A Circuit Node that Integrates Convergent Input from Neuromodulatory and Social Behavior-Promoting Neurons to Control Aggression in Drosophila

Abstract

Diffuse neuromodulatory systems such as norepinephrine (NE) control brain-wide states such as arousal, but whether they control complex social behaviors more specifically is not clear. Octopamine (OA), the insect homolog of NE, is known to promote both arousal and aggression. We have performed a systematic, unbiased screen to identify OA receptor-expressing neurons (OARNs) that control aggression in Drosophila. Our results uncover a tiny population of male-specific aSP2 neurons that mediate a specific influence of OA on aggression, independent of any effect on arousal. Unexpectedly, these neurons receive convergent input from OA neurons and P1 neurons, a population of FruM⁺ neurons that promotes male courtship behavior. Behavioral epistasis experiments suggest that aSP2 neurons may constitute an integration node at which OAergic neuromodulation can bias the output of P1 neurons to favor aggression over inter-male courtship. These results have potential implications for thinking about the role of related neuromodulatory systems in mammals.

Figure 1. R47A04 neurons are responsible for male-male aggressive behavior

(A) Overview of strategy for screen to identify OARNs involved in aggression. (B and C) The experimental setup (B) and the behavioral arena (C) used in this behavioral screen (Hoyer et al., 2008). (D) Strategy for conditional silencing or activation using tub-GAL80^ts. (E, F) Number of lunges in pairs of single- (E) or group-housed (F) male flies from GAL4 lines identified in the screen, during silencing using Kir2.1 (E) or activation using NaChBac (F). (G, H) Confocal images illustrating fly brains immunostained for mCD8::GFP expression (green) and the neuropil marker nc82 (magenta) in R47A04>mCD8::GFP (G) or R48B04>mCD8::GFP (H). Scale bars in panel G and H are 50 μm. For E and F, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed. P-values were adjusted for multiple comparisons using the Bonferroni correction. Here and throughout, *: P<0.05, **: P<0.01, ***: P<0.001 and ****: P<0.0001. Full genotypes for this and all subsequent figures are listed in Table S1.

Figure 2. Oamb acts in R47A04 neurons to control aggression

(A–C) RNAi-mediated Oamb knock-down in R47A04 neurons during male-male social behaviors. Frequency of lunges (A) or unilateral wing extensions (UWEs) (B) and the SBPI (Social Behavior Proportion Index, see the STAR Methods, C) in R47A04>UAS-Oamb RNAi flies. For C, error bars denote ± S.D. (D) Frequency of lunges in Oamb mutant (Oamb²⁸⁶/Oamb²⁸⁶) male flies rescued by R47A04-GAL4 driving UAS-Oamb-K3. (E) Frequency of lunges after overexpression of Oamb cDNAs encoding either of two alternative splice variants, with or without 10 mM OA feeding. (F) Time-course of OA feeding for behavioral experiments. Note that the time-line is not to scale. For A–E, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed.

Figure 3. fru-Expressing aSP2 neurons in R47A04-GAL4 control aggression

(A_1–4) Confocal images of the SMP region in male brains expressing R47A04-LexA>nls::GFP (green) and fru^GAL4>nls::tdTomato. (A₁) 3D render view, and (A_2–4) partial Z-stack images of the boxed region in A₁. (A₅) Schematic of aSP2 neurons in R47A04-GAL4. (B_1,2) Confocal images of R47A04>mCD8::GFP in the SMP region of male (B₁) and female (B₂) brain. (C–E) Confocal images of the SMP region in flies expressing R47A04-LexA>mCD8::GFP (green) and indicated GAL4 lines driving mCD8::RFP (magenta). (F) Confocal images of R47A04-GAL4>mCD8::GFP (green) male brain immunostained with anti-GABA antibody (magenta). (G_1,2) Confocal images of male brain containing split-GAL4 intersection between R47A04-AD and fru^P1.LexA>LexAop-DBD (R47A04 ∩ Fru) driving UAS-Kir2.1^eGFP; G₂: magnified image of the boxed region in G₁. (H) Raster plots illustrating bouts of lunges (blue) and UWEs (red) in control (no NaChBac) or aSP2-split GAL4>NaChBac fly pairs. (I–L) Average lunging rate (I), fighting frequency (see text) (J), unilateral wing extension rate (K) and the SBPI (L) of flies of the indicated genotypes; “fru^P1.LexA” indicates LexA expressed from the fru P1 promoter (Pan et al., 2011). (M) Raster plots illustrating bouts of lunges (blue) and UWEs (red) in control or aSP2-split GAL4-UAS>Kir2.1 fly pairs. (N–Q) Parameters as in (I–L) for flies of the indicated genotypes, except that in (O) the 95^th percentile is for single-housed control flies. Scale bars in panel A₁, C₁, and C₂ are 20 μm, A_2–4 are 10 μm, B₁ and B₂ are 50 μm, D is 25 μm. For I–L and N–Q, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed. For L and Q, error bars denote ± S.D.

