cryptochrome

The clock input to the first optic neuropil of Drosophila melanogaster expressing neuronal circadian plasticity

In the first optic neuropil (lamina) of the fly's visual system, two interneurons, L1 and L2 monopolar cells, and epithelial glial cells show circadian rhythms in morphological plasticity. These rhythms depend on clock gene period (per) and cryptochrome (cry) expression. This study found that rhythms in the lamina of Drosophila may be regulated by circadian clock neurons in the brain since the lamina is invaded by one neurite extending from ventral lateral neurons; the so-called pacemaker neurons. These neurons and the projection to the lamina were visualized by green fluorescent protein (GFP). GFP reporter gene expression was driven by the cry promotor in cry-GAL4/UAS-GFP transgenic lines. It was observed that the neuron projecting to the lamina forms arborizations of varicose fibers in the distal lamina. These varicose fibers do not form synaptic contacts with the lamina cells and are immunoreactive to the antisera raised against a specific region of Schistocerca gregaria ion transport peptide (ITP). ITP released in a paracrine way in the lamina cortex, may regulate the swelling and shrinking rhythms of the lamina monopolar cells and the glia by controlling the transport of ions and fluids across cell membranes at particular times of the day (Damulewicz, 2011).

This study showed a single projection from the pacemaker cells in the brain to the lamina, in which several structural circadian rhythms have been detected. Moreover, this input probably originates from the 5^th small LN_v. Since the 5^th s-LN_v does not express PDF, this cell is different from the other LN_vs. The possibility that this process originates from other clock cells, for example from the LN_ds, and extends to the aMe first, and next to the lamina cannot be excluded. A CRY-positive LN_d, which is immunoreactive to ITP, could invade the lamina by passing the aMe first. This neuron, however, is also immunoreactive to sNPF, but the projection detected in the lamina is immunoreactive to ITP only. It indicates that this projection originates from the 5^th s-LN_v, which is immunoreactive to ITP but not to sNPF. This study examined GFP expression driven by cry-GAL4 in thin, 20 microm cryostat sections and thick 100 microm vibratom sections of the Drosophila brain. In most earlier studies on clock neurons and their projections, whole-mount preparations of the Drosophila brain were used, or the lamina was cut-off during preparation. Such procedures from previous studies meant that the very fine projection from the brain to the lamina could not be observed. This study detected the projection by using 20 microm sections and collecting confocal optical sections at a 1 microm interval (Damulewicz, 2011).

In several previous studies, it has been suggested that CRY is present in different types of clock neurons. These results have been obtained using various methods; cry-GAL4 driven GFP expression, cry mRNA in situ hybridization, immunolocalization and cry deletion mutants. Using cry-GAL4 line and 20 microm sections of the D. melanogaster brain, it was found that CRY is located in all s-LN_vs, l-LN_vs, LN_ds, DN₁s and DN₃s but is absent in DN₂s and LPNs. These results only partly confirm the results of earlier studies. It has been shown that LN_vs but only some DN₁, and three or four from the six LN_d are CRY - positive, while DN₂, DN₃ and LPNs are CRY-negative. One study did d not detect CRY in DN₂s and DN₃s, and in about half of the LN_ds and DN₁, but cry promoter dependent reporter genes and cry mRNA can be detected in these neurons. In this study, all of LN_ds showed GFP fluorescence in the cry-GAL4 strain, but only 3-4 cells were found to be CRY-immunopositive using antibodies. In turn, using the in situ hybridization method, cry mRNA was not detected in those cells. Since the pattern of cry-GAL4 driven GFP expression depends on the transgene insertion site and whether the first intron of the transgene has been inserted, spatial and circadian regulation of cry was examined. A series of cry-GAL4 transgenes containing different portions of cry upstream and intron 1 sequences was examined. The first intron was shown to drive expression in eyes and antennae, and upstream sequences induce cry expression in brain clock neurons and in peripheral oscillators; in eyes and antennae. In addition, upstream sequences also induce expression of cry, in other non-clock cells in the optic lobe (Damulewicz, 2011).

The results obtained using various methods suggest that in the case of CRY, translation and cry transcription may be specifically regulated. CRY-positive labeling in the 4^th LN_d was observed in flies kept for 5 days in constant darkness. Flies kept longer in this condition brought on weak staining in one of the DN₂ neurons. Thus, the level of CRY in this neuron may be very low, and the CRY level may only be detected after it has accumulated for several days in DD. It is possible, that in some of the LN_ds, DN₁ and DN₃ cry expression is very low and protein is undetectable by the immunohistochemistry method, or that cry mRNA is unstable and CRY protein is not synthesized. Among six LN_ds, three neurons, that show a strong signal of GFP in the brain cryostat sections used in this study, may correspond to CRY-positive cells detected in the studies of other authors. In turn, three LN_ds with weak GFP in these preparations may correspond to CRY immunonegative cells. These cells had about a 50% lower GFP level than the rest of the LN_ds at all time points, except at ZT4 when their GFP fluorescence was lower by 20% (Damulewicz, 2011).

Beside neurons, clock genes have also been detected in glial cells. A subpopulation of glial cells in the brain of Drosophila have rhythmic expression of per gene, and they are necessary for maintaining circadian locomotor activity. However, the presence of CRY in glia was not detected in this study. In the optic lobes, GFP driven by cry-GAL4 was observed in many non-clock cells in which the localization pattern was very similar to the distribution of glial cells. But these non-clock cells were not labeled with the antibody against REPO protein, a specific marker for glial cells. The REPO protein is required for glia development and differentiation and has been detected in all types of glia in the adult brain of Drosophila. The analysis of cry-GAL4 driven GFP and REPO immunolabeling showed no co-localization between CRY and REPO. However, in the close vicinity of GFP-positive cells, REPO-positive glial cells were observed. A similar result was obtained using the antibody against the Drosophila vesicular monoamine transporter (DVMAT), which enabled labeling the fenestrated glia in the optic lobe. These results suggest that CRY is present in non-clock neurons in the optic lobe, but not in glial cells (Damulewicz, 2011).

In addition to localization of cry-GAL4 driven GFP in cell bodies of neurons, GFP processes were also detected invading three neuropils in the optic lobe. In the medulla, a dense network of processes originate from DN₃s and their terminals seem to form synaptic contacts with not-yet identified target cells. The regular network of processes was also detected in the lobula but their origin is unknown. The most interesting finding is the projection of CRY-positive processes to the lamina. Although the lamina showed robust circadian remodeling of neuron morphology, a circadian input had not been previously detected. In the lamina, per is probably expressed in the epithelial glial cells, however, maintaining the lamina structural rhythms also requires per expression in the retina photoreceptors and in the LNs (Damulewicz, 2011).

Beside PER, CRY is also important for circadian rhythms in the lamina. In an earlier study, it was shown that the circadian rhythm in morphological plasticity of L2 dendritic trees, is not present in per⁰¹ mutant while its phase depends on CRY. In cry^b mutant, the pattern of daily changes in size of the L2 dendritic tree was different than in wild-type Canton-S flies. In males and females of Canton-S wild-type flies, the largest L2 dendritic tree was found at the beginning of the day. This daily pattern of the structural changes of L2 dendrite resembles the pattern of cry mRNA cycling in Drosophila heads and bodies, and in the 5^th s-LN_v detected in this study. Although the L2 dendritic tree is the largest at the beginning of the day in the distal lamina, its axon, as well as the axon of L1 monopolar cell, swell at the beginning of both day and night. These changes have been detected in the proximal lamina. Moreover, the α-subunit of the Na⁺/K⁺-ATPase and subunits of the V-ATPase also show diurnal changes in abundance in the lamina. Such an occurrence indicates that circadian rhythms in cell structural plasticity are correlated with rhythmic changes in the level of proteins involved in the transport of ions. The rhythm in the α-subunit of the Na⁺/K⁺-ATPase level is bimodal with two peaks; in the morning and in the evening. This pattern is changed in the cry⁰ mutant. It indicates that CRY is not only important for the maintenance of the daily pattern of morphological changes of the L2 dendritic tree but CRY also helps to maintain cycling of the Na⁺/K⁺-ATPase in the epithelial glial cells in the lamina (Damulewicz, 2011).

