Neural Network Interactions Modulate CRY-Dependent Photoresponses in Drosophila

Abstract

Light is one of the chief environmental cues that reset circadian clocks. In Drosophila, CRYPTOCHROME (CRY) mediates acute photic resetting of circadian clocks by promoting the degradation of TIMELESS in a cell-autonomous manner. Thus, even circadian oscillators in peripheral organs can independently perceive light in Drosophila However, there is substantial evidence for nonautonomous mechanisms of circadian photoreception in the brain. We have previously shown that the morning (M) and evening (E) oscillators are critical light-sensing neurons that cooperate to shift the phase of circadian behavior in response to light input. We show here that light can efficiently phase delay or phase advance circadian locomotor behavior in male Drosophila even when either the M- or the E-oscillators are ablated, suggesting that behavioral phase shifts and their directionality are largely a consequence of the cell-autonomous nature of CRY-dependent photoreception. Our observation that the phase response curves of brain and peripheral oscillators are remarkably similar further supports this idea. Nevertheless, the neural network modulates circadian photoresponses. We show that the M-oscillator neurotransmitter pigment dispersing factor plays a critical role in the coordination between M- and E-oscillators after light exposure, and we uncover a potential role for a subset of dorsal neurons in the control of phase advances. Thus, neural modulation of autonomous light detection might play an important role in the plasticity of circadian behavior.SIGNIFICANCE STATEMENT Input pathways provide circadian rhythms with the flexibility needed to harmonize their phase with environmental cycles. Light is the chief environmental cue that synchronizes circadian clocks. In Drosophila, the photoreceptor CRYPTOCHROME resets circadian clocks cell-autonomously. However, recent studies indicate that, in the brain, interactions between clock neurons are critical to reset circadian locomotor behavior. We present evidence supporting the idea that the ability of flies to advance or delay their rhythmic behavior in response to light input essentially results from cell-autonomous photoreception. However, because of their networked organization, we find that circadian neurons have to cooperate to reset the phase of circadian behavior in response to photic cues. Our work thus helps to reconcile cell-autonomous and non-cell-autonomous models of circadian entrainment.

Keywords: Drosophila; behavior; circadian; photoreception.

Figure 1.

Light can reset the phase of locomotor activity in the absence of PDF-positive M-oscillators or their neuropeptide PDF. A, Pdf⁰ flies can undergo phase delays as well as phase advances in response to 5 min light pulses (1500 lux). Since rhytmhic Pdf⁰ flies have a short behavior period (Table 1), their PRC could be slightly advanced compared with wild-type flies. We thus first ran a single 5 point time course (left). We then focused on the time points with maximal phase delays (ZT15) and advances [ZT19 and ZT21; right, N (independent experiments) ≥ 4]. Phase changes are plotted on the y-axis; phase delays and advances are represented as negative and positive values, respectively. The x-axis represents the ZT of the light pulse (see Fig. 1-1A). B, Ablation of M-oscillators abrogates the morning anticipation of lights-on and advances the evening peak of activity under a 12 h/12 h LD cycle. Activity is plotted as a function of ZT. Gray bars in the histogram represent activity levels in the night, and the white bars represent activity during the day. For Pdf-GAL4 control flies, the solid arrow shows the morning anticipatory behavior. For Pdf-GAL4/UAS-hid flies, the dashed arrow indicates the lack of morning anticipation, and the gray arrow indicates the advanced evening peak of activity. C, Flies with ablated M-oscillators can respond to brief light pulses. Phase delay in response to a light pulse at ZT15 is shown on the left, and phase advance in response to a light pulse at ZT21 is on the right. Both the phase delay and advance responses in Pdf-GAL4/UAS-hid flies (pink bars) were similar to the Pdf-GAL4 control flies (black bars; Fig. 1-1B). Data are plotted as the mean ± SEM. Bars with the same letters do not significantly differ (p > 0.05 level, Student's t test).

Figure 2.

M-oscillators use PDF to modulate light-mediated phase responses. A, Knocking down jet in the M-oscillators (pink bar) reduces the phase delay and advance responses compared with the Pdf-GAL4 and jet RNAi control lines (black bars). Loss of PDF restores normal phase shifts in flies with jet downregulation in M-oscillators (blue bar). N = 4 for phase responses at ZT15; N = 6 for ZT19 and ZT21. Data represent the mean ± SEM. Different letters above the bars indicate significant differences revealed by one-way ANOVA coupled to post hoc Tukey's test for multiple comparisons (p < 0.05). PD2, Pdf-GAL4 UAS-dcr2 (Fig. 2-1A). B, Loss of PDF receptor restores normal behavioral phase shifts in flies with jet donwregulation in M-oscillators (blue bar), as they were similar to the control pdfr⁻; PD2 flies (black bar). The dotted lines in the blue bars are a reminder of the average phase shift observed for PD2; jetRNAi (see A). N ≥ 3. Data represent the mean ± SEM. The same letters indicate no significant difference. Statistical significance was tested as described above (Fig. 2-1B).

Figure 3.

Light can reset the phase of circadian behavior in the absence of E-oscillators. A, Ablation of E-oscillators (as well as a subset of dorsal neurons) abrogates the evening peak of activity under a 12 h/12 h LD cycle. The first two graphs (left and center) show the LD activity profiles of the control flies. Solid arrows indicate the evening anticipatory behavior. The dashed arrow in the last graph on the right shows the disruption of evening activity peak upon the ablation of E-cells. B, Ablation of E-oscillators (pink bar) had no effect on phase delay or advance responses. Black bars are the UAS and GAL4 control flies (Fig. 3-1). N = 4. Error bars represent SEM. Bars with the same letter do not differ significantly (p > 0.05) by one-way ANOVA and post hoc Tukey's test.

