Gal4	Bloomington Drosophila Stock Center	RRID: BDSC_39848
Drosophila: R91G04-Gal4	Bloomington Drosophila Stock Center	RRID: BDSC_40588
Drosophila: UAS-Dop1R1 RNAi	Bloomington Drosophila Stock Center	RRID: BDSC_3176589
Drosophila: UAS-Dop1R2 RNAi	Bloomington Drosophila Stock Center	RRID: BDSC_2601889
Drosophila: UAS-DopEcR RNAi	Vienna Drosophila Resource Center	VDRC: 10349425
Drosophila: TRiP background line	Bloomington Drosophila Stock Center	RRID: BDSC_36303
Drosophila: w1118	Vienna Drosophila Resource Center	VDRC: 60000
Drosophila: UAS-GCaMP6f	Bloomington Drosophila Stock Center	RRID: BDSC_42747
Drosophila: P{20XUAS-IVS-mCD8::GFP}attP2	Bloomington Drosophila Stock Center	RRID: BDSC_32194
Drosophila: y[1]w[1118]; Dop1R2-RA-7xGFP11/TM6C(Sb, Tb)	Gift from Shu Kondo	NA
Drosophila: y[1]cho[2]v[1]; attP40{UAS-GFP1-10}	Gift from Shu Kondo	NA
Drosophila: 10XUAS-IVS-mCD8::RFP	Bloomington Drosophila Stock Center	RRID: BDSC_32219
SP0	Gift from Mariana Wolfner	RRID: BDSC_77892
Df (3L)Δ130	Gift from Mariana Wolfner	NA
actin-Gal4	Bloomington Drosophila Stock Center	RRID: BDSC_4414
Software and algorithms
Fiji (version: 1.54f)	ImageJ90	RRID: SCR_002285, https://imagej.net/ij/
Audacity (version: 2.4.2)	Audacity Team	RRID: SCR_007198, https://www.audacityteam.org/
RStudio (version: 2024.04.2 + 764)	Posit, The open source data science company	RRID: SCR_000432, https://posit.co/download/rstudio-desktop/
Arduino IDE (version: 1.6.9)	ARDUINO	https://www.arduino.cc/en/software
Processing (version: 4.0b2)	Processing Foundation	https://processing.org
Monolix2023R1	Lixoft	https://lixoft.com/
R (version: 4.3.0)	R Core Team	RRID: SCR_001905, https://www.r-project.org/
Python (version: 3.9.7 or 3.9.13)	The Python Software Foundation	RRID: SCR_008394, https://www.python.org/
MATLAB (R2024b)	MathWorks	RRID: SCR_001622, http://www.mathworks.com/products/matlab/
Other
Virtual Fly Brain	Virtual Fly Brain Project	RRID:SCR_004229, https://www.virtualflybrain.org/
FlyCircuit	Chiang et al.41	RRID:SCR_006375,https://www.flycircuit.tw/
FlyWire (v783)	Dorkenwald et al.43; Schlegel et al.44; Eckstein, Bates et al.22	RRID:SCR_019205,https://flywire.ai/

Experimental model and study participant details

Experimental animals

Fruit flies were raised on yeast-based food at 25°C and 40–60% relative humidity under a 12 h light/12 h dark (12 h L/D) cycle unless otherwise noted. Transgenic fly strains were used for the Gal4/UAS⁹¹ and LexA/lexAop^92,93 systems. Canton-S was used as a wild-type D. melanogaster strain. For the transsynaptic analysis (Figure 2), we used two types of Gal4 driver strains: (1) promoter-Gal4 line, in which a promoter-Gal4 construct was inserted into the genome using transposase (Figure S4A),⁴⁰ and (2) a knock-in line where Gal4 was fused to the C-terminus of the endogenous tyrosine hydroxylase (TH) protein (Figure S4B).³⁶ Detailed information on the Drosophila strains is described in the key resources table. Genotypes and sexes of flies used for each experiment are listed in Table S1.