It is uncertain whether there is regulation of lamina rhythms by the brain pacemaker because connections between the pacemaker neurons in the accessory medulla and the lamina have not been observed. It was found, however, that rhythms in axon plasticity of neurons in the lamina are circadian, have two peaks (morning and evening) and are synchronized with locomotor activity. The present results now show, that thin neurite extends from the aMe and arborizes in the distal lamina. In the aMe, the s-LN_vs are regarded as the main pacemaker cells maintaining circadian rhythms. The l-LN_vs are involved in behavioral arousal and sleep. For these reasons, the LN_vs are good candidates as oscillators controlling lamina rhythms. Moreover, all LN_vs except the 5^th s-LN_v, express PDF which may synchronize central oscillators with each other and with peripheral ones. In the housefly, large PDF-immunoreactive neurons, similar to Drosophila's l-LN_vs, have terminals in the lamina which show circadian structural changes. Moreover, these neurons cyclically release PDF that affects circadian plasticity in the lamina. In Drosophila, release of PDF from PDF-immunoreactive processes in the medulla, where these processes form a dense network of varicose processes, is also possible. These processes, however, do not extend to the lamina. In the present study, PDF immunolabeling of the newly described Drosophila's CRY-positive terminals in the lamina was negative. This does not exclude PDF action in the lamina, particularly when PDF receptors have been detected in non-neuronal cells between the lamina and the retina. PDF may diffuse in the lamina after release from terminals in the distal medulla (Damulewicz, 2011).

Ion transport peptide (ITP) and short neuropeptide F (sNPF) have been detected in the LN_vs. Among the five s-LN_vs, ITP was found in the 5^th s-LN_v, while sNPF was observed in four other s-LN_vs which also express PDF. In the present study, ITP-immunoreactive fibers were detected, using the Schgr-ITP antisera, in the distal lamina, co-localized with cry-GAL4 driven GFP. The co-localization with ITP suggests that the projection into the lamina may originate from the 5^th s-LN_v. Little is known about the function of the 5^th s-LN_v. It has been suggested, that this neuron, together with LN_ds and some DN₁s, drive the evening peak of D. melanogaster bimodal activity. The finding indicates a possible new function of the 5^th s-LN_v in regulating circadian structural rhythms in the lamina, since this neuron is immunoreactive to ITP. Like other peptides in the optic lobe, ITP seems to be released from varicose terminals in a paracrine way. This conclusion was reached because no synaptic contacts between ITP-immunoreactive processes and cells in the lamina were detected. This peptide probably diffuses in the distal lamina and may facilitate chloride and/or other ion-dependent swelling and shrinking of the L1 and L2 axons. At least two ion pumps; the V-ATPase and Na⁺/K⁺-ATPase, show robust cyclical activity in the epithelial glial cells. The epithelial glial cells swell and shrink in anti-phase to the L1 and L2 interneurons. Preliminary results showed that in a transgenic line carrying RNAi to block ITP expression, the pattern of rhythmic changes in the level of the α-subunit of the Na⁺/K⁺-ATPase in the lamina glial cells of Drosophila is different than the pattern in wild-type flies. Thus, not only CRY but also ITP is important for maintaining rhythmic activity changes of the Na⁺/K⁺-ATPase (Damulewicz, 2011).

The function of ITP in the nervous system is unknown. In the lamina ITP may play a similar regulatory role as in hindgut of insects, transporting ions and fluids across cell membranes (Damulewicz, 2011).

Since the L1 and L2 monopolar cells swell in the morning and in the evening, ITP released from the 5^th s-LN_v may drive the evening peak of this rhythm. This is thought to be so, because the 5^th s-LN_v and LN_d are regarded as the lateral neurons' evening oscillator. In turn, PDF may drive the morning peak because PDF is thought to control the morning peak of locomotor activity, in a LD 12:12 regime. However, PDF's role in promoting locomotor activity in the evening has also been shown. The role of ITP as a neurotransmitter of circadian information to the lamina and as a possible regulator of rhythmic swelling and shrinking of the L1 and L2 monopolar cells, requires more experimentation and will be the subject of the next study (Damulewicz, 2011).

Transcriptional Regulation

How is Cryptochrome mRNA cycling affected by mutations in four clock genes implicated in gene regulation: per, tim, Clock, and cycle? In all single mutants and double mutant combinations, little or no mRNA cycling is found, indicating that cycling requires a functional pacemaker and is not merely light driven. cry mRNA levels are a function of the specific mutant or mutant combination. They are relatively low in the per or tim null mutants as well as in the per;tim double mutant combination, whereas they are relatively high in the Clock and cycle mutants. The double mutants per;Clock and per;cycle also show high cry mRNA levels, indicating an epistatic effect of Clock and cycle over per. Thus, CRY mRNA levels are low in per and tim null mutants, the opposite of what is observed for autoregulation of PER and TIM mRNA levels. CRY mRNA levels are high in clock or cycle mutants, contrary to the low PER and TIM mRNA levels found in these novel clock mutants (Emery 1998 and references).

A recessive third chromosomal mutation that abolishes bioluminescence rhythms has been identified, cry^b. cry^b is an apparent null mutation in a gene encoding Drosophila's version of the blue light receptor cryptochrome. To determine the mutation's effects on per and tim transcription, a per-luc or a newly generated tim-luc fusion gene (each encoding luciferase sequences only) were introduced into homozygous mutant genetic backgrounds. luc-reported expression in both cases is arrhythmic. In contrast to other recently identified mutations affecting per and tim expression (Allada, 1998 and Rutila, 1998), the new mutation does not give rise to profound subnormalities in overall levels of per and tim expression in mutant flies. Nevertheless, western blot analyses using head extracts of mutant flies maintained in LD show that the levels of Tim and Per protein remain at high levels throughout the day and night, relative to the very low troughs observed during the daytime in wild type. In addition, Tim and Per proteins are anomalously present in both hypo- and hyperphosphorylated forms in a temporally unchanging manner. That Tim stays at the same levels during the day and night in the mutant is especially interesting, because the rapid disappearance of this protein in response to light is the earliest response to this stimulus of a known component of Drosophila's rhythm system. Yet the absence of rhythmic clock gene transcription indicates that the mutant is doubly defective. This is because either of two regulatory phenomena is sufficient to drive Tim cycling (reviewed by Young, 1998): oscillating tim expression (which occurs in the absence of environmental fluctuations) or light suppression of TIM (in the absence of tim mRNA cycling). Against this background, the absence of effects of some (but not all) orthodox visual mutations on light-induced Tim degradation is notable (Yang, 1998), as is the fact that peak sensitivity for this light effect is in the blue range (Suri, 1998). Thus, the new mutation might uniquely affect elements of the light entrainment pathway, which would include extraocular reception and processing of blue light inputs. Alternatively, the mutation could affect a protease whose targets include Tim and Per (Stanewsky, 1998).

