The blue light photoreceptor CRYPTOCHROME is expressed in a subset of circadian oscillator neurons in the Drosophila CNS

Abstract

In the fruit fly Drosophila melanogaster, CRYPTOCHROME (CRY) functions as a photoreceptor to entrain circadian oscillators to light-dark cycles and as a transcription factor to maintain circadian oscillator function in certain peripheral tissues. Given the importance of CRY to circadian clock function, we expected this protein to be expressed in all oscillator cells, yet CRY cellular distribution and subcellular localization has not been firmly established. Here we investigate CRY spatial expression in the brain using a newly developed CRY antibody and a novel set of cry deletion mutants. We find that CRY is expressed in s-LN_vs, l-LN_vs and a subset of LN_ds and DN₁s, but not DN₂s and DN₃s. CRY is present in both the nucleus and cytoplasm of these neurons, and its subcellular localization does not change over the circadian cycle. Although CRY is absent in DN₂s and DN₃s, cry promoter activity and/or cry mRNA accumulation can be detected in these neurons, suggesting that CRY levels are regulated post-transcriptionally. Oscillators in DN₂s and DN₃s entrain to environmental light-dark cycles, which implies that they are entrained indirectly by retinal photoreceptors, extra-retinal photoreceptors or other CRY expressing cells.

Introduction

The fruit fly Drosophila melanogaster keeps circadian time via interlocked period/timeless (per/tim) and Clock (Clk) transcriptional feedback loops (reviewed in Hardin 2005). These feedback loops operate in the brain and many peripheral tissues to control behavioral, physiological and metabolic rhythms. Several groups of oscillator neurons in the lateral and dorsal brain control rhythms in locomotor activity (Helfrich-Forster 2005; Shafer et al. 2006). Oscillator neurons in the lateral brain include the dorsal lateral neurons (LN_ds), the large ventral lateral neurons (l-LN_vs), small ventral lateral neurons (s-LN_vs), and the lateral posterior neurons (LPNs). The s-LN_vs are further subdivided into four cells that express the neuropeptide Pigment Dispersing Factor (PDF), and a fifth s-LN_v that lacks PDF expression (the 5^th s-LN_v). Three additional clusters of oscillator neurons reside in the dorsal brain: dorsal neurons 1s (DN₁s), DN₂s and DN₃s. The s-LNvs are sufficient for locomotor activity rhythms in constant darkness (DD), whereas in 12h light: 12h dark (LD) conditions s-LN_vs control the morning peak of activity, and the 5^th s-LN_v and the LN_ds control the evening peak of activity (Grima et al. 2004; Rieger et al. 2006; Stoleru et al. 2004). Peripheral tissues that contain oscillators include sensory organs (e.g. compound eyes, ocelli, antennae, proboscis), testes, the gut, and Malpighian tubules (reviewed in Bell-Pedersen et al. 2005; Hall 2003). Oscillators in olfactory sensory neurons control rhythms in olfactory physiology (Krishnan et al. 1999; Tanoue et al. 2004), but rhythmic outputs from other peripheral oscillator tissues have not been characterized.

Circadian oscillators in the brain and peripheral tissues are entrained by environmental light-dark cycles. Photic entrainment in Drosophila is mediated by the blue light photoreceptor CRYPTOCHYROME (CRY), a conserved pterin/flavin containing protein found in other eukaryotes (Cashmore et al. 1999), and photoreceptors in compound eyes, ocelli and the Hofbauer-Buchner eyelet (Helfrich-Forster et al. 2001; Rieger et al. 2003; Veleri et al. 2007). CRY entrains the circadian oscillator by undergoing a light-dependent conformational change that promotes CRY-TIM binding, TIM ubiquitination, and ultimately the proteosome-mediated degradation of TIM (Busza et al. 2004; Ceriani et al. 1999; Koh et al. 2006; Lin et al. 2001; Peschel et al. 2006). Eliminating TIM destabilizes PER, which delays or advances the oscillator depending on whether light is presented during the early evening or late night, respectively. Consequently, mutants that alter or eliminate CRY show severe defects in their entrainment to light (Dolezelova et al. 2007; Emery et al. 2000b; Stanewsky et al. 1998).

In addition to its role as a circadian photoreceptor, CRY is necessary for circadian oscillator function in some peripheral tissues. In antennae, rhythms in per and tim expression and odor-dependent electroantennagram (EAG) responses are severely disrupted in cry^b mutant flies (Krishnan et al. 2001; Levine et al. 2002). Likewise, rhythmic PER and TIM accumulation in compound eyes and Malpighian tubules are abolished in cry^b flies (Ivanchenko et al. 2001; Stanewsky et al. 1998). Genes activated by CLK-CYC are expressed at high levels in cry^b flies and low levels in flies that overexpress CRY, which suggests that CRY represses CLK-CYC dependent transcription (Collins et al. 2006). Such a role in transcriptional repression is consistent with CRY function in mammals, where transcriptional activation by CLOCK-BMAL1 (the mammalian homolog of CLK-CYC) is repressed by mCry1 and mCry2 (Kume et al. 1999).

