Exquisite light sensitivity of Drosophila melanogaster cryptochrome

Abstract

Drosophila melanogaster shows exquisite light sensitivity for modulation of circadian functions in vivo, yet the activities of the Drosophila circadian photopigment cryptochrome (CRY) have only been observed at high light levels. We studied intensity/duration parameters for light pulse induced circadian phase shifts under dim light conditions in vivo. Flies show far greater light sensitivity than previously appreciated, and show a surprising sensitivity increase with pulse duration, implying a process of photic integration active up to at least 6 hours. The CRY target timeless (TIM) shows dim light dependent degradation in circadian pacemaker neurons that parallels phase shift amplitude, indicating that integration occurs at this step, with the strongest effect in a single identified pacemaker neuron. Our findings indicate that CRY compensates for limited light sensitivity in vivo by photon integration over extraordinarily long times, and point to select circadian pacemaker neurons as having important roles.

Figure 1. A double-plotted median actogram illustrating the basic paradigm used in this manuscript (A), and a Phase Response Curve (B).

A) A median actogram showing the results of a light pulse during late subjective night on circadian phase in constant darkness (DD). Flies were entrained to a 12 hr;12 hr light dark schedule with green light for 6 days, then given a 6 hr blue light pulse of intensity 33 nw/cm² at ZT18-24. The resulting phase advance (double arrow line) is shown by extrapolating from the computer called RMS line (blue) marking the activity off times (red dots) in DD. Phase shift data in this manuscript is computed from individual flies, relative to the phase shift of a control set of flies not receiving a light pulse. Data in this figure is double plotted for ease of viewing, but annotations are only shown for one of the two occurrences. B) A phase response curve, showing the averaged phase shifts resulting from a 1 hr light pulse at the indicated circadian times. Light pulses were of white light of intensity ∼150 µw/cm². Averaged data reprinted from .

Figure 2. Phase shift magnitude as a function of intensity and duration of blue light pulses.

LD-entrained flies were given light pulses of either: A) 10 min, or B) 120 min, at different intensities at times centered at ZT20. n = 10 per condition. For both duration light pulses, there are graded phase shift responses as a function of increasing light intensity. 10 min: 7,000 vs 50 nw/cm², P = 0.02; 120 min: 50 vs 4 nw/cm²: P = 0.04; 4 vs 0.3 nw/cm²: P = 0.03 (ANOVA). Asterisks indicate significantly different values from the no pulse control. 10 min: 600 nw/cm² vs no pulse: P = 0.02; 7,000 nw/cm² vs no pulse: P = 0.007. 120 min: 0.3, 4 nw/cm² vs no pulse: not significant; 50 nw/cm² vs no pulse: P = 0.0.006. Error bars = SEM. The dashed line indicates two light pulses with equal numbers of photons: not significantly different.

Figure 3. Phase shift magnitude as a function of blue light pulses with equal numbers of photons, reciprocally varying time and intensity.

A) w¹¹¹⁸; B) norpA^P41; C) cry⁰² LD-entrained flies were given light pulses of indicated duration centered at ZT20. Light intensities were controlled such that all flies received the same number of photons. Statistics, ANOVA, with post-hoc correction for multiple comparisons; n = 16–20 for each light pulse. A) All phase shifts were significant relative to the no pulse control (alpha <0.0125). The line through the light pulse points is a linear regression through the light pulse values (R² = 0.14), showing a significant positive slope of 0.55±0.13 (P = 4×10⁻⁵), using an X scale of log₁₀ light pulse duration. B) The 1, 10 and 100 min light pulses are all highly significantly different relative to the no pulse control or to the 0.1 min light pulse (P<10⁻⁸). C) Only the 100 min point approaches significance relative to the no pulse control (P = 0.018, alpha = 0.0125). Asterisks indicate significantly different values from the no pulse controls. Error bars = SEM.

Figure 4. Large amplitude phase advance responses to 6 hr ZT21-centered light pulses depend primarily on CRY.

A) w¹¹¹⁸ n = 15–18 per condition. B) norpA^P41; n = 6–8 per condition. C) cry⁰² n = 7–12 per condition. Asterisks: significant phase advances relative to the no pulse control (ANOVA, significance set at alpha = 0.017). Error bars = SEM.

Figure 5. TIM and PDF immunostaining in various LNv neurons.

Flies were subject to light pulses at initiating at ZT20 and assayed at ZT22. Blue light pulse conditions: A) No light pulse control. B) 7000 nw/cm², 10 min. C) 600 nw/cm² 10 min. D) 50 nw/cm² 120 min. Four of the five s-LNv neurons are labeled with PDF as well as TIM. The 5^th-sLNv is identified by TIM immunoreactivity and by position and morphology.

Figure 6. Quantitation of TIM levels in the specific LN subsets as shown in Figure 5.

A) s-LN_v. B) 5^th s-LN_v (PDF negative). C) l-LN_v. D) LN_d. Intensity/durations were designed based on the data in Figure 2. For each neuronal subset, we show the no light pulse control (0), 600 nw/cm² 10 min, representing a roughly half-maximal intensity-duration; 50 nw/cm² 120 min, an equal-photon exposure to the former, and 7,000 nw/cm², 10 min, yielding a stronger phase shift. All neuronal classes show highly significant TIM degradation in response to light pulses relative to the no pulse control (P<10⁻⁶). The PDF negative 5^th s-LN_v was visible in 10 of 11 LNv clusters in the no pulse control, and the absolute intensity of TIM IR in this neuron was indistinguishable from the s-LN_v or any of the other neuronal clusters: 209±21 (SEM, arbitrary fluorescence units), vs 192±19 for the other s-LN_v. Thus, the near disappearance of TIM immunoreactivity in the 5^th s-LN_v in response to the 50 nw/cm² 120 min light pulse cannot be explained by differential detection sensitivity. This neuron showed TIM immunoreactivity in 6 of 11 LN_v clusters following the 600 nw/cm² 10 min light pulse, but only 1 of 11 following the 50 nw/cm² 120 min light pulse. Statistics were performed using two way ANOVA on square root transformed data to better approximate normality (Stat Plus, Analystsoft). Two way ANOVA values for the interaction of specific light pulse x cell types are shown in Table S1. The six LNd neurons are heterogeneous, with only three showing detectable CRY immunoreactivity , but our analyses do not detect a subset showing differential TIM degradation.