Figure 4. R47A04^aSP2 neurons respond to OA

(A, B) Experimental setup for calcium imaging of R47A04^aSP2 neurons in brain explants using 2PM. Brains were perfused with saline or 500 μM OA. (C) Representative images of GCaMP responses during pre-stimulation (Pre (t=−2 min)) and post-stimulation (Post (t=10 min)) in R47A04^aSP2 neurons. (D) Responses (%ΔF/F) of R47A04^aSP2 neurons to 500μM OA perfusion in brain explants from R47A04-GAL4>UAS-GCaMP6m flies, with or without 500μM mianserin (red line) or UAS-Oamb RNAi (green line). (E) Fold changes in integrated %ΔF/F (∫ΔF/Fdt) during indicated time periods. (F) Confocal image of Tdc2-GAL4>mCD8::GFP (green) and R47A04-LexA>myr::tdTomato in the lateral protocerebral complex. Magenta arrow: cell bodies of R47A04^aSP2 neurons, white arrow: OA-ASM, white double arrows: OA-AL and white arrowhead: OA-VL. (G₁,₂) Native GRASP signals (G₁: green and G₂: GRASP only) with Tdc2-GAL4>CD4::spGFP1-10 and R47A04-LexA>CD4::spGFP11 and the fibers of R47A04-LexA neurons labeled with LexAop2-myr::tdTomato (G₁, magenta). R47A04^aSP2 fibers are delineated by dashed line in G₂. (H) Schematic of aSP2 neurons with putative input site from Tdc2⁺ neurons. Scale bars in panel C are 10 μm, and F, G₁ and G₂ are 20 μm. For E, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed.

Figure 5. aSP2 neurons are physiological and functional targets of P1 neurons

(A) Confocal image illustrating close proximity between aSP2 (green) and P1 neurons (magenta) in the lateral protocerebral complex. (B) GRASP signals between P1 and aSP2 neurons. Arrows: arch region. Arrowheads: ring region (see Fig. 3A₅). (C) Schematic of aSP2 neurons with putative input sites from Tdc2⁺ and P1 neurons. Regions labeled with syt::GFP (green) and DenMark (magenta) are depicted. See also Fig. S4B, C. (D) Representative images during pre-stimulation (t=0s) and P1>ReaChR stimulation (t=40s) of GCaMP6s responses (ΔF/F) in aSP2 neurons.. Circles: cell bodies of aSP2 neurons. (E) Responses (%ΔF/F) of aSP2 neurons to P1 stimulation. No ReaChR: control flies lacking UAS-ReaChR. Thick solid lines represent average trace and thin grey lines individual responses. (F) Fold change in integrated ΔF/F (∫ΔF/Fdt) during indicated 10 sec time bins. (G–N) Epistatic relationship between P1 and aSP2 neurons. (G) Experimental design. (H, I) Raster plots of control (H, P1>CsChrimson + BDPG4U>Kir2.1) and experimental (I, P1>CsChrimson + R47A04>Kir2.1) flies. Blocks of frequency and intensity-titrated 30 s photostimulation trials (grey bars, 655 nm) with 2.5 min inter-trial intervals were delivered as indicated. C, continuous stimulation. Blue arrows indicate Post-Stimulation Intervals (PSIs) showing statistically significant differences between experimental and control flies. (J–L), Frequency of lunges (J) and UWEs (K), and SBPI (L) during PSIs showing significant differences. For L, error bars denote ± S.D. (M, N) Frequency of lunges (M) and UWEs (N) for all PSIs. Data points indicate mean ± S.E.M. Asterisks indicate statistically significant pairwise comparisons, bracket indicates significant difference between curves (*: P<0.05, ***: P<0.001). Scale bars in panel A and B are 20 μm, D are 10 μm. For F and J–L, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed. For M and N, two-way ANOVA and post hoc Mann-Whitney U tests were performed.