If that is not the case, and the new mutation causes a specific defect in the light entrainment pathway, protein oscillations in temperature cycles should not be affected. Western blots of extracts from mutant and normal heads showed that Per and Tim fluctuated robustly in 12 hr:12 hr, 25°C:20°C cycles; such cyclings continued in constant conditions. The daily mobility shifts of Per and Tim are apparent in both wild type and mutant genetic backgrounds, indicating that the phosphorylation program can function in the mutant (Stanewsky, 1998).

Targets of Activity

Although most circadian clock components are conserved between Drosophila and mammals, the roles assigned to the Cryptochrome (Cry) proteins are very different: Drosophila Cry functions as a circadian photoreceptor, whereas mammalian Cry proteins (mCry1 and 2) are transcriptional repressors essential for molecular clock oscillations. This study demonstrates that Drosophila Cry also functions as a transcriptional repressor. RNA levels of genes directly activated by the transcription factors Clock (Clk) and Cycle (Cyc) are derepressed in cryb mutant eyes. Conversely, while overexpression of Cry and Period (Per) in the eye repressed Clk/Cyc activity, neither Per nor Cry repressed individually. Drosophila Cry also represses Clk/Cyc activity in cell culture. Repression by Cry appears confined to peripheral clocks, since neither cry^b mutants nor overexpression of Per and Cry together in pacemaker neurons significantly affected molecular or behavioral rhythms. Increasing Clk/Cyc activity by removing two repressors, Per and Cry, leads to ectopic expression of the timeless clock gene, similar to overexpression of Clk itself. It is concluded that Drosophila Cry functions as a transcriptional repressor required for the oscillation of peripheral circadian clocks and for the correct specification of clock cells (Collins, 2006).

Several pieces of evidence point to Drosophila Cry, like its mammalian counterparts, functioning as a repressor of Clk/Cyc-activated transcription: (1) expression of four Clk/Cyc target genes is derepressed in cry^b mutants; (2) overexpression of cry together with per is sufficient to repress tim and vri expression in the eye, and this is supported by Cry repressing Clk/Cyc-activated transcription in transfected cells, either alone or in conjunction with Per; and (3) removing both Cry and Per leads to ectopic tim expression in the brain (Collins, 2006).

Although Cry and Per seem to function together to repress Clk/Cyc activity, the results do not imply a direct interaction between Cry and Per proteins. Drosophila Cry-Per interactions have been detected in yeast, but Cry and Per appear to interact only via Tim in vivo. Furthermore, Per continues to repress Clk/Cyc activity in vivo during the first half of the day, presumably after Cry has been degraded by light. Thus, Cry and Per seem to control distinct steps in repression of Clk/Cyc activity, with Cry probably initiating, and Per maintaining, repression. Further experiments will be required to test whether Tim also facilitates repression. While in vitro studies indicated that Tim helps remove Clk/Cyc from DNA, in vivo studies of the tim^UL mutant suggests that Tim does not participate in repression of per and tim transcription and instead stabilizes Per and facilitates its nuclear entry. Given that Drosophila Tim interacts with both Per and Cry in vivo, it will be interesting to test whether the Per-Cry interactions detected in mammalian clock cells are mediated via mTim (Collins, 2006).

Very little is known about the developmental specification of clock neurons. Per and Cry normally restrict tim expression to cells that adopt a circadian cell fate. The results complement experiments in which overexpression of Clk led to ectopic tim expression, since they reveal that cells not normally destined to develop as clock cells repress Clk/Cyc activity during development. However, there must be additional factors that contribute to clock cell fate, since the ectopic Tim+ve cells in per⁰¹; cry^b double mutant larvae did not produce PDP1. Similarly, there must be unidentified factors that maintain repression of tim in nonclock cells, since repression of Clk/Cyc activity will prevent further per expression. The presence of extra Tim-expressing cells may also explain the Tim-dependent rhythmic behavior of per⁰¹; cry^b in LD cycles, since ectopic Tim expression influences LD behavior (Collins, 2006).

The findings that Cry functions as a repressor in Drosophila are supported by the high conservation across species of the “core” photolyase-like domain of Cry, which is sufficient for repression in Xenopus. TheDrosophila cry^m mutation removes most of the Cry C terminus and interferes with Cry's response to light. However, Cry^M still supports a functional clock in the eyes, suggesting that the remaining core of Cry^M functions as a transcriptional repressor (Collins, 2006).

Cry's homology with DNA photolyases has led to the suggestion that Cry was the original molecule that allowed organisms to respond to light -- primitive organisms could detect light and regulate gene expression with one molecule (Cry) to avoid damage by sunlight during light-sensitive processes such as DNA replication. While ancestral Cry may have acted as both a light sensor and repressor, non-Drosophilid insects such as the monarch butterfly Danaus plexippus have two cry genes and divide repressor/light sensor function between them. Thus, circadian clocks may well have their origins in rapid responses to light, and the anticipatory clock gene networks could have subsequently been built around Cry, a light-responsive protein and a transcriptional repressor, the function of which has gradually become specialized (Collins, 2006).

The cryptochrome (cry) gene and a mating isolation mechanism in tephritid fruit flies

Two sibling species of tephritid fruit fly, Bactrocera tryoni and Bactrocera neohumeralis, are differentiated by their time of mating, which is genetically determined and requires interactions between the endogenous circadian clock and light intensity. The cryptochrome (cry) gene, a light-sensitive component of the circadian clock, was isolated in the two Bactrocera species. The putative amino acid sequence is identical in the two species. In the brain, in situ hybridization shows that cry is expressed in the lateral and dorsal regions of the central brain where Per immunostaining is also observed and in a peripheral cell cluster of the antennal lobes. Levels of cry mRNA were analyzed in whole head, brain, and antennae. In whole head, cry is abundantly and constantly expressed. However, in brain and antennae the transcript cycles in abundance, with higher levels during the day than at night, and cry transcripts are more abundant in the brain and antennae of B. neohumeralis than in that of B. tryoni. Strikingly, these results are duplicated in hybrid lines, generated by rare mating between B. tryoni and B. neohumeralis and then selected on the basis of mating time, suggesting a role for the cry gene in the mating isolation mechanism that differentiates the species (An, 2004).

Identification of novel genes involved in light-dependent CRY degradation through a genome-wide RNAi screen

Circadian clocks regulate many different physiological processes and synchronize these to environmental light:dark cycles. In Drosophila, light is transmitted to the clock by a circadian blue light photoreceptor Cryptochrome (Cry). In response to light, Cry promotes the degradation of the circadian clock protein Timeless (Tim) and then is itself degraded. To identify novel genes involved in circadian entrainment, an unbiased genome-wide screen was performed in Drosophila cells using a sensitive and quantitative assay that measures light-induced degradation of Cry. The expression of ~21,000 genes was systematically knocked down and those that regulate Cry stability were identified. These genes include ubiquitin ligases, signal transduction molecules, and redox molecules. Many of the genes identified in the screen are specific for Cry degradation and do not affect degradation of the Tim protein in response to light, suggesting that, for the most part, these two pathways are distinct. The effect of three candidate genes on Cry stability was further evaluated in vivo by assaying flies mutant for each of these genes. This work identifies a novel regulatory network involved in light-dependent Cry degradation and demonstrates the power of a genome-wide RNAi approach for understanding circadian biology (Sathyanarayanan, 2008).

