Manipulation of an innate escape response in Drosophila: photoexcitation of acj6 neurons induces the escape response

Abstract

Background: The genetic analysis of behavior in Drosophila melanogaster has linked genes controlling neuronal connectivity and physiology to specific neuronal circuits underlying a variety of innate behaviors. We investigated the circuitry underlying the adult startle response, using photoexcitation of neurons that produce the abnormal chemosensory jump 6 (acj6) transcription factor. This transcription factor has previously been shown to play a role in neuronal pathfinding and neurotransmitter modality, but the role of acj6 neurons in the adult startle response was largely unknown.

Principal findings: We show that the activity of these neurons is necessary for a wild-type startle response and that excitation is sufficient to generate a synthetic escape response. Further, we show that this synthetic response is still sensitive to the dose of acj6 suggesting that that acj6 mutation alters neuronal activity as well as connectivity and neurotransmitter production.

Results/significance: These results extend the understanding of the role of acj6 and of the adult startle response in general. They also demonstrate the usefulness of activity-dependent characterization of neuronal circuits underlying innate behaviors in Drosophila, and the utility of integrating genetic analysis into modern circuit analysis techniques.

Figure 1. Optical control of muscle depolarization and contraction in Drosophila larvae.

(A) CO₂-anesthetized L3 larvae producing ChR2 in muscles (twi-GAL4; 24B-GAL4) display a marked head retraction in response to light stimulation. The larva is immobile during dark periods (0–30 sec, 60–90 sec) but undergoes a rapid contraction in response to brief periods (2–3 second-long light periods at the indicated times) of illumination with blue light (0.2 mW/mm²). Moving larvae of the same genotype show complete motion arrest in response to illumination (see Movie S1). Animals were raised in vials supplemented with all-trans retinal. (B–D) One-electrode recordings from larval muscle 6 (for a map of muscle fibers see [31]). (B) Inward current in voltage-clamped muscle fibers, evoked by pairs of 0.5 second light pulses. Several such pairs are overlaid, with a delay varying from 1 to 10 seconds; the first pulse marked by a large arrowhead, and the second pulse from each pair is marked by a small arrowhead. (C) Enlarged view of inward current generated by a single 0.5 sec light pulse. (D) Voltage traces showing membrane depolarization in a current-clamped muscle evoked by 15 ms light pulses delivered at 10 Hz.

Figure 2. Precise temporal control of motor neuron activity in Drosophila larvae.

Animals contain one copy of the D42-GAL4 motor neuron driver and three copies of UAS-chr2::yfp. (A) Brain stem of L3 larva producing ChR2-YFP in the cell bodies of motor neurons. (B) A freely moving larva shows rapid body contraction in response to illumination (upper panel prior to illumination, lower panel during illumination) with blue light (0.2 mW/mm²). Also see Movie S2. (C–G) One-electrode recordings from muscle 5 of a dissected third instar larva. Short pulses (as short as 2 ms) of blue light (5 mW/mm²) generate excitatory junction potentials (EJP's). Precise temporal control of motor neuron activity is demonstrated by varying the frequency of light pulses (lines above traces represent individual light pulses). Trains of EJP's in response to 10 ms light pulses delivered at (C) 5 Hz, (D) 10 Hz and (E) 20 Hz are shown. (F) The responses of the neurons are highly reproducible. This panel shows an overlay of current traces generated by 7 sweeps of 20 light pulses delivered at 5 Hz. (G) Addition of glutamate receptor blockers (CNQX and D-APS) to the preparation blocks light-evoked EJPs.

Figure 3. Optical control of behavior in adult flies.

(A) CO₂-anesthetized adult flies that express ChR2 in muscles (MHC-gs-GAL4) respond to sustained illumination by contracting abdominal muscles and extending their abdomen. Illumination periods are indicated (‘Light’) and arrows indicate equivalent positions for reference. Also see Movie S3. Adult flies were fed 5 mM ATR and 200 µM RU486 for 3 days prior to assaying. (B) CO₂-anesthetized adult females expressing ChR2 in motor neurons were assayed for light-induced muscle contractions in leg and abdominal muscles. Animals with one copy of D42-GAL4 and 3 copies of UAS-chr2::yfp display a reproducible light response when they are fed all-trans-retinal (ATR). Little or no response to light was observed with flies that were not fed ATR (0 mM ATR) or that contained only UAS-chr2::yfp (not shown). Three groups of 20 flies were assayed. (C) Production of ChR2 in cholinergic neurons produces a light-evoked seizure response in adult flies that are fed ATR (also see Movie S4). Few or no light-evoked responses were observed for animals not raised on ATR-supplemented food. Assays were done in triplicate (n = 20). The seizure response was not observed in animals that contained only one or two copies of UAS-chr2::yfp (data not shown). All behaviors were scored blindly with respect to ATR supplementation. Mean±s.d shown for all panels.

Figure 4. Genetic and physiological disruption of the escape response.

(A) Flies were tested for a jump response within 3 s of bumping their vial on the table; the average number of jumps over three trials is shown. Hemizygous mutant males (acj6-GAL4/Y) show an attenuated startle response. Data shown as mean±s.e. (B) Silencing acj6-GAL4 expressing neurons with the temperature-sensitive dynamin allele shibire ^TS in acj6 neurons recapitulates the acj6 mutant phenotype, but expressing ChR2 with or without ATR generates no deficit. (C) acj6-GAL4 expression as revealed by UAS-mcd8::GFP in the anterior and posterior brain, as well as the VNC. Arrowheads mark notable cell clusters. For all figures, the double-asterix shows significance at p<0.001 vs. the relevant control.

Figure 5. Induction of a synthetic escape response.

(A) ChR2-mediated excitation of acj6-GAL4 expressing neurons in female flies; number of jumps pooled over three 15 s illumination periods. Photoexcitation of acj6 neurons is sufficient to induce a startle response. Pooled controls are all flies lacking ATR or one transgene. Data shown as mean±s.d. (B) ChR2-mediated excitation of acj6 neurons in mutant flies, as well as photoexcitation restricted to non-cholinergic acj6 neurons by the cha-GAL80 transgene. The light-induced jump response is attenuated in females held over deficiency, but not in hemizygous male flies. Data shown as mean±s.d (C) acj6-GAL4 expression in non-cholinergic neurons, with expression limited by the cha3.1-GAL80 transgene. White arrowheads mark cell clusters retained in cha-GAL80 flies, as compared to Fig. 4C, and black arrowheads mark those largely absent; see text for details. Asterix denotes p<0.05 vs. the relevant control, double-asterix denotes p<0.001.