The dual roles of CRY as a circadian photoreceptor and, in many peripheral tissues, a transcriptional repressor implies that cry expression will encompass all oscillator cells. Several methods have been used to track cry expression in the brains (primarily) of adults including cry promoter driven Gal4 (cry-Gal4) reporter genes (Emery et al. 2000b; Klarsfeld et al. 2004; Zhao et al. 2003), cry RNA in situ hybridization (Egan et al. 1999; Zhao et al. 2003), and CRY immunohistochemistry (Klarsfeld et al. 2004), but disparate results within and between these methods leave no clear consensus on cry spatial expression. For instance, cry-Gal4 lines containing different amounts of cry upstream sequences are expressed in virtually all oscillator neurons in the brain and many peripheral tissues (Zheng et al., submitted), whereas cry-Gal4 lines that contain the first intron have a more restricted distribution (Emery et al. 2000b; Zheng et al., submitted). Moreover, different insertion sites of the same cry-Gal4 construct can produce substantial differences in spatial expression (Zhao et al. 2003). Among all cry-Gal4 lines the only common sites of expression are the LN_vs and LN_ds, though most cry-Gal4 lines are also expressed in subsets of DN₁s and DN₃s (Emery et al. 2000b; Helfrich-Forster et al. 2007; Klarsfeld et al. 2004; Shafer et al. 2006; Stoleru et al. 2004; Zhao et al. 2003). In situ hybridization experiments varied substantially in the number and location of cry positive cells, though at maximum cry signal was detected in LN_vs, LN_ds, DN₂s and a subset of DN₁s (Zhao et al. 2003). A novel peptide antibody detected strong CRY immunostaining in s-LN_vs and two DN₁s from adult brains, and weaker staining in l-LN_vs, LN_ds and additional DN₁s (Klarsfeld et al. 2004). In these neurons, CRY was detected in both the cytoplasm and nucleus, but time dependent changes in subcellular localization were not assessed. A weakness of previous cry mRNA and protein localization studies was the lack of a negative control in the form of a cry null mutant, though such a mutant was recently generated (Dolezelova et al. 2007).

Here we generate a novel CRY antiserum and cry deletion mutants to investigate CRY spatial expression and subcellular localization in adult brains. CRY immunostaining is detected in the LN_vs and a subset of LN_ds and DN₁s, but not DN₂s or DN₃s from wild-type flies, or any oscillator cells in cry^b mutant and cry deletion controls. CRY is detected in both the nucleus and cytoplasm at all times of the circadian cycle even though its light dependent molecular target, TIM, translocates from the cytoplasm to the nucleus during the night. These results reveal that CRY is only expressed in a subset of brain oscillator cells, indicating that some oscillator cells entrain to light indirectly. CRY is absent in DN₂s and DN₃s even though cry promoter activity and/or cry mRNA are detected in these neurons, suggesting that cry is under post-transcriptional control in DN₂s and DN₃s.

Materials and Methods

Fly stocks and transgenic flies

Flies that were wild-type with regard to clock function were either Canton-S or w¹¹¹⁸. The cyc⁰¹, per⁰¹, tim-Gal4, cry^b, and hscry strains have been described previously (Emery et al. 1998; Hazelrigg et al. 1984; Konopka and Benzer 1971; Park et al. 2000; Rutila et al. 1998; Stanewsky et al. 1998). cry^b+ flies, which are wild-type flies having the same genetic background as cry^b mutants, were generously provided by Dr. R. Stanewsky (School of Biological and Chemical Sciences Queen Mary College, University of London). Flies bearing the P-element P{XP}cry^d10630 (BL 19331) were obtained from the Bloomington Stock Center (Wilson et al. 2008). All fly strains were reared on medium containing corn meal, molasses, yeast, agar, and Tegosept (a mold inhibitor) at 25°C.

Generation and characterization of cry excisions

To generate cry deletions via imprecise excision, P{XP}cry^d10630, a w⁺ transposable element insert 57bp upstream of the cry transcription start site, was excised as described (Greenspan 2004). The approximate size of deletions due to imprecise excision was determined by PCR. To precisely determine the deletion endpoints in three different excision lines, PCR products containing the deleted sequences were cloned into pCR-Blunt II-TOPO vector (Invitrogen). For A2, the primer combination used was cryjumpF 5'-ATCTTTTCTCGCAGCCATTAT-3' and cryjumpR5 5'-CGATATTCAGCTCCCGACAT-3', and for B1 and I1 the primers used were cryjumpF3 5'-CCTGGCGTATGGGTAAGTG-3' with cryjumpR5. Inserts for each of the deletions were sequenced on both strands.