Of the ubiquitin-conjugating enzymes identified in this screen, a composite E2-E3 ligase, Bruce, and a novel HECT domain-containing E3 ligase CG17735 had the strongest effects on light-dependent Cry degradation. Neither of these ligases had a significant effect on the Tim light response in flies. Bruce exerts its anti-apoptotic role through its N-terminal IAP domain or by targeting proapoptotic proteins, such as SMAC2, for degradation using its C-terminal E2-E3 ligase activity. In flies, apart from its anti-apoptotic activity, it also plays an important role in regulating sperm individualization, a process by which the sperm nuclear syncytium differentiates into individual sperm (Huh, 2004). This is similar to the apoptotic pathway proteins, DIAP1, DIAP2, and Dredd, which perform a nonapoptotic role in the IMD immune pathway. Thus a nonapoptotic role for Bruce in regulating Cry stability is not surprising. It is interesting to note that Bruce knockdown in S2 cells has a small effect on baseline levels but primarily blocks light-induced Cry degradation, suggesting that Cry is rendered stable and nonresponsive to light in the absence of Bruce. The genetic evidence from loss-of-function mutations of Bruce in the fly suggest that Bruce affects the stability of Cry in LD cycles. The same appears to be true for the dual specific phosphatase SSH. This effect on Cry in LD cycles most likely also represents impaired degradation in response to light. It is inferred that the mechanisms involved in the Cry response to parametric and nonparametric entrainment may be somewhat different, and the cell culture assay may allow detection of molecules required for either mechanism. The fact that all these candidates identified in the cell culture screen have the same general effect on Cry stability in cells and flies demonstrates the power of the screen for identifying novel circadian regulatory molecules (Sathyanarayanan, 2008).

CG17735 affects the degradation of Cry by light in cell culture as well as in flies. Given that this gene encodes a putative E3 ligase, it is tempting to speculate that this is the E3 ligase that targets Cry for degradation in response to acute light exposure. However, until a null mutant of CG17735 can be obtained, the extent to which it is required for the regulation of Cry remains unknown. Nor can the behavioral consequences of completely blocking Cry degradation be known. It is intriguing that the knockdown of CG17735 is associated with a long period, but at this point it is unclear if this is due to the effect on Cry or on some other, as yet unidentified, clock target (Sathyanarayanan, 2008).

Although various components of the Cry degradation complex were identified, the importance of light-dependent Cry degradation to circadian resetting behavior is not clear. Bruce mutants demonstrate a strong defect in Cry degradation; however, due to the weak behavioral rhythms in these mutants, the effect on circadian photic sensitivity could not be addressed. Also it is noted that light-dependent Cry degradation, although reduced, still persists in BruceRB mutants, suggesting that it is regulated by redundant pathways. The other two mutants assayed (ssh and CG17735) did not completely abolish Cry degradation either, but in this case, it may be because neither is a null mutant. Interestingly, Tim degradation was not affected in any of these mutants, suggesting that they specifically affect Cry degradation. Tim is targeted for light-dependent degradation by an F-box-containing protein, Jet, which has minimal effect on Cry degradation. This result again supports the existence of two distinct destruction complexes that regulate Cry and Tim degradation. However, some aspects of the degradation pathways may be similar. The demonstration that SkpB, a component of the SCF complex, is critical for both Cry and Tim degradation suggests that some components are shared between the two destruction complexes. In addition, components required for the activation of Cry are presumably also required for the Tim light response. It is likely that other classes of molecules identified in the screen, such as those that affect redox status, are involved in Cry activation and therefore also affect Tim and the behavioral response. Further analysis of these molecules in clock neurons should address the role of redox and Cry degradation in circadian photosensitivity (Sathyanarayanan, 2008).

Protein Interactions

The biological clock synchronizes the organism with the environment, responding to changes in light and temperature. Drosophila Cryptochrome (Cry), a putative circadian photoreceptor, interacts with the clock protein Timeless (Tim) in a light-dependent manner. Although Tim dimerizes with Period (Per), no association between Cry and Per has previously been revealed, and aspects of the light dependence of the Tim/Cry interaction are still unclear. Behavioral analysis of double mutants of per and cry suggest a genetic interaction between the two loci. To investigate whether this is reflected in a physical interaction, a yeast-two-hybrid system was employed that revealed a dimerization between Per and Cry. This is further supported by a coimmunoprecipitation assay in tissue culture cells. The light-dependent nuclear interactions of Per and Tim with Cry require the C terminus of Cry and may involve a trans-acting repressor. Thus, as in mammals, Drosophila Cry interacts with Per, and, as in plants, the C terminus of Cry is involved in mediating light responses (Rosato, 2001).

At 25°C, per^S;cry^b flies display predominantly 24 hr cycles in a LD 12:12 regime, although ~40% of the flies also have a minor 19 hr per^S component. These two periodic components are not found together in either single mutant (Rosato, 2001).

Locomotor activity was monitored for per^S, cry^b, and the per^S;cry^b double mutants at 18°C and 28°C. Single mutant per^S or cry^b flies entrain to the LD 12:12 regime at both temperatures, showing a 24 hr period and distribution of activity around the times of light/dark transitions. In DD, they free-run with a period of about 24 hr for cry^b and a period of about 19 hr for per^S, with a modest temperature dependence. In DD conditions, per^S;cry^b flies behave virtually identically to per^S mutants at both temperatures. However, the behavior of the double mutant changes dramatically in LD, in a temperature-dependent manner. At 18°C, all per^S;cry^b flies show a periodic component of about 24 hr, but about 60% of them also display a minor 19 hr component. At 28°C, 79% of the rhythmic flies display the endogenous 19 hr period as the main rhythmic component. The breakdown of entrainment at 28°C in double mutant flies could reflect a genuine genetic interaction between the cry^b and per^S mutations. Alternatively, perhaps the limits of entrainment at high temperature are reduced in cry^b mutants so that per^S;cry^b flies might indeed entrain to a T cycle of 20 hr at this temperature (which is closer to the 19 hr endogenous period of per^S), whereas cry^b individuals (whose endogenous period is ~24 hr) might not. To test this hypothesis, the locomotor activity rhythms of single and double mutant flies was monitored at 28°C under an LD 10:10 regime. Both per^S and cry^b flies entrain under this condition. However, the double mutants may show some evidence of entrainment during the first two cycles of the new light/dark regime, but any entrainment soon breaks down, and the per^S;cry^b flies free-run, with their daytime activity advancing by about 90 min on each successive day. Therefore, the entrainment defect at high temperature shown by per^S;cry^b flies is the product of a specific interaction between the two mutations rather than a defect in the entrainment of cry^b alone. In Drosophila, the visual system is involved in the reception of circadian-relevant light information. This system is perfectly functional in the double mutant and is revealed by the startle response that is evident at the transition points from dark to light (and vice versa) at both temperatures. Therefore, per^S;cry^b flies are able to detect light but are deficient in the transmission of light information to the clock mechanism in a temperature-dependent manner (Rosato, 2001).

The genetic interaction between per^S and cry^b prompted an investigatation of the possibility of a physical interaction between Per and Cry using a yeast-two-hybrid system. A full-length Cry protein, directly fused to LexA (bait), was challenged with Per(233-685) as prey. This fragment includes the major protein/protein interaction domains described for Per. A fragment of Tim(377-915) that is known to bind to Per and contains the relevant regions for Per/Tim dimerization as prey was also tested. No interactions were observed between LexA-Cry and both Per(233-685) and Tim(377-915) fragments in the dark. Cry has been shown to interact with full-length Tim, but not Per, under constant light. In light, LexA-Cry binds strongly to Per(233-685), but not to Tim(377-915). LexA-Cry was also challenged with full-length Per and Tim, both in darkness and light. No interactions were observed in the dark. Under constant light, only full-length Tim showed evidence of dimerization with LexA-Cry. Three conclusions are drawn from these results: Per and Tim interactions with LexA-Cry are light dependent; the N and/or the C terminus of Tim are required for the association with LexA-Cry, and there is an inconsistency between the results obtained from full-length Per and the fragment Per(233-685). In regard to the latter, the well-established Per/Tim interaction was retested using LexA-Tim bait with Per and Per(233-685) preys in darkness and light. No interactions were observed using full-length Per. Subsequent Western blot analysis has revealed that, in this system, full-length Per is poorly expressed, thereby explaining the lack of interactions in yeast with this construct. Nevertheless, a strong interaction between LexA-Cry and Per(233-685) could be demonstrated. This discrepancy between the current results and contradictory published results must reside in the different yeast-two-hybrid systems employed. Evidence was also found for a Tim-independent Cry/Per complex using coimmunoprecipitation (Rosato, 2001).