To prepare unmated and mated females for calcium imaging experiments and split-GFP analysis, adult females collected within 8 h after eclosion were raised in groups of 3–10. Unmated females were kept in vials for 5–7 days before being used for the experiments. Mated and pseudo-mated females were kept in vials for 4–6 days, after which three females each were transferred to a food vial containing 3–10 unmated wild-type or SP null (SP0/Df(3L)Δ130) males, where they were kept for 12–24 h. This procedure resulted in fertilization of over 98% of females, as confirmed by quantifying the population of females that produced larvae after mating with wild-type males. To prepare newly eclosed females, adult females collected within 6 h after eclosion were raised in groups of 3–10 for 20–30 h.

For the L-DOPA feeding experiment, adult females collected within 8 h after eclosion were raised in groups of 3–10 for 3–5 days. Females were starved for 15–18 h by placing in a vial containing a paper towel moistened with water. The females were then paired with wild-type males as described above and placed in a vial containing food with L-DOPA (0.3% (w/v) L-DOPA (3-(3,4-Dihydroxyphenyl)-L-alanine, #D0600, Tokyo Chemical Industry Co., Ltd), 0.073% (w/v) HCl (#37314-15, NACALAI TESQUE, INC.), 9% (w/v) glucose (Glu Final, Sanei Toka Co., Ltd), 5% (w/v) Extract of yeast (#103753, Merck KGaA) and 1% (w/v) agarose (#312–01193, NIPPON GENE CO., LTD.) in water). For controls, food without L-DOPA was used. Flies were kept for 12–24 h and used for calcium imaging.

At the time of calcium imaging, unmated and mated females were 5–7 days old after eclosion, while newly eclosed females were 26–36 h after eclosion.

Our study focused on female Drosophila. It remains to be determined whether our findings are similarly applicable to male Drosophila. Future studies are warranted to investigate analogous phenomena in males and to assess potential sex-specific differences.

Method details

Immunohistochemistry

Immunohistochemistry was performed as described previously with minor modifications.¹⁸ Adult females 5–10 days after eclosion were ice-anesthetized, washed in 99.5% ethanol (14033-81, KANTO CHEMICAL CO., INC.) for about 30 s to remove the oil on the body surface, and dissected in phosphate-buffered saline to obtain brains and antennae (PBS; #T9181, Takara). After dissection, these tissues were fixed for 1 h in 4% paraformaldehyde/PBS (163–20145, Wako Pure Chemical Industries, Ltd) on ice or 30 min in glyoxal solution [3% (w/v) formaldehyde (#28908, Thermo Fisher Scientific), 1% (w/v) glyoxal (#077–05291, FUJIFILM Wako Pure Chemical Corporation), 0.1% (w/v) methanol (#34885, Sigma-Aldrich) in PBS] at room temperature.^94,95 Samples were then washed twice with PBS containing 0.5% Triton X-100 (PBT) (#X100-500 ML, Sigma-Aldrich), and incubated overnight in PBT at 4°C. Subsequently, samples were incubated with primary and secondary antibodies for three days each at 4°C. The antibodies are listed in key resources table. After washing the samples with 0.5% PBT, samples were transparentized in 50% and 80% Glycerol (075–04751, Wako Pure Chemical Industries, Ltd) sequentially.

Confocal microscopy and image processing

Serial optical sections of brains and antennae were obtained at 0.50 to 0.84 μm intervals with a resolution of 512 × 512 pixels using an FLUOVIEW FV1200, FV1000D, or FV3000 laser-scanning confocal microscope (Olympus) equipped with a silicone-oil-immersion lens UPLSAPO 30xS (NA = 1.05; Olympus) or UPLSAPO 60xS (NA = 1.30; Olympus). Image size, brightness, and contrast were adjusted using Fiji (version: 2.0.0-rc-69/1.52p).⁹⁰

Analysis of expression of T2A-Gal4 knock-in driver lines in the antenna

To quantify the ratio of Gal4-positive JO neurons, we randomly extracted three optical slices from each serial-image dataset of JO. In each section, we counted the number of the Gal4 > tdTomato signals overlapping with Elav signals to examine the percentage of JO neurons expressing Gal4 signals using the “Cell Counter” plugin in Fiji (version: 2.0.0-rc-69/1.52p).