Figure 6. Convergent activation of R47A04^aSP2 neurons by P1 activation and OA

(A, B) OAergic modulation of the calcium response of aSP2 neurons to P1 optogenetic activation. All flies contained R71G01-GAL4>ReaChR and R47A04-LexA>GCaMP6s unless otherwise indicated. (A) Time courses of average (thick line) and individual (thin lines) responses (%ΔF/F). OA or mianserin feeding (500 μM each) are indicated. The data for No ReaChR and ReaChR are reprinted from Fig. 5 using the same y-axis scale, to facilitate visual comparisons. Statistical analyses (B and Fig. 5F) were performed using all pooled data and P-values were adjusted for multiple comparisons with all conditions using the Bonferroni correction. (B) Fold change in integrated ΔF/F (∫ΔF/Fdt) during light stimulation period. (C–G) Experiment to test epistatic interactions between P1 stimulation and endogenous OA signaling with or without aSP2 neuronal activation. (C–E) (Left) experimental design, drug treatment and genotypes for each condition. (Right) raster plots showing bouts of lunges (blue) and UWEs (red) evoked by P1 activation without (C) or with (D, E) 1 mM mianserin feeding, together with (E) or without (C, D) NaChBac-mediated aSP2 activation.. Blocks of frequency (1 Hz, 5 Hz, 20 Hz and Continuous (C)) and intensity (5.0 mW/cm², 9.5 mW/cm²)-titrated 30 s photostimulation (grey bars, 655 nm) with 2.5 min inter-trial intervals were delivered as indicated above raster plots. (F, G) Frequency of lunges (F) and UWEs (G). Data points indicate mean ± S.E.M. For bar graphs and SBPI scores see Figure S6A–C. For F–G, two-way ANOVA and post hoc Mann-Whitney U tests were performed.

Figure 7. OAergic modulation enhances P1-stimulated aggression in an aSP2-dependent manner

(A–F) Optogenetic activation of P1 neurons was performed without (A) or with (B, C) 5 mM OA feeding, with (C) or without (A, B) silencing of R47A04 neurons. (A–C) (Left), experimental design, drug treatments and genotypes. (Right), raster plots showing bouts of lunges (blue) and UWEs (red) in fly pairs subjected to P1 activation alone (A), P1 activation + 5mM OA feeding (B) or P1 activation +5 mM OA feeding + aSP2 silencing (C). Blocks of frequency (1 Hz, 5 Hz) and intensity (5.0 mW/cm², 9.5 mW/cm²) titrated 30 s photostimulation trials (grey bars, 655 nm) with 2.5 min inter-trial intervals were delivered as indicated (right). Aggression in P1-stimulated flies (A) was low in this experiment, possibly due to leakage expression of BDPG4U>Kir2.1 in the control genetic background. (D–F) Frequency of lunges (D), UWEs (E) and SBPI (F). For clarity, only PSIs showing statistically significant differences are illustrated; data for all PSI intervals are shown in Figure S7A–C. For F, error bars denote ± S.D. (G) Summary of experiments showing the interaction between OA neurons, P1 neurons and aSP2 neurons. “√” indicates experiments performed; “N/A,” Not Applicable; “ND,” Not Done. “§” indicates that experiments were part of the design illustrated in G3 (Figs. 6 and 7). Arrows in illustrations are not meant to imply monosynaptic connectivity. It is uncertain whether the behavioral, physiological and anatomical interactions between P1 and aSP2 cells documented here (G2, G3) are all mediated by the same subset of P1 neurons. For D–F, Kruskal-Wallis one-way ANOVA and post hoc Mann-Whitney U tests were performed.