Functions of Four Light-Sensing Proteins from Water Animals Tested in Fruit Fly Body Clock Cells

To generate pUAST-attB-cry1a, cry3, and cry4, the zebrafish cry genes were subcloned from pGAD-cry1a, pGEM-cry3, and pBS-cry4 into pUAST-attB (Brand and Perrimon, 1993). The full coding sequence of the SpuCry gene (S. purpuratus genome 3.1: SPU_000282, WHL22.613873; S. purpuratus genome 5.0: LOC581225) was amplified using the Forw-EcoRI-CCGGAATTCATGCCTGGCGGTGCCT and Rev-XhoI-TCCGCTCGAGATTAAGAAAAAGGAACAAAC primers and a full-length cDNA clone derived from S. purpuratus total RNA at the stage of 33 hours after fertilization (early gastrula). A purified fragment was cloned into the pGemT vector (Promega) according to manufacturer’s instructions. Recombinant clones were sequenced using T7 and SP6 primers to confirm the correct fragment had been cloned. To generate pUAST-attB-SpuCry, SpuCry was subcloned from pGemT-SpuCry into pUAST-attB. All constructs were verified by sequencing before injection into fly embryos.

Flies were raised in 12 h:12 h light-dark (LD) cycles on a standard Drosophila medium (0.7% agar, 1.0% soya flour, 8.0% polenta/maize, 1.8% yeast, 8.0% malt extract, 4.0% molasses, 0.8% propionic acid, 2.3% nipagen) at 25 °C and 40%-60% humidity. Pdf-gal4 (Renn et al., 1999), Clk856-gal4 (Gummadova et al., 2009), and tim-gal4:27 (Kaneko and Hall, 2000) were crossed into a homozygous mutant cry^b background (Stanewsky et al., 1998) using appropriate balancer chromosomes and dominant markers. UAS-cry24.5 and UAS-per:16 lines have been described (Blanchardon et al., 2001; Emery et al., 1998) and are located on chromosomes 2 and 3, respectively. pUAST-attB vectors containing zebrafish cry1a, cry3, cry4, and SpuCry were transformed into y¹v¹nos-Φ31, attP40/attP40 flies using standard procedures. Transformants of each cry gene were then crossed into a homozygous cry^b mutant background. BG-luc60 and plo3b-1 transgenics are located on chromosome 1 and 3, respectively, and have previously been described (Stanewsky et al., 1998, 2002).

Analysis of locomotor activity of 4- to 5-day-old male flies was performed using the Drosophila Activity Monitor System (DAM; Trikinetics). Individual flies were placed into glass tubes filled with 2% agar and 4% sucrose and loaded into the DAM system. The monitors were located inside a light- and temperature-controlled incubator (Percival) where the fly’s activity was monitored for 1-2 weeks depending on different experimental conditions. Plotting of behavioral activity, rhythmicity, and period calculations was performed using a signal-processing tool-box (Levine et al., 2002) implemented in Matlab (MathWorks). For phase determination, activity data were transferred to an Excel macro (Microsoft), and the position (phase) of the evening activity peak for each individual fly was determined for every day of the experiment (phase plots in Figure 2A and 2B) as described (Sehadova et al., 2009). To calculate how long a certain genotype requires for re-synchronizing to the shifted LD cycle, daily activity profiles of individual flies were plotted, and the number of days where the evening peak showed transient delays before reaching a stable phase was determined manually for each fly (Figure 2C and 2D).