Gene expression analyses using single-nucleus RNA-seq data

Fly Cell Atlas (www.flycellatlas.org), a platform for single-nucleus RNA sequence dataset of fruit flies,²³ was used for gene expression analyses. To estimate the proportion of JO neurons that express neurotransmitter receptors, the antennal dataset (antenna.h5ad; collected from male and female antennae)²³ was loaded using SeuratDisk (ver. 0.0.0.9020) for R (ver. 4.3.0). The cell cluster previously annotated as “Johnston Organ Neuron”²³ was extracted using the subset function of the Seurat package (ver. 4.3.0.1). The proportion of JO neurons that express each dopamine receptor was calculated using the WhichCells function of Seurat package (ver. 4.3.0.1). To estimate the expression level of SP and JH receptors in JO-AB neurons, we extracted Dop1R2-expressing cells from annotated “Johnston Organ Neuron”²³ as Dop1R2-T2A-Gal4 signals are exclusively expressed in JO-AB neurons (Figure 1). Following a previous study,²³ the normalized expression level was calculated as follows:

$$\mathit{log}\left(\frac{readcountofeachgene}{totalcountsforthecell}\times 10,000+1\right)$$

trans-Tango analysis

For trans-Tango analysis, we raised flies for 13–15 days after eclosion at 18°C in accordance with a previous study reporting that rearing at 18°C typically yields the best S/N ratio for trans-Tango signals.³⁹ The number of TH>trans-Tango positive JO neurons was then quantified by counting tdTomato signals that represent TH-Gal4 > trans-Tango signals in JO neuron somata across all confocal optical sections through the second antennal segment using the “Cell Counter” plugin in Fiji (version: 2.0.0-rc-69/1.52p). Signals in the external sensory neurons, located underneath the antennal external sensory bristles, were excluded from the count. To validate that the trans-Tango signals in the AMMC were above the autofluorescence level, we acquired confocal images with the same scan settings for the laser power and the detector sensitivity across samples and genotypes (Figure S4G).

Detecting dopaminergic neurons innervating the AMMC

To extract dopaminergic neurons innervating the AMMC, we used both EM and LM database, Flywire, FlyCircuit, and Virtual Fly Brain.^{22,41,42,43,44,96} In the EM database, we searched for dopaminergic neurons using codex, the explorer of the Flywire brain dataset (v783), with the following conditions: input_neuropils = = AMMC_L or AMMC_R, nt_type = = DA. In the LM database, we searched for dopaminergic neurons using Text-based Search in Flycircuit with the following conditions: Driver = = TH-Gal4, Innervation sites = = AMMC or ammc. From the hit neurons, we extracted dopaminergic neurons innervating AMMC zone A or zone B. We also observed the morphology of the neurons in Virtual Fly Brain to validate this extraction.

RT-qPCR

Adult females 5–7 days after eclosion were ice-anesthetized, washed in 99.5% ethanol (#14032-08, KANTO CHEMICAL CO., INC.) for about 30 s to remove the oil on the body surface, and dissected in RNA later (#R0901 Sigma-Aldrich) with whole head tissue being collected. Dissected samples were homogenized (Motorized tissue grinder, #12-141-361, Fisher Scientific) in RNAiso (#9109, Takara Bio Inc.). Additional RNAiso was added to the lysed samples to make up a final volume of 1 mL. 200 μL chloroform (#07278-00, KANTO CHEMICAL CO., INC.) was added to each sample and mixed well before incubating at room temperature for 2–15 min. Samples were then centrifuged at 12,000 × g for 15 min at 4°C before the addition of 0.5 mL 2-propanol (#32435-00, KANTO CHEMICAL CO., INC.). Samples were stored in −20°C for 1 h followed by centrifugation at 12,000 × g for 10 min at 4°C, the supernatant was discarded, and RNA pellet was washed with 0.5 mL of 99.5% ethanol. Samples were centrifuged at 12,000 × g at 4°C for 5 min. RNA pellet was then washed with 0.5 mL of 75% ethanol and then centrifuged again at 12,000 × g at 4°C for 5 min before the RNA pellet was dissolved in Nuclease Free Water (ReverTra AceTMqPCR RT Master Mix with gDNA Remover, FSQ-301, TOYOBO). RNA sample quality was confirmed using a Nanodrop and stored in −80°C until use. RNA was reverse transcribed (ReverTra AceTMqPCR RT Master Mix with gDNA Remover, FSQ-301, TOYOBO) prior to conducting qPCR (THUNDERBIRD SYBRTM qPCR Mix, QPS-201, TOYOBO). The housekeeping gene ribosomal protein 49 (RP49) was used as the internal control. Dopamine receptor primers were designed using the Primer BLAST (Table S3). For each qPCR 96-well plate (Bio-Rad), three technical repeats for each sample and primer were tested, with the median of these technical values being used for analyses. In total, five biological repeats were conducted.