Fly heat shock

hs-cry flies were placed in an empty vial, pushing the cotton lid down so that the flies were close to the bottom. The vial was incubated on a 37°C water bath for 30 minutes. Immediately after that, flies were transferred to a 25 °C dark incubator, recovered for 2hr, collected and stored at −80 °C.

CRY antibody generation

Full length CRY was expressed in E. coli, purified and used to produce antisera in Guinea pigs. A DNA fragment containing the complete cry coding region was amplified from a head cDNA library (Hamilton'91) using the crypETs1 (5' GGAATTCACATATGGCCACGCGAGG 3') and crypETa1 (GGAATTCACTCGAGTCAAACCACCACGTC 3') primers. This fragment was inserted into the NcoI and XhoI sites of pET28a (+) to generate pET28a-cry, which expresses an N-terminal HIS-tagged CRY fusion protein. pET28a-cry was transformed into BL21 cells and protein expression was induced using IPTG. CRY fusion protein was purified from cell lysates using a copper affinity column (Amersham) as described (Houl et al. 2006). A single band at the predicted molecular weight of the CRY fusion protein (∼60kDa) was detected, and used to immunize guinea pigs for antibody production (Cocalico Biological Inc).

Whole mount immunofluorescence

Flies were entrained in a 12 hour Light: 12 hour Dark cycle (LD) for at least three days, released into constant darkness (DD), and collected at the indicated time points. Adult brains were dissected in 1XPBS pH7.4 followed by fixation with 5% formaldehyde in 0.1M Phosphate buffer during 20 min at room temperature. The specimens were then washed twice with 1xPBS (10 minutes each), and once with PBT (1XPBS pH7.4, 0.05% Triton-X100) for 10 minutes. Blocking was performed with 5% goat serum, 2% BSA in PBT at room temperature for 3hr or at 4°C O/N. Primary antibody dilutions were prepared in 2% goat serum in PBT, added to the brains, and incubated at 4°C for at least 48 hours. Dilutions used for the primary antibodies were 1/300 for rabbit anti PER that was preadsorbed against per⁰¹ (Cheng and Hardin 1998), 1/1000 for rabbit anti PDF, and 1/100 for guinea pig anti CRY. Brains were left for 2 hr at room temperature before washing off the primary antibody with PBT for at least 6 times (30 minutes each) at room temperature. Specimens were incubated with fluorescently labeled secondary antibody (1:300) diluted in PBT at 4°C overnight. Secondary antibodies used were: Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 488 and 568 goat anti-guinea pig (Molecular Probes, Invitrogen), Cy3 goat anti-rabbit and Cy3 goat anti-guinea pig (Jackson ImmunoResearch Laboratories Inc.). After equilibration for 2 hr at room temperature, brains were washed with PBT for 30 minutes at least 6 times at room temperature and mounted on glass slides using Vectashield (Vectorlab). Visualization of whole mounts was performed using an Olympus FV100 confocal microscope. Serial Z-stack images were obtained at 1−2um intervals. Images were processed using Adobe Photoshop. Images shown in the figures are overlays of several confocal stacks or single optical sections.

Behavioral analysis

Locomotor activity was assessed using Drosophila Activity Monitors (TriKinetics Inc.). Flies were entrained in LD cycles at 25°C for 3 days and then kept in constant darkness or constant light for at least 7 days. The data collected was analyzed with Fly Activity Analysis Suite for Mac OSX (FaasX), version 0.7.1 (Copyright 2002−2003 Boudinot, CNRS-INAF). Flies having powers >10 and widths >2 were considered rhythmic.

Real time PCR

cry mRNA levels were assayed by Quantitative Real Time PCR (RT-PCR). Flies were entrained in a 12hour light: 12hour dark cycle for at least three days and collected at Zeitgeiber Time (ZT) 5 and ZT17 (where ZT0 corresponds to lights on and ZT12 to lights off). Total RNA was isolated from fly heads using TRIZOL (Invitrogen). To eliminate genomic DNA contamination, each sample was treated with DNase (Promega). First-strand cDNA was synthesized from 2ug of RNA using random primers and Superscript™ II (Invitrogen) in an Eppendorf Mastercycler® gradient (Eppendorf). Ribosomal protein 49 (rp49) was used as the internal control. Reactions were carried out on an ABI Prism® 7000 Sequence Detector System (ABI). The sequences of the primers and probes used for the different messages are: cry forward 5'-CACCGCTGACCTACCAAA-3', cry reverse 5'-GGTGGAAGCCCAATAATTTGC-3', cry probe 5'6 FAM- TGTTCCTGCACACGGT-3'MGBNFQ; rp49 forward 5'-CTGCCCACCGGATTCAAG-3', rp49 reverse 5'-CGATCTCGCCGCAGTAAAC-3', rp49 probe 5'VIC-CCTCCAGCTCGCGCACGTTG-3'MGBNFQ. For mRNA quantification, a standard curve was prepared for cry and rp49. Target mRNAs and rp49 relative values were calculated based on the standard curve and the studied mRNA values were normalized to rp49.