Flies were entrained in 12 h:12 h LD at 25 °C for 3-4 days before fixation. Ten flies of each genotype were fixed at ZT21 or 2 h after a light pulse (LP) given at ZT19; therefore, all the flies were collected at the same time. After the 2.5-h fixation in 4% paraformaldehyde in phosphate buffered saline (PBS) + 0.1% Triton-X100, fly brains were dissected and washed in PBS + 0.1% Triton-X100, followed by incubation with primary antibodies as described (Chen et al., 2015). Rat anti-TIM (1:1000) (Rush et al., 2006), mouse anti-Pigment Disp ersing Factor (PDF) (1:1000, Developmental Studies Hybridoma Bank, DSHB), rat anti-HA (1:1000, Roche), and secondary rat AlexaFluor-594 and mouse AlexaFluor-647 antibodies (1:400, Invitrogen) were applied. Mounted brains were scanned using a Leica TCS SP5 confocal microscope. Quantification of TIM signals was performed (Gentile et al., 2013) with minor modifications: Pixel intensity of stained neurons and background staining in each neuronal group was measured using Image J. Background signal was determined by taking the average signal of two surrounding fields of each neuronal group and was subtracted from the neuronal signal. For each group of clock neurons, at least 6 hemispheres from each genotype were measured. Data were normalized by setting the peak value to 1, and the value from each time point was then divided by the peak value.

Luciferase expression of individual flies was measured as described (Stanewsky et al., 2002). Briefly, 2- to 3-day-old males were ether-anesthetized and loaded in a 96-well microtiter plate in which every other well contained 100 µl of 5% sucrose, 1% agar, and 15mM luciferin. Flies were measured in a Packard Topcount Multiplate Scintillation Counter for 6-7 days during 12 h:12 h LD and DD at 25 °C as indicated in the figures. Data were plotted using BRASS software (Version 2.1.3) (Locke et al., 2005) and analyzed using Chronostar software (Klemz et al., 2017). In particular, data were first de-trended using a running average with a 24-h window. After trendline subtraction, data were subjected to a sinus fit operation, and the resulting curves were plotted in Figure 4B (see Klemz et al., 2017 for details).

The circadian clock of Drosophila fails to operate in constant light (LL), presumably because of constitutive light-dependent degradation of one of its key components, the clock protein Timeless (Tim) (Price et al., 1995; Zeng et al., 1996). As a consequence, while Drosophila locomotor activity rhythms are sustained in conditions of constant darkness (DD), wild-type flies become arrhythmic in LL and constant temperature (Konopka et al., 1989; Konopka and Benzer, 1971). Interfering with light-input pathways to the clock can restore clock function in LL, leading to molecular and behavioral rhythmicity (Chen et al., 2011; Emery et al., 2000a), while LL rhythmicity induced by the cry^b mutation can be reversed to wild-type LL arrhythmicity by driving UAS-cry expression in all clock cells (Emery et al., 2000b). As a quick and straightforward assay to test if the various cry genes can replace light-dependent Cry functions in flies, we exposed cry^b mutant flies heterologously expressing one of the different cry genes in all clock neurons (Clk856-gal4/UAS-cry; cry^b/ cry^b) to bright LL (~1500 lux). As expected, wild-type flies were arrhythmic in LL, while homozygous cry^b/ cry^b flies displayed robust rhythmicity (Figure 1, Supplementary Table S1↗). The LL-rhythmicity of cry^b mutants could be fully rescued by driving Drosophila UAS-cry expression with Clk856-gal4. Based on sequence homology to Drosophila Cry, we predicted that SpuCry and zebrafish Cry4 could at least partially replace its function, but cry^b flies expressing these cry genes in all clock neurons remained thoroughly rhythmic in LL. As expected, none of the more distantly Drosophila-related cry genes (zebrafish cry1a and cry3) restored LL arrhythmicity (Figure 1). Similar results were obtained in dim LL (~100 lux) and with the Pdf-gal4 driver, where expression of Drosophila cry resulted in 50% of LL arrhythmic flies as previously reported (Emery et al., 2000b) and none of the other cry genes had any effect (Supplementary Table S1↗). To rule out whether the LL-assay may not be suitable to detect potential partial photoreceptive functions of the heterologously expressed cry genes, we next turned to a more sensitive assay.