Sound file preparation

Artificial pulse songs and 100-Hz pure tone were generated using Audacity (The Audacity Team, version 2.0). For artificial pulse songs, we generated two song types, carrying IPIs of 35 ms and 75 ms, respectively. Both songs are comprised of pulses with a carrier frequency of 167 Hz.¹⁸ Sound amplitude was adjusted using NR23158-000 microphone (Knowles, NR Series) to 11.1–13.1 mm/s (for pulse songs) or 15.5–18.4 mm/s (for 100-Hz pure tone) particle velocity for the Ca²⁺ imaging experiment and 8.08 mm/s particle velocity for the copulation assay at the peak-to-peak amplitude. For the calcium imaging of the L-DOPA fed females, the sound amplitudes for both pulse songs and 100-Hz pure tone were adjusted to 4.6 and 9.2 mm/s.

Ca²⁺ imaging - Data acquisition

Calcium imaging was performed on female flies as described previously with minor modifications.^13,18 F-Gal4 and iav-Gal4, fly strains that label most JO neurons spanning all subgroups, were combined to express a Ca²⁺ indicator GCaMP6f⁹⁷ and each dopamine receptor-knockdown construct simultaneously. A group of 3–10 female flies expressing both transgenes were collected in a food vial within 8 h after eclosion and maintained at 25°C and 40–60% relative humidity under a 12 h L/D cycle until the experiment was conducted. For the imaging, unmated or mated flies five to seven days after eclosion were used; they were anesthetized on ice and stabilized ventral side up onto an imaging plate using silicon grease (SH 44M, Toray, Tokyo, Japan). The mouthparts were removed from the fly head using fine forceps to open a window to monitor GCaMP fluorescence from the brain. A drop of a saline solution (108 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 8.2 mM MgCl₂, 4 mM NaHCO₃, 1 mM NaH₂PO₄, 5 mM trehalose, 10 mM sucrose, and 5 mM HEPES, pH 7.5, 265 mOsm)⁹⁸ was added to prevent dehydration. A 35 mm dish (35-mm lummox-film bottom dish, SARDTEDTAG & Co) with a hole at the bottom was set on the window opened at the fly mouthparts and filled with 2 mL of saline solution.

GCaMP fluorescence was detected using a fluorescent microscope (Axio Imager.A2, Carl Zeiss, Oberkochen, Germany) equipped with a water-immersion 20× objective lens (W N-Achroplan, numerical aperture = 0.5; Carl Zeiss), a spinning disk confocal head CSU-W1 (Yokogawa, Tokyo, Japan), a dichroic mirror (405/488/561/640 Di01-T405/488/568/647-13 × 15 × 0.5, Semrock, Rochester, NY), and a bandpass filter (FF01-528/38-25, Semrock). An OBIS 488 LS laser (Coherent Technologies, California, US) was used to excite GCaMP6f at 488 nm. The fluorescent images were captured at 10 fps with an exposure time of 100 ms for 10 s using an EM-CCD camera (iXon Ultra 897; Oxford Instruments, Abingdon, UK) at a resolution of 512 × 512 pixels in water-cooled mode. The imaging setup was operated using Micro-Manager 2.0 (version 2.0).

For sound stimulation, a loudspeaker (Fostex, FF225WK; Foster Electric) equipped with an amplifier (TB10A; Fosi Audio) was located 11 cm from the antenna of the fly. The sound files were played through Arduino Uno R3, using the Adafruit Wave Shield for Arduino Kit v1.1. The onset of fluorescent image recording and sound playback were synchronized using Processing (version: 4.0b2), Arduino IDE (version: 1.6.9), and Arduino UNO R3. Sound playback began 2 s after the onset of calcium imaging.

