P1 interneurons promote a persistent internal state that enhances inter-male aggression in Drosophila
- Janelia Research Campus, Howard Hughes Medical Institute, United States
- California Institute of Technology, United States
- Howard Hughes Medical Institute, California Institute of Technology, United States
Figures
Figure 1 with 3 supplements
Thermogenetic activation of P1 neurons promotes inter-male aggression and wing extension.
(A) Overview of thermogenetic activation screen to identify GAL4 lines that promote aggression. (B) Lunge frequency across the 2223 GAL4 lines that were analyzed for increased aggression. The GAL4 …
Figure 1—figure supplement 1
Anatomical expression of parental GAL4s and P1a split-GAL4.
(A, B) Anatomy of R15A01-GAL4 (A), R71G01-GAL4 (B) parental GAL4 lines in the female and male central brain and ventral nerve cord (VNC). White arrowheads denote P1 neuron cell bodies, and yellow …
Figure 1—figure supplement 2
Additional split-GAL4 intersections that target P1 neurons.
(A–C) Anatomical expression of R22G11-GAL4 (A), R15A01-AD;R22G11-DBD spGAL4 (B) and R71G01-AD;R22G11-DBD spGAL4 (C) in the male brain. Image properties are the same as in Figure 1—figure supplement …
Figure 1—figure supplement 3
dsx+/R71G01+ P1 neurons drive both aggression and wing extension.
(A) Positve intersection strategy for targeting P1 neurons using R71G01-LexA and dsxGAL4 to drive a Flp-dependent myc-tagged dTrpA1 (UAS>stop>dTrpA1myc; von Philipsborn et al., 2011; Pan et al., …
Figure 2
TrpA1 activation of FruM+ subset of P1 neurons in parental GAL4 lines promotes aggression but not wing extension.
(A) Intersectional strategy for targeting FruM+ neurons in the parental GAL4 lines. (B, C), Intersectional expression of dTrpRPA1myc in male brains of R15A01-GAL4 ∩ FruFLP (B) and R71G01-GAL4 ∩ FruFL…
Figure 3 with 3 supplements
Optogenetic stimulation of P1 neurons acutely suppresses aggression and promotes wing extension.
(A–B) Frequency titration of 685 nm light on pairs of P1a>CsChrimson::venus (A) or BDPG4U>CsChrimson::venus (B) males. Blocks of 30 s photostimulation (PS, gray bars) with increasing stimulation …
Figure 3—figure supplement 1
Functional imaging of P1 neuronal responses to optogenetic activation.
(A, B) Experimental setup of calcium imaging experiments. See Materials and methods for details. (A) Calcium responses of P1 neurons to Chrimson activation were recorded with a two-photon laser …
Figure 3—figure supplement 2
Photoactivation of P1a>CsChrimson flies.
(A) Comparison of P1a>CsChrimson lunges (blue bars) and wing extensions (red bars) per min at different PS frequencies with pre-stimulation baseline levels. Data are replotted from experiments shown …
Figure 3—figure supplement 3
Effect of silencing TKFruM neurons on behavioral phenotype of P1 neuron optogenetic activation.
(A) Plots in these panels are from the P1/TKFruM epistasis experiment shown in Figure 3H. Raster plots show lunging and wing extension for each genotype tested; genotypes and number of pairs tested …
Figure 4 with 1 supplement
Transient activation of P1 neurons induces persistent aggression.
(A) Activation of P1a spGAL4 with UAS-ReaChR. Pairs of males were photostimulated for 1 min with 530 nm light (30 Hz, 5 ms pulse-width, 0.25 mW/mm2). Plot properties are the same as in Figure 3A. (B,…
Figure 4—figure supplement 1
Wing extension is not necessary for persistent post-PS aggression.
(A) Raster plots relate to photostimulation experiment of P1a>ReaChR pairs with 627 nm light at 50 Hz that is shown in Figure 4D−F. Since there was a small but significant increase in wing extension …
Figure 5 with 1 supplement
P1 activation induces a persistent state that is fly intrinsic.
(A, B) Assay and experimental design used to separate optogenetic activation from social interaction. Males were placed on opposite sides of a removable opaque barrier (left; experimental design on r…
Figure 5—figure supplement 1
Persistent aggression after optogenetic activation of P1 neurons.
(A) Raster plot of lunging and wing extension by P1a>ReaChR male towards a WT male ('mixed genotype pair'). Both males were group-housed to reduce baseline aggression. Flies were photostimulated …
Figure 6
Models for how P1 neurons may regulate courtship and aggression.
(A, B) Alternative models for how P1 neurons influence aggression ('ag') vs. courtship ('ct')-promoting circuitry. (A) 'Direct' model. P1 neurons exert parallel and independent influences on …
Videos
Video 1
dTrpA1 activation of R15A01-GAL4.
Pairs of R15A01>dTrpA1 males at 29°C exhibit interspersed bouts of aggression (lunging) and courtship (unilateral wing extension).
Video 2
P1a>dTrpA1 behavioral phenotype.
Pairs of P1a>dTrpA1 males at 29°C exhibit interspersed bouts of aggression and wing extension. Note that wing extensions occurring before lunging as shown in this clip is not the invariant sequence …
Video 3
Activation of P1 neurons induces aggression between wingless males.
Pair of P1a>dTrpA1 males at 29°C with wings removed still show aggression.
Video 4
P1a>CsChrimson phenotype in pairs of males during strong photostimulation.
Pairs of P1a>CsChrimson males at the onset of 50 Hz PS (see Figure 3) show a rapid suppression of aggression and locomotion and an increase in wing extensions. At the offset of the PS, males switch …
Video 5
P1a>ReaChR males interacting following barrier removal 10 min after photostimulation.
Males were stimulated as described in Figure 5A.
Additional files
-
Supplementary file 1
Genetic intersections used to target P1 neurons.
- https://doi.org/10.7554/eLife.11346.022
-
Supplementary file 2
Full genotypes of flies in each experiment.
- https://doi.org/10.7554/eLife.11346.023
-
Supplementary file 3
Training datasets and performance of JAABA classifiers.
(A) The videos and genotypes of flies used to train the lunging and wing extension classifiers. The total number of bouts of each behavior, along with the total number of frames scored as positive or negative for the behavior are shown. (B) Performance of the classifiers against an independent set of 'ground-truth' videos. The framewise error rates and performance of the classifiers was calculated from a set of between 4,000 and 32,000 frames of manually scored ground-truth labels (total frames and total bouts). Frames were labeled as 'Important' if there was no uncertainty about the behavioral label to distinguish them from frames where the human scorer was unsure of the label. For each classifier the error rates, precision and recall for ‘important’ vs. all frames is shown. See Materials and methods for detailed description of ground-truth validation.
- https://doi.org/10.7554/eLife.11346.024