Resynchronization to altered LD cycles (i.e. a jetlag assay) is a very sensitive behavioral assay to determine functionality of the different light-input pathways to the circadian clock. For example, compared to wild-type flies, which require only 1-2 days to resynchronize their behavioral activity pattern to an 8-h-delayed LD cycle, cry^b mutants need 4-5 days, while flies with an additionally impaired visual system (norpA^P41cry^b) require >7 days to achieve this task (Emery et al., 2000b). Because it allows for detection of partially functional light input to the circadian clock, we exposed cry^b mutant flies expressing one of the different cry genes in all clock cells to such a jetlag assay. In particular, Clk856-gal4/UAS-cry; cry^b/cry^b flies were first kept in a 12 h:12 h LD cycle for 5 days, after which the LD cycle was delayed by 6 h. After exposure to this delayed LD cycle for 7 days, flies were released into DD for an additional 3-5 days. During the first 5 days, flies from all genotypes synchronized their activity pattern to the LD cycle, with activity peaks in the morning and evening. In addition, all flies “anticipated” the environmental light transitions in the morning and evening by increasing their locomotor activity several hours before the actual light transition, indicative of light synchronization of the underlying circadian clock (Wheeler et al., 1993). As expected, control flies rapidly adjusted their activity pattern to the 6-h-delayed LD regime within 1-2 days, while homozygous cry^b/ cry^b flies required ~4-5 days before adjusting their evening activity peak to the shifted LD regime (Figure 2a). The slow resynchronization of cry^b mutants could be fully rescued by driving Drosophila UAS-cry expression with Clk856-gal4 (Figure 2a). We predicted that in this more sensitive assay, SpuCry and zebrafish cry4 would at least partially restore cry function in the fly and speed up resynchronization to delayed LD cycles in cry^b mutants, but this was not the case (Figure 2a). Quantifying the days required to reach a stable activity pattern in the shifted LD regime (i.e., after the jetlag), revealed no significant differences between homozygous cry^b/cry^b flies expressing none and those expressing any of the different cry genes (Figure 2c). To rule out the possibility that cry expression driven by Clk856-gal4 may not be strong enough, or spatially too restricted, we repeated these experiments using tim-gal4. Again, no improvement of light resynchronization was induced by any of the heterologous cry genes, while expression of Drosophila cry resulted in wild-type behavior (Figure 2b and 2d). Taken together, our results suggest that the zebrafish and sea urchin cry genes analyzed here are not able to restore light- and cry-dependent behavior in flies (Figures 1 and 2).

To investigate if any of the Cry proteins encoded by the sea urchin and zebrafish cry genes can support light responses of the molecular clock, we measured light-induced degradation of Tim protein in clock neurons. The various cry genes were expressed in cry^b mutant flies using the Clk856-gal4 driver, and Tim levels were determined by immunofluorescence late at night, when Tim levels reach their maximum (at Zeitgeber Time [ZT] 21, meaning 3 h before the lights came on in a 12 h:12 h LD cycle). These values were then compared to Tim levels in flies which were exposed to a 2 h of bright LP starting at ZT19. As expected, in cry^b mutant flies, Tim levels in all clock neurons were similar between the control and LP-treated flies (Yoshii et al., 2015). In contrast, Clk856-gal4-driven expression of Drosophila cry led to a strong reduction of Tim in all clock neuronal groups, indicating a rescue of Tim stabilization induced by cry^b (Figure 3). In agreement with our behavioral results (Figures 1 and 2), none of the zebrafish or sea urchin Cry proteins induced a clear reduction of Tim levels in the clock neurons of LP-treated flies. Although only the expression of zebrafish cry4 consistently showed a trend toward reduced Tim levels in all neuronal groups in the LP-treated flies, the difference to the non-pulsed controls was not significant, indicating that none of the heterologously expressed Cry proteins supports light-dependent Tim degradation.