Each individual was randomly presented with the following combinations of sounds four times: 100 Hz, 35 ms IPI and 75 ms IPI, each for 2 s 100-Hz pure tone served as a control sound as it evokes strong responses in both JO-A and JO-B neurons.⁴⁵

Ca²⁺ imaging - Data analysis

Imaging data were analyzed using Fiji (version: 2.0.0-rc-69/1.52p) and R (version 4.2.1) as described previously with minor modifications.^13,18 Brain movements were corrected using a custom Python script (version: 3.9.7)⁹⁹ or the NoRMCorre algorithm¹⁰⁰ with MATLAB, which was manually validated by reviewing all the corrected images. To measure fluorescence intensity, regions of interest (ROIs) were defined using Fiji (version: 2.0.0-rc-69/1.52p). ROIs were placed to cover the axonal region of either JO-A or JO-B neurons.

To quantify the calcium responses, the relative fluorescence change ($\Delta F/F$) of GCaMP6f was normalized as follows: $\Delta F/F$ = ($F_n-F_{Base}$) $/F_{Base}$, where the mean fluorescence intensity within an ROI at each time point n ($F_n$) was divided by the mean fluorescence intensity of the Ca²⁺ response during the 1.9 s preceding the stimulus onset ($F_{Base}$).¹⁸ A three-frame moving average of the $\Delta F/F$ values was used for the time trace graphs and analysis. To correct for fluorescence bleaching, fluorescence from frames 1 to 19 and frames 61 to 100 was used to fit a linear model using the stats package (version 4.3.2) in R.¹³ Fluorescence intensity was corrected by subtracting its fitting function.

Modeling calcium response dynamics in sound stimulus

We developed the following mathematical model to describe the dynamics of calcium response before ($t<2$), during ($2\le t\le 4$), and after ($4\le t$) sound stimulation, starting from the onset of calcium imaging ($t=0$):

$$\frac{dR\left(t\right)}{dt}=0\left(t<2\right)$$ (Equation 1)

$$\frac{dR\left(t\right)}{dt}={r-d}_{1}R\left(t\right)\left(2\le t<4\right)$$ (Equation 2)

$$\frac{dR\left(t\right)}{dt}={-d}_{2}R\left(t\right)\left(4\le t\right)$$ (Equation 3)

where the sound stimulation started at $t=2$. The variable $R\left(t\right)$ is the normalized GCaMP6f fluorescence value ($\Delta F/F$) at time $t$ in each fly. The parameters $r$, $d_1$, and $d_2$ are the increase rate of calcium response during the stimulus and the decrease rates of calcium response during and after the stimulus, respectively. To estimate these parameters in each fly, we used a nonlinear mixed-effect modeling approach, which includes fixed effect (constant across flies) and random effect (different in flies) for all parameters to consider the variability of the responses among flies. Each individual parameter of fly $i$, ${\theta }_i\left(=\theta \times e^{{\pi }_i}\right)$, is decomposed as fixed effect $\left(\theta \right)$ and random effect ($e^{{\pi }_i}$, where ${\pi }_i$ is assumed to be drawn from a normal distribution, $\mathcal{N}\left(0,\mathrm{\Omega }\right)$. Parameters of fixed effect and the standard deviation of random effects, $\theta$ and $\mathrm{\Omega }$ respectively, were estimated with the stochastic approximation Expectation/Maximization (SAEM) algorithm. Then, parameters of individual flies considering random effects were estimated as the mode of their conditional distributions using the empirical Bayes method. Monolix 2023R1 [Monolix 2023R1, Lixoft SAS, a Simulations Plus company] was used as a tool for maximum likelihood estimation applying a nonlinear mixed-effect model to fit the calcium response data.

Quantification of Split-GFP images – image acquisition

Groups of 3–10 female flies were collected in each food vial within 8 h after eclosion and maintained at 25°C and 40–60% relative humidity under a 12 h L/D cycle for seven days. Brains were mounted on a glass slide immediately after dissection, and incubated in SeeDB2G¹⁰¹ (OMNIPAQUE 350 INJECTION, GE Healthcare Pharma, 7219415H4062) for at least 30 min to increase transparency. Serial optical sections of brains were then obtained using an FLUOVIEW FV1200 laser-scanning confocal microscope (Olympus) equipped with a silicone-oil-immersion lens UPLSAPO 30xS (NA = 1.05; Olympus) at 0.84 μm intervals with a resolution of 512 × 512 pixels. The same scan settings for the laser power and the detector sensitivity (HV; high voltage) were used across samples (GFP, Laser power: 2.6%, HV: 560; RFP, Laser power: 0.7%, HV: 560).