Cryptochromes are believed to interact with a signaling factor after light exposure, and evidence has been found in plants for a role of the C-terminal domain in signaling. Since the coimmunoprecipitation result supports the view that the interaction between LexA-Cry and Per(233-685) in yeast reflects a meaningful association between Per and Cry, the power of yeast genetics was exploited to test the regulatory role of the C terminus of Drosophila Cry. Twenty residues were deleted from the Cry C terminus to create CryDelta and it was challenged with Per(233-685) and full-length Tim in darkness and light. An interaction was evident in both conditions, with no obvious difference between them. It has been suggested that LexA-Cry^b is unable to interact with Tim in yeast cells because it may have lost its photoresponsiveness. Both LexA-Cry^b and LexA-Cry^bDelta, which are strongly expressed in yeast, are nevertheless unable to interact with Per(233-685) or with Tim. Given the light independence of CryDelta, it is suggested that the D[410]N substitution in Cry^b probably confers a gross structural defect to LexA-Cry^b, rather than simply affecting its photoreceptor ability (Rosato, 2001).

To further map the interaction between Cry and Per, LexA-CryDelta was challenged with several overlapping Per fragments. It was confirmed that LexA-Tim (377-915) interacts with the PAS A domain (Per[233-390]) and Per(233-685). LexA-CryDelta does not associate with Per(233-390), nor with the PAS A + B region (Per[233-485]), but interacts with the downstream C domain (Per[524-685], which includes the per^S site. From these results, it is speculated that Tim and Cry may interact with different regions of the Per protein and, since Cry associates with region(s) of Tim external to the (377-915) fragment, it is hypothesized that Per, Tim, and Cry can be found in the same complex (Rosato, 2001).

LexA-Cry requires light in order to interact with Per(233-685) and Tim. However, it cannot be ruled out a priori that it is the temperature increase, caused by the continuous light exposure, rather than light per se, that triggers Cry's interactions. LexA-Cry was therefore challenged with Per(233-685) and Tim at 37°C in the dark, but no interactions were observed. Furthermore, since LexA-CryDelta does not require light, this variant was used to investigate the effect of temperature on Cry interactions (Rosato, 2001).

Yeast patches were grown on X-gal plates at 30°C and 37°C in parallel. It was noted that at 37°C, the LexA-CryDelta interaction with Per(233-685) is considerably weakened, whereas the control LexA-Tim(377-915)/Per(233-685) dimerization does not show any substantial temperature differences. The same temperature dependence is also observed when LexA-CryDelta is challenged with Tim and Per (524-685) (Rosato, 2001).

To further identify those regions of Cry that could suppress the negative effect of darkness, random Taq-induced mutations were introduced into full-length cry by PCR, and LexA-Cry* mutants were created by in vivo gap repair. The putative LexA-Crys* were challenged with Per(233-685) in the dark. A total of 14 bona fide light-independent mutations were identified that generated a Cry/Per interaction in darkness. The sequencing of these variants shows that all of these light-independent Crys* carry either a translational stop or a frame-shift at their C termini. Some of the mutants have additional amino acid substitutions scattered across the entire sequence, but because of their sporadic nature, it is very unlikely that these missense mutations are contributing to the light-independent phenotype (Rosato, 2001).

The results reported above support the view that the C terminus of Cry is responsible for the light dependence of the interactions with Per and Tim. Perhaps the removal of the C terminus changes Cry conformation to a form that is active in darkness. Alternatively, there could be a carboxy-terminal-bound, light-inhibited nuclear repressor of Cry in yeast. In fact, trans-acting factor in yeast can be mutated to disinhibit nuclear Cry activity in darkness, and currently, attempts are being made to identify the gene(s) involved (Rosato, 2001).

Thus, it has been shown that Cry binds Per in yeast and in a Drosophila cell culture system. As in yeast, the light-dependent activities of Cry in S2 cells have been reported only in the nucleus, where Cry is suggested to undergo a conformational change after light absorption, allowing it to bind to Tim (and now Per). However, Cry coimmunoprecipitates with Tim and Per in the cytoplasm of S2 cells under darkness, suggesting that light is not required to change Cry into its active conformation. Consequently, both in yeast and S2 cells, it is predicted that a nuclear factor may interact with the Cry C terminus in darkness to prevent it from interacting with the two clock proteins. Cry itself is probably not its own repressor, because full-length Cry was tested in a yeast-two-hybrid assay and it does not significantly self-associate in light or dark. However, mutagenesis of the yeast genome has identified two variants that can derepress the Cry/Per interaction in darkness. Isolation of this gene(s), irrespective of its function in yeast, will provide candidates for this nuclear repressor(s), which might have a clock relevant homolog(s) in Drosophila. An analogous situation has been reported in which Saccharomyces cerevisiae casein kinase I, HRR25 (without known clock function in yeast) binds and phosphorylates Per with affinities similar to the Drosophila casein kinase Iepsilon, Doubletime (Dbt). The signaling mechanism of cryptochrome is also mediated through the C terminus in Arabidopsis. A fusion between ß-glucuronidase (GUS) and the C-terminal domain (CCT) of either Cry1 or Cry2 (to create CCT1 and CCT2) mediates a constitutive light response. This means that 'isolated' CCTs display properties in the dark that are strikingly similar to those of light-activated Crys. Within the Cry molecule, the C-terminal domain is repressed under darkness, and light activation might be achieved either by an intramolecular or an intermolecular redox reaction, but the details of the light-induced activation of CCT are not known. In this study, it has been shown that the intermolecular model is the more appropriate to explain observations with Drosophila Cry. Light-induced activation of Cry removes a regulatory molecule, enabling the binding of Per and Tim, although the possibility exists that the regulatory molecule itself, rather than Cry, could act as the primary photopigment. It will be of interest to see if this model also applies to Arapidopsis cryptochromes. The C-terminal domain of Cry thus becomes a focal point for further studies, and it is probably not a coincidence that it is this region of the otherwise evolutionary conserved Cry molecule that is the most variable (Rosato, 2001).

It cannot be unequivocally concluded that the physical interaction revealed between Per and Cry is responsible for the genetic interaction that occurs in per^S;cry^b mutants at high temperature, even though this was the experiment that led the authors to test a possible Per/Cry dimerization. However, the Per/Cry interaction is temperature-sensitive in yeast, and it is the Per C domain (which includes the site of per^S) that dimerizes with Cry, providing further circumstantial evidence that the genetic interaction between Per and Cry may correlate with the physical interaction. Furthermore, it is tempting to speculate that differences in the Per^S/Cry physical interaction may be at the heart of reports that the per^S mutants are hypersensitive to light and that flies carrying a small deletion (amino acids 515-568) within the Per C domain display short, poorly temperature-compensated rhythms and an altered behavioral response to light pulses. Perhaps a reduction in the strength of the Per^S/Cry association, further decreased at 28°C below a critical threshold, might account for the entrainment defect of per^S;cry^b double mutants at high temperature. Finally, genetic interactions between short-period mutations per^S and per^T and the arrhythmic mutation dbt^ar implicate the C domain in the dynamics of Per phosphorylation by DBT. Taken together, these results suggest that the Per C domain may provide a convergence point for both Cry and Dbt, and it is anticipated that future research may disclose a prominent role for Cry in the fly circadian clock (Rosato, 2001).