Drosophila Cry also mediates molecular synchronization of peripheral clock cells to LD cycles (Ivanchenko et al., 2001). In fact, the original cry^b mutation was isolated in a screen for altered period-luciferase (per-luc) oscillations in peripheral clock cells during LD cycles. While wild-type flies displayed robust per-luc oscillations in LD, cry^b abolished these oscillations (Stanewsky et al., 1998). Because daily temperature cycles restored Per and Tim protein as well as per-luc oscillations in cry^b mutant flies, it followed that Cry is required for light-resetting of peripheral circadian clocks in flies (Glaser and Stanewsky, 2005; Ivanchenko et al., 2001; Stanewsky et al., 1998). To test if the zebrafish and sea urchin Crys support light synchronization of peripheral clocks in flies, we expressed them individually in cry^b mutant flies expressing the same per-luc reporter (BG-luc, containing 4 kb of 5’-flanking regulatory sequences and about two-third of the PER coding region fused to luciferase cDNA) used to isolate cry^b. Flies were placed individually in the wells of 96-well microtiter plates, and luminescence originating from each fly was measured once per hour during 2.5 days of LD followed by 4 days of DD. As expected, cry^b mutants showed no or low-amplitude per-luc oscillations when looking at average raw bioluminescence counts or de-trended and curve-fitted data, respectively, while wild-type flies expressed robust and light-dependent luciferase rhythms (Figure 4a and 4b). Expressing Drosophila Cry with the tim-gal4 reporter restored per-luc rhythms in cry^b mutant flies, confirming that this assay can be used to test the function of the zebrafish and sea urchin Crys in light synchronization of peripheral clocks (Figure 4A and 4B). As described for clock neuronal light responses earlier, only zebrafish cry4 showed limited ability to restore per-luc rhythmicity in cry^b flies (Figure 4B). Interestingly, SpuCry, zebrafish cry1a, and cry3 led to trough levels of per-luc expression during the LD part of the experiment, suggesting that the respective proteins can act as repressors of per expression (see below).

The mammalian type 2 Cry proteins function as essential, light-independent repressor proteins in the circadian clock. Mouse Cry1 and Cry2 repress transcription by binding to the transcription factors Clock and Bmal1 (Shearman et al., 2000). To directly test the possibility that the zebrafish and sea urchin Cry proteins can act as repressors of per transcription, we applied a per-luc reporter (plo), which faithfully reports per transcriptional rhythms (Brandes et al., 1996; Stanewsky et al., 1997). This plo reporter contains the same 4-kb upstream regulatory DNA sequences as BG-luc, which are directly fused to the luciferase gene (so no per coding sequences) (Brandes et al., 1996). plo transgenics exhibit robust oscillations in luminescence in LD, which rapidly dampen in DD (Figure 5). We expressed the various cry genes in plo flies using the tim-gal4 driver to see if this would cause a reduction of overall plo luminescence levels. As a positive control, we also expressed Drosophila UAS-per using the same tim-gal4 driver because Per is a known repressor of its own transcription (Zeng et al., 1994). As expected, and in agreement with previous observations (Zeng et al., 1994), overexpression of Per resulted in a drastic reduction of plo luminescence levels in the LD and DD parts of the experiment, while rhythmic expression was only maintained during LD (Figure 5a and 5b). Although Drosophila Cry has been shown to act as a transcriptional repressor (Collins et al., 2006), we did not observe a reduction of plo oscillation amplitude, nor decreased levels after overexpressing Drosophila cry, indicating that Drosophila Cry does not act as a repressor of per expression in the tim-expressing cells contributing to the bioluminescence signal. Strikingly, expression of zebrafish cry1a, cry3, and SpuCry had essentially the same effect on plo expression as overexpression of Per, strongly indicating that the Cry proteins encoded by these three genes can function as repressor of per transcription in Drosophila. The result for cry3 was surprising given its lack of repressive function when expressed in human cells (Kobayashi et al., 2000; Liu et al., 2015). Compared to wild-type controls, zebrafish cry4 also reduced plo levels, but this reduction was not significant, indicating that Cry4 has only weak repressor function in flies, if any (Figure 5a and 5b). In agreement with this result, Cry4 mainly localizes to the cytoplasm when expressed in fly clock neurons (Supplementary Figure S1↗).