Western blots

During the initial western blot experiments, head extracts were prepared as described (Emery et al. 1998), and blots were developed for chemiluminescence using ECL (Amersham). Subsequently, head extracts were prepared by homogenizing heads from ∼50 flies in 3 volumes of EB4 buffer (5mM Tris-HCl pH 7.5, 50mM KCl, 10mM HEPES pH 7.5, 10% Glycerol, 1mM EDTA, 1mM DTT, 0.05%Triton X100, 0.5mM PMSF, 0.01mg/ml Aprotinin, 0.01mg/ml Leupepetin, 0.002 mg/ml Pepstatin) using sterile pestles, centrifuged at 4°C for 15 minutes, and protein levels in the supernatant were quantified using the DC Protein Assay (Bio-Rad) or the Coomassie (Bradford) Protein Assay Kit (Pierce). 40−60 μg of protein per sample was separated on a 12% acrylamide gel, transferred to nitrocellulose, and blots were developed for chemiluminescence using Advanced Plus ECL (Amersham). Nitrocellulose membranes were blocked in 5% skim milk (or 2% blocking powder if using Advanced Plus ECL) for 1 hr at room temperature or overnight at 4°C, and incubated with 1:5000 dilution of CRY antiserum (or a dilution of 1:7500 for Advanced Plus ECL) overnight at room temperature. Blots were washed six times for 1 hr using 0.1% Tween 20 in TBS (TBST), and incubated with secondary antibody (HRP goat anti-guinea pig, SIGMA) diluted 1/1000 (for detection using ECL) or 1/5000 (for detection using Advanced Plus ECL) in blocking solution for 1 hr at room temperature. Blots were then washed six times for 1 hr in TBST, developed, and exposed to film.

Results

Generation of a CRY antibody (* Place Fig. 1 here, 3 inches high)

Figure 1.

GP23 antiserum detects CRY in fly head extracts. A. Western containing head extracts from hs-cry flies that were heat induced (+) or uninduced (−) at 37°C for 30 min. A ∼60 kDa band corresponding to CRY is detected. B. Western containing head extracts from wild-type flies that received (+) or did not receive (−) a 6hr light pulse at CT12 on the third day of DD before being collected at CT18. C. Western blot of white, cry^b and heat induced hs-cry flies collected at the times shown during a 12hr light: 12hr dark cycle and probed with GP23. The CRY and non-specific (ns) bands are shown. D. Western blot of white, cry^b and heat induced hs-cry flies collected at the times shown during the first day of DD and probed with GP23. The CRY band is shown. In panels B-D, proteins were extracted using the lab's standard procedure (see Material and Methods). All experiments were replicated at least three times with similar results.

Full length HIS-tagged CRY protein was produced in bacteria, purified, and used to immunize Guinea pigs. One antiserum, designated GP23, detected a protein of approximately 60 KDa on western blots that was induced to high levels in hs-cry flies upon heat shock, which suggests that this protein corresponds to CRY (Fig. 1A). CRY is destabilized by both light and the cry^b mutation (Emery et al. 1998; Stanewsky et al. 1998). After entrainment in LD, wild-type flies were given a 6hr light pulse on the third day of DD and collected at CT18 along with untreated controls. GP23 detected a strong band at ∼60kDa in control flies kept in DD, but this band was virtually eliminated after light treatment (Fig. 1B), consistent with detection of CRY. Likewise, when flies were collected every four hours during LD, the band corresponding to CRY rose to high levels during the night and fell to low levels by the end of the day (Fig. 1C). This band was detected at very low levels in cry^b flies (Fig. 1C), further demonstrating that it corresponds to CRY. Consistent with a very long half-life for CRY in DD, the band detected by GP23 accumulates during the first day of DD (Fig. 1D).

Generation and confirmation of cry null mutants (* Place Fig. 2 here, 1 column wide)

Figure 2.

Expression and structure of cry deletion mutants. A. qPCR of cry mRNA from the heads of w (WT), cry^b and cry excision flies (A1, A6, B1, B2, D2, D3, E2, I1) collected at ZT5 (white bars) or ZT17 (black bars) under LD conditions. Relative cry mRNA levels were quantified as described in Materials and Methods. Error bars are ±SD. B. Western blots of head extracts from w (WT), cry^b and excision flies (A1, A6, B1, B2, D2, D3, E2, I1) collected at CT19 during the first day of DD were probed with GP23. In panels A and B, B2 is a precise excision control. C. Structural characterization of cry deletion mutants. Diagram of the cry gene (top) showing the first two exons (white boxes), intron 1 (dashed line), upstream sequence (black line), the P{XP}cry^d10630 insert (triangle), the transcription start site (+1) and the translation start site (ATG). The extent of deleted sequences the A2, B1 and I1 strains are shown as gray lines.