Oscillations of the period (per) and timeless (tim) gene products are an integral part of the feedback loop that underlies circadian behavioral rhythms in Drosophila melanogaster. Resetting this loop in response to light requires the putative circadian photoreceptor Cryptochrome (Cry). The early events in photic resetting were dissected by determining the mechanisms underlying the Cry response to light and by investigating the relationship between Cry and the light-induced ubiquitination of the Tim protein. In response to light, Cry is degraded by the proteasome through a mechanism that requires electron transport. Various Cry mutant proteins are not degraded, and this suggests that an intramolecular conversion is required for this light response. Light-induced Tim ubiquitination precedes Cry degradation and is increased when electron transport is blocked. Thus, inhibition of electron transport may 'lock' Cry in an active state by preventing signaling required either to degrade Cry or to convert it to an inactive form. High levels of Cry block Tim ubiquitination, suggesting a mechanism by which light-driven changes in Cry could control Tim ubiquitination (Lin, 2001).

The profile of Cry protein expression in light-dark (LD) cycles suggests that the protein is unstable in the presence of light. Light-induced instability of Cry is supported by experiments in which Cry was expressed in S2 cells. Cells were transfected with a pIZ-cry construct, in which Cry is tagged with a V5 epitope, and it was noted that levels of the protein were reduced by light treatment. To identify the mechanisms that degrade Cry, Cry-transfected cells were treated with light in the presence of proteasome inhibitors MG115 and lactacystin. Both these agents were effective in blocking Cry degradation. This effect of light on both Tim and Cry suggests that the action of a ubiquitin/proteasome degradation pathway may be one of the first events in photic resetting in Drosophila (Lin, 2001).

Light-induced ubiquitination of Tim takes place in S2 cells and CRY mRNA is expressed in S2 cells. Consistent with the presence of endogenous Cry in S2 cells, it was found that light-dependent inhibition of Per-Tim feedback activity, which is known to be Cry dependent, occurs to a significant extent in the absence of transfected Cry. The presence of endogenous photic signaling mechanisms in S2 cells provides a system in which the degradation of transfected Cry could be studied regardless of its ability to transduce a photic signal (Lin, 2001).

To determine whether a specific region of Cry mediates its degradation, different Cry mutants were transfected and their responses to light were assayed. Cry-N (amino acids 1 to 423) has the C-terminal 119 amino acids deleted and Cry-C (amino acids 244 to 542) has the N-terminal 243 amino acids deleted. The cry^b mutation is a missense mutation within the sequence that encodes the highly conserved flavin-binding region of Drosophila Cry. It corresponds to the original cry mutation isolated through a genetic screen of Drosophila. Treatment with light does not reduce the levels of any of these mutant Cry proteins. For Cry-N the levels are consistently low with and without light treatment, indicating a general instability of the protein. Cry-C is expressed at high levels, and levels of Cry^b are equivalent to those of the wild type. However, none of these proteins show a response to light. Since there is no common sequence that is deleted in all these constructs, this indicates either that Cry degradation requires more than one part of the molecule or that the overall conformation of the molecule is important for its recognition by the degradation system. The large deletions in Cry-N and Cry-C may prevent such a conformation change. For Cry^b, the mutation is thought to prevent association with flavin, which may be required for a redox-mediated conformation change (Lin, 2001).

Cry signaling in plants requires redox activity and is mediated, at least in part, by the flavin moiety bound to Cry. This is based on the finding that DPI, which inhibits the transport of electrons from reduced flavin, is effective in blocking Cry-mediated photic signaling in Arabidopsis. In addition, flavins participate in electron transport in other systems, the most relevant system being that of the photolyases, which are homologous to Crys and which are known to repair DNA through an electron transfer mechanism (Lin, 2001).

Based on the data implicating electron transport in Arabidopsis Cry signaling, the effect of electron transport inhibitors on degradation of Drosophila Cry were determined. DPI attenuates Cry degradation. Thus, photic signaling by Drosophila Cry involves redox activity, most likely mediated by the flavin (Lin, 2001).

The presence of endogenous Cry in S2 cells supports the idea that the light-dependent Tim ubiquitination is mediated by Cry. To determine if the Tim response to light requires Cry, Tim levels were assayed in light-pulsed and unpulsed cry^b flies. While wild-type flies show the characteristic decrease in Tim levels with light treatment, this response is lacking in cry^b flies (Lin, 2001).

The S2 cell system was used to determine the relationship between light-induced Cry degradation and Tim ubiquitination and degradation. One possibility considered was that Cry is required for Tim stability. In this model, light-induced Cry degradation would lead to Tim degradation, perhaps by exposing relevant sites on Tim to phosphorylation and ubiquitination events. Although Tim and Cry do not bind each other in the dark in the yeast two-hybrid system, they can be coimmunoprecipitated from S2 cells, suggesting that they are present in the same complex. Thus, removal of Cry in response to light could affect Tim processing. Alternatively, light exposure may lead to some conformational and/or redox changes in Cry that trigger downstream events, including Tim ubiquitination and Cry degradation. To distinguish between these two possibilities, the time course of Tim ubiquitination and that of Cry were examined for degradation in S2 cells. An increase in Tim ubiquitination is detected within 5 min of light exposure, while Cry levels in the same extracts remain unchanged up to the end of a 30-min light pulse. Thus, overall degradation of Cry does not appear to be required for Tim ubiquitination. The possibility that Cry is removed from a complex with Tim cannot be excluded: it is more likely that in response to light, Cry transmits a signal that leads to Tim ubiquitination (Lin, 2001).

No degradation of Tim in S2 cells could be detected in response to light. This may be due, in part, to the HA tag on the ubiquitin, which could interfere with proteasomal digestion. However, other researchers have also noted that Tim is not turned over upon light exposure in S2 cells. Extended incubation (up to 6 h post-Tim induction) of transfected cells results in Tim degradation in both dark- and light-treated cells (Lin, 2001).

Tim ubiquitination was examined in the presence of electron transport inhibitor DPI. Tim ubiquitination is increased by DPI, although Cry degradation is blocked, which is consistent with the idea that Tim ubiquitination does not require degradation of Cry. In fact, the increased Tim ubiquitination is most likely due to the accumulation of activated Cry, effected through a block either in degradation or in the reconversion of Cry to an inactive form (Lin, 2001).

Light-induced Tim ubiquitination in S2 cells is thought to be mediated by endogenous Cry. To test the effects of increasing Cry levels on light-induced Tim ubiquitination, S2 cells were cotransfected with hs-tim, hs-Ub, and different concentrations of either pIZ-cry or hs-cry and Tim ubiquitination was assayed 2 h after light exposure (Lin, 2001).

High concentrations of both hs-cry and pIZ-cry decrease Tim ubiquitination. However, ubiquitination of Tim is enhanced when hs-cry is transfected. pIZ-cry does not increase Tim ubiquitination when transfected at low concentrations, most likely because this plasmid yields higher levels of Cry expression. Taken together these observations indicate that small increases in Cry promote Tim ubiquitination after 2 h of light exposure but that high levels attenuate it. However, even in the presence of high levels of Cry, Tim ubiquitination increases during the first 10 to 15 min of light treatment. The block at later time points in Cry-overexpressing cells is indicative of a deficit in the maintenance of Tim ubiquitination, which may be due to enhanced deactivation of Cry (Lin, 2001).