Figure 4.

Phase advance responses are defective when the E-oscillators—but not the dorsal neurons—are electrically silenced. A, Electrical silencing of E-oscillators disrupts the evening peak of activity. The first row shows the locomotor activity under a 12 h/12 h LD cycle. The second row shows locomotor activity rhythms under a 16 h/8 h LD cycle at the permissive temperature of 29°C (inducing KIR expression in experimental flies), and the third row shows the activity under a 16 h/8 h LD cycle at the restrictive temperature of 20°C (no KIR expression). The evening peak of the UAS control and the experimental flies is shifted in the dark phase of the 12 h/12 h LD cycle because of their long circadian period of 26 h (Table 1), but the presence of an evening peak is clearly revealed in these two genotypes when exposed to a 16 h/8 h LD cycle. Note that at 20°C, the evening peak of activity is present in control as well as in experimental flies, but it is absent in experimental flies at 29°C. Solid arrows indicate the normal evening anticipation and dashed arrows represent abrogation of the evening peak in the electrically silenced flies. B, Specific electrical inactivation of the E-oscillators does not block the response to early night light pulses. Black bars, Control flies; pink bars, flies in which the E-oscillators are electrically silenced. The phase shift in response to a light pulse at ZT15 is shown on the left, and that for ZT17 on the right. Note that at ZT15 both the UAS control and the experimental flies have a smaller phase shift than GAL4 control flies due to their long-period phenotype (Table 1). At ZT17, both the GAL4 and UAS controls shift normally despite the difference in period length. The E-oscillator silenced flies also phase shift in response to a ZT17 light pulse, although with a very slightly reduced amplitude, which was statistically significant from control flies as determined by one-way ANOVA followed by Tukey's test (Fig. 4-1A). N = 5. Different letters above the bars represent significant difference between genotypes (p < 0.05). C, Specific electrical inactivation of the E-oscillators compromises the response to late-night light pulses. The phase response to ZT21 light pulses is shown on the left, and ZT23 on the right. Silencing the E-oscillators with the DvPdf-GAL4; Pdf-GAL80 combination (pink bar) disrupts the phase advance response (Fig. 4-1A). Note that the dorsal neurons are not affected by this manipulation. Phase shifts at ZT23 were tested to ensure that the reduction seen in experimental flies at ZT21 was not caused by the long-period phenotype of flies carrying the tub-GAL80ts; UAS-Kir combination (Table 1). ZT23 is not the time point for maximum advances for flies with a normal ~24 h period rhythm; therefore, the GAL4 control shows a reduced response compared with ZT21. However, the phase shifts in electrically silenced E-oscillators and the UAS control flies with a similar endogenous long period were statistically different at both ZT21 and ZT23. N = 5. Statistical analysis was performed as described above. D, Flies carrying tub-GAL80ts; UAS-Kir and DvPdf-GAL4; Pdf-GAL80 show a normal phase advance response at ZT23 at the restrictive temperature of 20°C (Fig. 4-1B). N = 3. Statistical analysis was performed as described above.

Figure 5.

Light can reset the phase of circadian behavior when E-oscillators as well as a subset of dorsal neurons are electrically silenced. A, Inhibition of firing from E-oscillators and a subset of dorsal neurons using the Pdf-GAL80; cry-GAL4(13) combination prevented the evening anticipation peak. The column on the left shows the locomotor activity in the 12 h/12 h LD cycle and the second column on the right shows activity in 16 h/8 h LD cycle. Solid arrows indicate evening anticipation, and dashed arrows indicate loss of the evening peak. B, Silencing the E-oscillators with Pdf-GAL80; cry-GAL4(13) does not reduce the phase delay response to ZT17 light pulses (top) or the phase advance response to ZT21 light pulses (bottom; Fig. 5-1). Black bars, Control flies; pink bar, electrically silenced flies. N = 4. Error bars represent the SEM.

Figure 6.

Photic phase responses in peripheral oscillators resemble those of circadian behavior. A, Whole fly luciferase signal generated by the ptim-TIM-LUC transgene primarily comes from peripheral tissues. Both at ZT18 and ZT19 (time points for the peak TIM-LUC bioluminescence), the majority of the TIM-LUC signal is emitted from the bodies, with only a small contribution from the heads. The difference between the TIM-LUC signal from heads and bodies is statistically significant, as determined by Student's t test at both time points. B, Phase response curve of luciferase rhythms in whole ptim-TIM-LUC flies. Sixteen flies were tested for each time point and in each experiment. Phase shifts in the TIM-LUC bioluminescence rhythms are plotted on the y-axis, and the time at which the light pulse was administered is on the x-axis (N = 3; Fig. 6-1A–D). C, Amplitude of TIM-LUC rhythms. The amplitude of the light-pulsed flies is plotted relative to non-light-pulsed (NLP) flies on the y-axis. x-axis, Different ZTs for light pulses. Note: luciferase rhythm amplitude is significantly reduced when the light pulse is administered at time points closer to the middle of the night (Fig. 6-1D). Different letters indicate significant difference as determined by ANOVA followed by post hoc Tukey's test, p < 0.05.