Most functional studies of CRY are based on the cry^b mutant, which produces low levels of CRY that retain some CRY activity (Dolezelova et al. 2007; Emery et al. 1998; Stanewsky et al. 1998). However, cry^b may not serve as a useful negative control for immunolocalization studies since CRY antisera may detect the low levels of CRY^b protein. To avoid this potential problem, mutants having no detectable CRY expression were generated via imprecise P-element excisions (see Materials and Methods). Of 16 precise excision and 30 cry deletion lines, one precise excision line (B2) and seven deletion lines were selected for further characterization. The abundance of cry mRNA was measured in these eight lines at times when cry mRNA was at peak (ZT5) and trough (ZT17) levels. In the wild-type and B2 controls, cry mRNA is expressed at ∼2-fold higher levels at ZT5 than ZT17 (Fig. 2A). In contrast, cry mRNA from the seven deletion strains ranged from undetectable to 4-fold lower than cry mRNA in wild-type and B2 flies at ZT17 or cry mRNA in cry^b flies at ZT5 and ZT17 (Fig. 2A). Consistent with these cry mRNA results, CRY levels were very low or undetectable in the seven deletion strains and cry^b, but were high in wild-type and B2 flies (Fig. 2B). These results argue that the seven cry deletion strains examined are either null or severely hypomorphic.

Three lines that produce extremely low levels of cry mRNA, A2, B1 and I1, were characterized in detail. The A2 deletion removed 50bp of cry upstream sequence and 963bp of cry transcribed sequence, the B1 deletion removed 835bp cry upstream sequence and 881bp of cry transcribed sequence, and the I1 deletion removed 1209bp of cry upstream sequence and 1279bp of cry transcribed sequence (Fig. 2C). In each case, the deleted sequences remove the cry transcription and translation initiation sites, consistent with the loss of cry mRNA and protein expression in these strains. We will refer to these new cry alleles as cry^ΔA2, cry^ΔB1 and cry^ΔI1. (* Place Table 1 here, 1 column wide)

Table 1.

Locomotor activity rhythms of cry excision strains under LL conditions. Period and power were calculated for all rhythmic animals as described in Materials and Methods. Activity was calculated for all flies that survived to the end of the experiment, as described in Materials and Methods. Homozygous deletions are shown in bold. n.a.; not applicable.

Genotype	N	% Rhythmic (N)	Tau +/−SEM	Power	Activity
white	16	0	n.a.	n.a.	22.01+/−2.5
cryb	14	85.7 (12)	25.4+/−0.2	77.1+/−10.1	14.75+/−1.5
B2/+	10	0 (0)	n.a.	n.a.	19.41+/−1.7
B2/B2	17	0 (0)	n.a.	n.a.	24.3+/−2
A2/+	10	0 (0)	n.a.	n.a.	22.01+/−1.4
A2/A2	12	83.3 (10)	25.01+/−0.4	42.8+/−7.4	30+/−1.4
B1/+	8	0 (0)	n.a.	n.a.	30.2
B1/B1	17	70.6 (12)	26.3+/−0.7	50.3+/−8	33.7+/−2.4
I1/+	11	0 (0)	n.a.	n.a.	22.8+/−1.76
I1/I1	20	90 (18)	25.1+/−0.4	51.1+/−7.1	25.9+/−2.4

The loss of CRY function in cry^b flies enables rhythmic behavioral activity with slightly long periods in constant light and disrupts phase shifts to pulses of light (Emery et al. 2000a; Emery et al. 2000b; Stanewsky et al. 1998). Nevertheless, robust circadian activity rhythms with ∼24hr periods in cry^b flies indicates that cry is not required for oscillator function in neurons that mediate locomotor activity rhythms (Stanewsky et al. 1998). To determine whether the behavioral phenotypes of the cry^ΔA2, cry^ΔB1 and cry^ΔI1 mutants are similar to, or more severe than, cry^b, we tested behavioral rhythms under LL and DD conditions after entrainment in LD. Wild-type and B2 control flies were arrhythmic in LL, while ∼85% of cry^b flies were rhythmic with a period of 25.4hr (Table 1). Like cry^b flies, ∼70% to 90% of cry^ΔA2, cry^ΔB1 and cry^ΔI1 flies exhibited rhythmic locomotor activity under LL having periods of ∼25hr (Table 1). Under DD conditions, cry^ΔA2, cry^ΔB1 and cry^ΔI1 flies were mostly rhythmic with periods of ∼24.5−25.5hr., which was similar to the B2 control flies having the same genetic background, but longer period than either the wild-type or cry^b controls (Table 2). The similarity in LL and DD locomotor activity phenotypes between cry^b and the cry deletion lines suggests that cry^ΔA2, cry^ΔB1 and cry^ΔI1 are severely hypomorphic or null. (* Place Table 2 here, 1 column wide)

Table 2.