The data on the differential effects of low and high Cry concentrations are supported by results of Cry overexpression in transgenic flies. Flies that overexpress Cry under control of the tim promoter show enhanced resetting, while those that express Cry under the actin 5c promoter show a reduction of light-induced phase delays. The difference in the phenotypes of these two overexpression strains may lie in the level of overexpression. To determine whether the reduced resetting in the actin 5c line correlates with reduced Tim degradation in response to light, flies carrying a UAS-cry construct were crossed to others carrying an actin 5c promoter-GAL4 transgene and the resulting progeny was assayed for Tim expression. Tim expression was examined at different times of day by Western blotting of adult fly head extracts. In Cry overexpression flies Tim levels are considerably higher than wild-type levels at time points early in the day but equivalent to wild-type levels at all other time points. Thus, the effect is specific for the early part of the day, when Tim is normally turned over in response to light (Lin, 2001).

Degradation of Cry by light invokes analogies with the plant photoreceptors, phytochrome (PHY) and Cry, both of which are degraded in response to light. Thus, it may be a common mechanism to control levels of the photoreceptor and thereby the strength of the photic response. Moreover, as noted here for Cry, PHY is known to be degraded by the proteasome (Lin, 2001).

The role of the proteasome in degradation of both Cry and Tim also underscores similarities with the cell cycle. The cell cycle is characterized by cycling proteins that undergo phosphorylation and subsequent degradation, in many cases by the proteasome. Both Per and Tim are cyclically phosphorylated and phosphorylation plays a role in turnover of both proteins. For Tim, light-induced degradation is effected through an increase in phosphorylation and ubiquitination. Thus, as for the cell cycle, multiple proteins in the circadian cycle are turned over by the ubiquitin-proteasome pathway. However, Per turnover may utilize a different pathway since ubiquitination of Per has not been observed (Lin, 2001).

The mechanisms that lead to Cry degradation in response to light are not clear, but it is hypothesized that a conformation change in Cry is required. A light-induced conformation change is supported by the following lines of evidence: (1) in the yeast two-hybrid system Cry interacts with full-length Tim in the presence of light but not in the dark; (2) sequences that mediate Cry degradation do not appear to map to a unique part of the molecule, suggesting that the tertiary structure is important; (3) the Cry^b protein is not degraded by light. All these mutants were tested in the presence of endogenous Cry, and so their ability to signal was dissociated from their degradation. Although the single amino acid mutated in the flavin-binding region in Cry^b could play a direct role in the degradation process, it is far more likely that it affects a flavin-mediated conformation change. The fact that Cry^b does not associate with Tim in the yeast two-hybrid system is consistent with an inability to undergo a conformation change (Lin, 2001).

Cry has been shown to block Per and Tim autoregulation of their own RNA synthesis in a light-dependent manner in S2 cells. Since Tim degradation is not detectable in S2 cells, it has been suggested that the inhibition of Tim activity by Cry, rather than its degradation, is the primary response to light. This block in Per-Tim activity may be the immediate response of the clock to light. Presumably this block persists as long as the photic signals are present and Cry is not degraded. However, a phase change of several hours, which can be produced with a pulse of <1 min of light, must require an irreversible change in a clock component. Tim is ubiquitinated in S2 cells within 5 min of light treatment. In flies, Tim degradation (which presumably follows ubiquitination) occurs within 30 to 60 min of light treatment and is apparently critical for resetting the clock. A Cry molecule with a functional flavin-binding domain is required for this response (Lin, 2001).

Signaling by flavins frequently involves a redox change. In fact, a reagent that blocks the transfer of electrons from reduced flavin prevents Cry degradation by light. At the same time, it increases Tim ubiquitination. Based on the recently proposed models for Arabidopsis Cry, DPI may block either intramolecular electron transport required for a change in Cry conformation or intermolecular transport to a signaling pathway that effects degradation. Assuming that active Cry, which promotes Tim ubiquitination, is produced by a conformation change, it is suggested that the DPI-sensitive step occurs after the conformation change. It should be noted that DPI can also block the activity of other flavoproteins, such as NADPH oxidase and nitric oxide synthase, that play a role in redox processes (Lin, 2001).

Cryptochromes (CRYs) are flavoproteins important for the molecular clocks of animals. The Drosophila cryptochrome (Cry) is a circadian photoreceptor, whereas mouse cryptochromes (mCRY1 and mCRY2) are essential negative elements of circadian clock transcriptional feedback loops. The Drosophila circadian clock involves an autoregulatory feedback loop, in which Period (Per) and Timeless (Tim) inhibit their own transcription by associating with the transcriptional activators Clock (Clk) and Cycle (Cyc). Light affects the feedback loop by causing rapid degradation of Tim by the proteasome. Drosophila cryptochrome appears to mediate this light effect, as suggested by its binding to Tim and Per in yeast and in Drosophila Schneider 2 (S2) cells, its own light-dependent degradation by the proteasome in S2 cells, and its light-dependent ability to inhibit Per:Tim transcriptional repression in S2 cells. It has been proposed that reduction/oxidation (redox) reactions are important for Cry light responsiveness and mCRY1 transcriptional inhibition. Therefore the role of redox in light-dependent activation of Cry and in mCRY1 transcriptional inhibition was evaluated in Drosophila Schneider 2 cells. Using site-directed mutagenesis, three of the four conserved flavin binding residues in dCRY were found to be essential for light responses, whereas three of the four corresponding residues in mCRY1 did not abolish transcriptional responses. Two tryptophan residues in dCRY are critical for its function and are likely involved in an intramolecular redox reaction. The corresponding tryptophan residues do not play a redox-mediated role in mCRY1 function. The data provide a multistep redox model for the light-dependent activities of Cry and suggest that such a model does not apply to mCRY1 transcriptional responses (Froy, 2002). Intramolecular redox is critical for Cry functions. Photolyases repair UV light-induced pyrimidine dimers through intermolecular redox between the reduced flavin, FADH^-, and DNA. However, in bacterial photolyases, an intramolecular redox pathway has also been discovered. Under purification conditions or oxidative stress, the flavin molecule is oxidized into FAD^- and becomes catalytically inert. Upon irradiation with white light, the photolyase regains its activity by transferring to the flavin an electron through an internal chain of three tryptophans. To study whether electron transfer through the corresponding tryptophans might be involved in Drosophila Cry function, these tryptophan residues were mutated to alanine in Cry (W342A, W397A, and W420A) and the activity of the mutant proteins was evaluated in S2 cells, using the transcriptional assay and a proteolysis assay (Froy, 2002).

Cry-W342A and Cry-W397A are deficient in mediating light-induced transcriptional derepression, showing no significant light responsiveness , whereas Cry-W420A was significantly light responsive and was not significantly different from wild-type Cry in this response. Similarly, Cry-W342A and Cry-W397A are not light responsive in the proteolysis assay, whereas Cry-W420A was consistently degraded by light. Even though Cry-W342A lost its light responsiveness and, in most assays, proved to be inactive, in some assays, this mutant had limited ability to relieve Per:Tim inhibition, independent of the lighting conditions. The residual activity could be due to the intact W³⁹⁷ that may still be able to transfer electrons to the flavin, but at very low efficiency (Froy, 2002).

Because Cry-W397A was poorly expressed, another mutant, Cry-W397L, was generated to determine whether the loss of activity in Cry-W397A was due to its low level of expression or an inability to mediate redox. Leucine was selected for this substitution because it is a small hydrophobic residue, like alanine, and is unlikely to participate in redox-mediated reactions. Cry-W397L lost its light responsiveness and activity, while retaining a high level of expression (Froy, 2002).