Quantification of Split-GFP images – Analysis

For quantification, one side of the AMMC was used for each individual. The confocal optical images were cropped to 300 × 300 pixels around the AMMC using Fiji (version: 2.0.0-rc-69/1.52p) and Python (version: 3.9.13). To extract the axonal region of JO-A neurons, we binarized RFP images using Imbinarize function for MATLAB. The threshold set by Otsu’s method¹⁰² in a randomly selected sample was applied to all samples. The GFP intensity was divided by the RFP intensity for normalization in each pixel. Normalized GFP intensity values were averaged in each individual and used for statistical analysis.

Female copulation assay

Female song response behavior was evaluated using a copulation assay.¹⁸ F-Gal4,⁵² JO15-Gal4,⁵⁴ R74C10-Gal4,¹⁹ and R91G04-Gal4²¹ were used as driver lines to knockdown the expression of each dopamine receptor in all subgroups of JO neurons, JO-AB neurons, JO-A neurons, and JO-B neurons, respectively (key resources table and Table S1). Wild-type males and genetically manipulated females (knockdown and control groups) were used. Both sexes of flies were collected under ice anesthesia within 8 h after eclosion. Males were wing-clipped and kept in groups of 3–10 until experiments were conducted. Females were kept in groups of 3–10. These flies were maintained at 25°C and 40–60% relative humidity under a 12 h L/D cycle. Males and females five to seven days after eclosion were used for the assay.

The assays were conducted at 25°C and 40–60% relative humidity during zeitgeber time = 0 to 4. A pair of male and female was gently aspirated into a chamber (1.5 cm diameter, 3 mm deep) without anesthesia. Fly behaviors were recorded for 35 min at 15 fps with a web camera (Logiteck Logicool , model: HD Webcam C270). During recording, the artificial courtship song carrying a 35-ms IPI were played back from a loudspeaker (Datio Voice AR-10N, Tokyo Cone Paper MFG. Co. Ltd) via digital amplifier (Lepy LP-2020A; Lepy LP-2024A; Lepy LP-V3S), in which a cycle of 1 s of pulse sound followed by 2 s silence was repeatedly broadcasted for the 35-min observation period.¹⁸ The assays were also performed without sound playback as the no-sound condition.

After the assay, the copulation onset in each pair was identified manually. Copulation was judged based on the male being on top of the female for at least 5 min and the female decreasing her spontaneous behavior.^18,103

Locomotor activity analysis

To analyze the locomotor activities of F-Gal4>Dop1R2 RNAi females and their control flies, data from the top five pairs with the earliest mating initiation times in each group were used. We tracked the positions of the females during the initial phase of the copulation assay. Using the “Manual Tracking” plugin in Fiji (version: 2.0.0-rc-69/1.52p), we traced their positions over 300 frames following the moment when males oriented and began chasing the females. Total walking distances and walking speeds were calculated, and a ten-frame moving average of the walking speed was applied for plotting.

Quantification and statistical analysis

For all qPCR experiments, median RP49 values were used to calculate DeltaCt (ΔCt) values for each gene. DeltaDeltaCt (ΔΔCt) values were then calculated within a repeat using control (actin-Gal4 > TRiP BG attP2 for Dop1R1 and Dop1R2; actin-Gal4 > w¹¹¹⁸ for DopEcR) samples as references; all control ΔΔCt values were thus equal to 0. -ΔΔCt to the power of 2 was used for the values in Figure S6. For the copulation assay, RMTL¹³ with the Benjamini and Hochberg method were used to compare the song response behavior between control and experimental groups in each sound condition. For the calcium imaging, Welch’s t-test with the Benjamini and Hochberg method was performed to compare peak calcium responses, increase rates, and decrease rates. For the split-GFP analysis, Welch’s t-test was performed to compare normalized GFP pixel values. stats package (version 4.3.2) and survRM2 package (version 2.2.1) for R (version 4.3.0) was used in these analyses. Cohens’d was calculated using the effsize package (version 0.8.1). Asterisks for statistical significance in figures use the following format: ns, not significant; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Published: July 29, 2025