Locomotor activity rhythms of cry excision strains under DD conditions. Period and power were calculated for all rhythmic animals as described in Materials and Methods. Activity was calculated for all flies that survived to the end of the experiment, as described in Materials and Methods. Homozygous deletion strains are shown in bold.

Genotype	N	%Rhythmic (N)	Tau	Power	Activity
white	25	96 (24)	23.8+/−0.05	93.1+/−7.3	26.0+/−1.4
cryb	19	74 (14)	23.9+/−0.09	44.5+/−5.9	14.1+/−1.7
B2/+	18	83.3 (15)	24.4+/−0.13	61.7+/−8.5	24.4+/−2.2
B2/B2	29	55.2 (16)	24.8+/−0.11	64.77+/−7.9	25.7+/−1.7
B1/+	18	66.7 (12)	24.4+/−0.07	79+/−18.7	21.7+/−1.4
B1/B1	19	63.2 (12)	25.5+/−0.41	93.8+/−15.5	30.0+/−2.8
I1/+	26	88.5 (23)	24.3+/−0.08	50.6+/−6	18.8+/−1.6
I1/I1	32	62.5 (20)	24.7+/−0.11	61.0+/−4.7	27.3+/−1.9
A2/+	14	85.7 (12)	24.6+/−0.18	58.2+/−9.8	23.3+/−2.1
A2/A2	15	13.3 (2)	24.3+/−0.5	18.5+/−3.9	23.9+/−2

CRY localization in adult brains (* Place Fig. 3 here, 5 inches high)

Figure 3.

CRY expression in adult brains. Brains were dissected from w flies collected at CT19 on the third day of DD, immunostained with CRY (GP23) and PER antisera, and imaged via confocal microscopy. A minimum of six brains were analyzed with similar results. A. Images of a 2μm z-stack projection through the left hemisphere are shown, where dorsal is at the top. Colocalization of CRY (green) and PER (red) immunofluorescence is seen in the merged image as yellow. Arrows indicate CRY and/or PER immunostaining in oscillator neurons. s-LN_v, small ventrolateral neurons; l-LN_vs, large ventrolateral neurons; LN_ds, dorsolateral neurons; DN1, dorsal neuron1s; DN2, dorsal neuron 2s; DN3, dorsal neuron 3s. B. A compressed stack of six 1 μm optical sections through the lateral protocerebrum is shown. Colocalization of CRY (green) and PER (red) immunofluorescence is seen in the merged image as yellow. Arrows indicate CRY and/or PER immunostaining in oscillator neurons. Oscillator neurons are labeled as in panel A. C. A compressed stack of six 1μm optical sections through the dorsal protocerebrum is shown. Colocalization of CRY (green) and PER (red) immunofluorescence is seen in the merged image as yellow. Arrows indicate CRY and/or PER immunostaining in oscillator neurons. Oscillator neurons are labeled as in panel A. D. Images of a 2μm z-stack projection through the right hemisphere are shown, where dorsal is at the top. CRY and PER immunostaining and oscillator cells are as described in panels A-C, CRY-IR is also detected in the Pars Intercerebralus (PI) and the medial protocerebrum (MP).

In situ hybridization and promoter analyses show that cry is expressed almost exclusively in brain oscillator cells, though the number and identity oscillator cells that express cry vary among studies (Emery et al. 2000b; Zhao et al. 2003; Zheng et al., submitted). To determine whether GP23 can detect CRY expression in oscillator cells, we used this antibody along with antiserum against the clock cell marker PER to immunostain brains from flies collected at CT19 on the third day of DD. Strong CRY immunoreactivity (CRY-IR) was detected in many, but not all PER expressing brain neurons (Fig. 3A-C). A magnified image of the lateral brain region shows CRY-IR in all of the s-LN_vs, l-LN_vs, and some LN_ds (Fig. 3B). In the dorsal brain region, CRY-IR is detected in a majority of the DN₁s, but not in DN₂s or DN₃s (Fig 3C). Two of the DN₁s show more intense CRY-IR than other DN1s, and likely correspond to DN₁a neurons based on their position (Shafer et al. 2006). CRY-IR was uniform within all CRY positive oscillator neurons, suggesting that CRY is present in both the cytoplasm and the nucleus at this time. Projections from oscillator neurons also showed CRY-IR (Fig. 3C), consistent with previous results (Klarsfeld et al. 2004). However, CRY-IR could only be detected in neuronal projections from flies kept for at least 3 days in DD, suggesting that such staining is due to inordinately high levels of CRY that have accumulated over several days. CRY-IR is also detected in non-PER expressing cells in the Pars Intercerebralis (PI) and medial protocerebrum (Fig. 3D). These cells were quite large, and CRY-IR was cytoplasmic. (* Place Fig. 4 here, 5 inches high)

Figure 4.

CRY expression in cry^b and cry deletion strains. Brains were dissected from cry deletion strains cry^ΔI2 and cry^ΔA2, cry^b mutant flies, and cry B2 (precise excision) flies collected at CT19 on the third day of DD, immunostained with CRY (GP23) and PER or PDF antisera, and imaged via confocal microscopy. A minimum of four cry^ΔI2, four cry^ΔA2, and six cry^b brains were analyzed with similar results. A. A 2μm z-stack projection through the left hemisphere of a cry^ΔI2 fly is shown. Colocalization of PER (green) and CRY (red) immunofluorescence is seen in the merged image as yellow. Arrows indicate PER and/or CRY immunostaining in oscillator neurons, which are labeled as in panel A of Fig. 3. B. A compressed stack of 6−8 1μm optical sections through the Pars Intercerebralus is shown. A merged image of CRY (green) and PER (red) immunofluorescence is shown. C. A compressed stack of 3−6 1μm optical sections through the lateral protocerebrum of cry B2 and cry^ΔA2 flies is shown. Colocalization of PDF (green) and CRY (red) immunofluorescence in s-LN_vs is seen in the merged image as yellow. D. A compressed stack of 3−4 2μm optical sections from the right brain hemispheres of cry^b flies is shown. Colocalization of CRY (green) and PER (red) immunofluorescence in the merged image is shown as yellow. Arrows indicate PER and/or CRY immunostaining in oscillator neurons, which are labeled as in panel A of Fig. 3.

To confirm that CRY-IR in adult brains corresponds to CRY, GP23 and PER were used to immunostain brains from cry^ΔI1 adults collected on the fourth day of DD. No CRY-IR was detected in neurons that were immunostained with PER (Fig. 4A). However, large PER negative cells in the PI and medial protocerebrum of cry^ΔI1 flies continued to show CRY-IR similar to that in wild-type flies (Fig. 4B), indicating that this immunostaining is non-specific. Brains from cry^ΔA2 and control B2 excision flies collected on the third day of DD were immunostained with PDF and GP23 antisera to confirm the lack of CRY-IR in LN_vs using a separate cry deletion strain. Although CRY-IR was seen in LN_vs from B2 controls, no CRY-IR was detected in LN_vs from cry^ΔA2 flies (Fig. 4C). In addition, no CRY-IR was detected in oscillator cells from cry^b flies collected on the third day of DD (Fig. 4D). The lack of CRY-IR in cry^ΔA2, cry^ΔI1, and cry^b flies demonstrates that GP23 specifically detects CRY.

Localization of CRY in brain oscillator neurons during the circadian cycle (* Place Fig. 5 here, 5 inches high)

Figure 5.

CRY subcellular localization over the circadian cycle. A. Brains were dissected from w flies collected every 4 hours starting at CT1 on the third day of DD, immunostained with CRY (GP23) and PER or PDF antisera, and imaged via confocal microscopy. A minimum of five brains were analyzed for each region with similar results. A. 1μm optical sections of DN₁s in the dorsal protocerebrum and LN_ds, l-LN_vs and s-LN_vs in the lateral protocerebrum. Colocalization of CRY (green) and PER (red) immunofluorescence is seen in the merged images as yellow. B. A compressed stack of 2−31μm optical sections showing s-LN_vs in the lateral protocerebrum. Colocalization of PDF (magenta) and CRY (green) immunofluorescence is seen in the merged image as pink.

Since TIM localization shifts from the cytoplasm to the nucleus during the night and is degraded during the morning (Hunter-Ensor et al. 1996; Meyer et al. 2006; Myers et al. 1996; Shafer et al. 2002; Zeng et al. 1996), it is possible that CRY undergoes similar changes in subcellular localization so that it can always access TIM. To test this possibility, brains from flies collected every 4hr during the third day of DD were immunostained with PER and CRY antisera. Consistent with previous results (Shafer et al. 2002; Veleri et al. 2003; Yang and Sehgal 2001), high amplitude rhythms in PER abundance and nuclear localization were seen in s-LN_vs, LN_ds and DN₁s, but not l-LN_vs (Fig. 5A). However, CRY abundance and subcellular localization to both the nucleus and cytoplasm were constant in s-LN_vs, l-LN_vs, LN_ds, and DN₁s. Constant levels of CRY in l-LN_vs and s-LN_vs is supported by experiments using PDF as a cytoplasmic marker; CRY abundance and subcellular localization do not cycle in PDF positive cells during the third day of DD (Fig. 5B). These results demonstrate that in brain oscillator neurons, the levels and subcellular localization of CRY remain constant in DD.

Discussion

CRY expression in brain oscillator cells

Here we employed a novel CRY antibody and newly generated cry deletion mutants to define the pattern of CRY expression in adult brains. CRY immunostaining was detected in most, but not all, groups of oscillator neurons in the adult brain. Specifically, CRY is present in all of the s-LN_vs, l-LN_vs, but only about half of the LN_ds and DN₁s, and none of the DN₂s or DN₃s (Fig. 3AD). Our CRY immunostaining results are consistent with those of Klarsfeld et al., in which CRY-IR could be detected in LN_vs, and some LN_ds and DN₁s on the third day of DD, but not DN₂s or DN₃s (Klarsfeld et al. 2004). We also detect CRY immunostaining in s-LN_v projections into the dorsal brain. The significance of this observation is not known, but it is likely this staining is simply a consequence of high CRY accumulation in DD over several days since we did not observe such staining under LD cycling conditions (data not shown).

The lack of CRY expression in half the DN₁s, and all DN₂s and DN₃s suggests that the oscillator in these neurons is light-entrained indirectly by other CRY expressing neurons, retinal photoreceptors in the compound eye or ocelli, or extraretinal photoreceptors in the Hofbauer-Buchner eyelet. When retinal and extraretinal photoreceptors are removed using the gl^60j mutant, oscillator function persists in DN₂s, demonstrating that DN₂ oscillators are not light-entrained by retinal or extraretinal photoreceptors (Helfrich-Forster et al. 2001). However, in cry^b flies oscillator function is lost in DN₂s, suggesting that these neurons entrain indirectly through other CRY expressing neurons (Helfrich-Forster et al. 2001). Since s-LN_vs are not able to reset oscillator phase in DN₂s (Stoleru et al. 2005), other CRY expressing neurons must entrain DN₂s.

We expected cry mRNA to be expressed in all CRY-expressing brain cells. However, different cry promoter driven Gal4 lines (with or without cry intron 1) give variable patterns of GFP reporter gene (i.e. UAS-GFP) expression, with some lines expressing exclusively in a single cluster of brain oscillator neurons, and others expressing in most or all of the brain oscillator neuron clusters along with different groups of non-oscillator cells (Emery et al. 2000b; Klarsfeld et al. 2004; Zhao et al. 2003). Recent analysis of different portions of the cry promoter (with or without the first intron) driving Gal4 show consistent expression of a β-galactosidase (i.e. UAS-lacZ) reporter gene within all six clusters of brain oscillator neurons (Zheng et al., submitted). In situ hybridization analysis detects cry mRNA in all three groups of LNs, DN₁s and DN₂s, but not DN₃s (Zhao et al. 2003). Since cry-Gal4 driven reporter genes and cry RNA in situ hybridizations detect cry expression in s-LN_vs, l-LN_vs, LN_ds and DN₁s, it is not surprising that CRY is present in these neuronal clusters. However, we do not detect CRY in DN₂s even though cry expression can be detected via reporter gene and RNA in situ analyses. Similarly, CRY is not detected in a group of non-oscillator cells in the dorsal margin of the optic lobe (DOL cells) that express cry mRNA and cry promoter dependent reporter genes (Zhao et al. 2003). These results suggest that CRY is either not translated or is unstable in DN₂s and DOLs. Although cry-Gal4 driven reporter gene expression can be detected in DN₃s (Klarsfeld et al. 2004; Zheng et al., submitted), no cry mRNA is detected by in situ hybridization (Zhao et al. 2003), indicating that cry mRNA is unstable in these cells. Taken together, these studies indicate that cry expression is regulated at the transcriptional or post-transcriptional level depending on the cluster of brain oscillator neurons.

CRY-IR was consistently observed in non-oscillator cells within the Pars Intercerebralis (PI), a region of the dorsal protocerebrum containing neurosecretory cells (de Velasco et al. 2007), and the medial protocerebrum. Unlike brain oscillator cells, CRY-IR in these cells was very intense and always cytoplasmic. However, in contrast to a loss of CRY-IR in oscillator cells from cry^b and cry deletion flies, CRY-IR remained in the cytoplasm of cells in the PI and medial brain (Fig. 4B), suggesting that CRY-IR in these cells is not due to CRY. These results suggest that the GP23 antibody cross-reacts with a protein present in these cells.

Subcellular localization of CRY during the circadian cycle

During the circadian cycle, TIM accumulates in the cytoplasm during the early evening, then moves into the nucleus until it is degraded during the early morning (in LD) or the subjective early morning (in DD) (Hunter-Ensor et al. 1996; Meyer et al. 2006; Myers et al. 1996; Shafer et al. 2002; Zeng et al. 1996). However, CRY does not move from the cytoplasm to the nucleus each day, but is present in both the nucleus and cytoplasm at all times of the circadian cycle (Fig. 5A, B). CRY nuclear and cytoplasmic localization provides constant access to TIM, thus allowing for phase resetting in response to light at any time during the circadian cycle. Since CRY controls oscillator function in some peripheral tissues by rhythmically repressing CLK-CYC activity in the nucleus (Collins et al. 2006), it is possible that rhythms in CRY nuclear translocation could occur in peripheral tissues.