To determine whether the loss of activity of Cry-W342A and Cry-W397A/L was secondary to a structural change, or whether their inactivity was due to a block in intramolecular redox, each of these tryptophan residues were replaced with tyrosine or phenylalanine, residues that are structurally unlike tryptophan but are still capable of electron transfer. Importantly, the light-dependent activity was fully restored by all four mutations (W342Y, W342F, W397Y, W397F) in both the transcriptional assay and the proteolysis assay. These results strongly suggest that intramolecular redox is involved in the light-induced activity of Cry in cell culture and that W⁴²⁰ is less critical than the other two tryptophans for this redox pathway. Furthermore, generation of a double mutant, Cry-W342A-R381A, suggests that flavin excitation is accompanied by intramolecular redox for Cry activation (Froy, 2002).

Serotonin modulates circadian entrainment in Drosophila

Entrainment of the Drosophila circadian clock to light involves the light-induced degradation of the clock protein timeless (Tim). This entrainment mechanism is inhibited by serotonin, acting through the Drosophila serotonin receptor 1B (5-HT1B). 5-HT1B is expressed in clock neurons, and alterations of its levels affect molecular and behavioral responses of the clock to light. Effects of 5-HT1B are synergistic with a mutation in the circadian photoreceptor cryptochrome (Cry) and are mediated by Shaggy (Sgg), Drosophila glycogen synthase kinase 3beta (GSK3beta), which phosphorylates Tim. Levels of serotonin are decreased in flies maintained in extended constant darkness, suggesting that modulation of the clock by serotonin may vary under different environmental conditions. These data identify a molecular connection between serotonin signaling and the central clock component Tim and suggest a homeostatic mechanism for the regulation of circadian photosensitivity in Drosophila (Yuan,2005).

Serotonin regulates the entrainment of circadian behavioral rhythms in Drosophila by affecting the molecular response to light. By modulating the expression of the 5-HT1B receptor in clock neurons, a role of this receptor subtype has been established in the regulation of Drosophila circadian photosensitivity. The data also demonstrate that the molecular connection between 5-HT1B signaling and the clock is GSK3β, which directly phosphorylates the central clock component Tim. It is proposed that serotonin signaling is a part of the homeostatic regulation that prevents dramatic fluctuations in the phase of the circadian clock. In addition, given the altered levels of serotonin in extended DD, it may confer selectivity on the response of the clock to light under different environmental conditions (Yuan, 2005).

The expression pattern of 5-HT1B, as determined by both UAS-Gal4 experiments and by immunostaining, provides some clues to its functions in Drosophila. Besides LNvs and SE5HT-IR neurons, major compartments of the fly brain that express the 5-HT1B receptor include the optic lobes, PI neurons, and mushroom bodies. Interestingly, expression in each of these locations is consistent with functions proposed for serotonin signaling in other organisms. In the housefly, the neuropil of the optic lobes undergoes daily structural changes regulated possibly by serotonin and PDF. PI neurons are neurosecretory cells that may also participate in the ocellar phototransduction pathway. The mushroom body is important for olfactory learning and memory in Drosophila. Therefore, in addition to its postsynaptic function in the LNvs, 5-HT1B may be involved in other aspects of physiology and behavior (Yuan, 2005).

The effect of 5-HT1B on Tim was especially pronounced in the small LNvs. One of the differences between the large and small LNvs is in the timing of nuclear entry, which is delayed in the small subgroup. If delayed nuclear entry accounts for the increased resistance of Tim to light in the small LNvs, it would suggest that 5-HT1B signaling largely affects cytoplasmic Tim (Yuan, 2005).

In addition to its effect on the light response, 5-HT1B overexpression influences free-running behavioral rhythms of cry^b flies. It is speculated that this is due to the loss of synchrony among LNs. The mutual coupling of oscillators within an organism is important for the generation and synchronization of circadian rhythms, and serotonin is implicated in this process in some insects. Decreased synchrony may also result from the reduced photosensitivity produced by 5-HT1B overexpression. Interestingly, a significant number of glass, cry^b double mutants, which lack CRY as well as all visual photoreceptors, are arrhythmic in DD (Yuan, 2005).

5-HT1B not only affects circadian photosensitivity when over- or under-expressed, it also appears to be the major receptor subtype required for the inhibitory effects of serotonin on entrainment. Notably, when 5-HT1B was knocked down with the RNAi transgene driven by tim-Gal4, the effect on photosensitivity was not as pronounced as with the 5-HT1B-Gal4 driver. This might be due to some background differences in flies carrying the tim-Gal4 transgene, or to nonspecific effects produced by expressing the RNAi construct in irrelevant cells. Also, the possibility that cells other than clock neurons participate in the regulation of light sensitivity via 5-HT1B cannot be excluded. However, clock cells clearly have a major role in this effect, in particular since the circadian response to serotonin is eliminated in the tim-Gal4/RNAi flies (Yuan, 2005).

Effects of serotonin on circadian photosensitivity have been demonstrated in other systems, but the underlying mechanisms were not identified. These studies in Drosophila address this issue by demonstrating an effect of 5-HT1B signaling on the posttranslational modification of Tim via Sgg. In 5-HT1B-overexpressing flies, Tim phosphorylation is reduced, and its stability is increased. In contrast, Sgg phosphorylation is increased (i.e., its activity is decreased) in response to elevated levels of 5-HT1B as well as in response to serotonin treatment. Consistent with this effect of 5-HT1B on Sgg, increased Sgg activity abolishes effects of 5-HT1B overexpression on circadian photosensitivity, while 5-HT1B attenuates the period shortening produced by excess Sgg activity. These reciprocal effects in genetic experiments strongly support the regulation of Sgg activity by 5-HT1B. Expression data indicate that Sgg is expressed predominantly in the cytoplasm. The regulation of cytoplasmic Sgg by 5-HT1B is predicted to affect the phosphorylation status of Tim mainly in the cytoplasm; Sgg-phosphorylated Tim is transported to the nucleus more effectively and is also a better substrate for light-induced degradation (Yuan, 2005).

5-HT1B alone does not significantly affect circadian period, suggesting that its effects on Sgg are limited. In this context, it is noted that, while sgg hypomorphs have a period of ~26 hr, flies hemizygous for the locus have wild-type periods. It is inferred that small (up to 50%) changes in Sgg activity do not alter circadian period but can affect circadian photosensitivity. A role for Sgg in circadian photosensitivity was previously suggested by Martinek (2001) who found that forms of Tim phosphorylated by Sgg were selectively degraded in response to light. In fact, phosphorylated Tim is known to be more sensitive to light. While Sgg appears to be the primary kinase that increases photic sensitivity of Tim, the actual process of light-induced Tim degradation involves the activity of a tyrosine kinase (Yuan, 2005).

These results provide a new mechanism for circadian regulation by a G protein-coupled signaling pathway. A role for GSK3β in the mammalian circadian system was recently reported (Iwahana, 2004). In addition, the mammalian 5-HT1A receptor affects phosphorylation of GSK3β in the mouse brain. It is possible that inhibition of GSK3β activity is a conserved mechanism in the regulation of circadian entrainment in mammals and insects (Yuan, 2005).

Slow dark adaptation has been described in Drosophila, whereby circadian sensitivity to light increases more than 10-fold over 3 days in DD. Increased light responsiveness during dark adaptation occurs in rodents, but the mechanism underlying these effects has not been addressed. Elevated responsiveness to light after prolonged exposure to darkness could be due either to a gain in sensitivity in the sensory system or to an increase in sensory output, which may be caused by a reduction in an inhibitory mechanism. In this study, lower serotonin levels were observed in flies maintained in DD. Given that serotonin signaling modulates circadian light sensitivity, it may be the reduction in this inhibitory mechanism that at least partially accounts for the enhanced light response in prolonged DD (Yuan, 2005).

It is proposed that serotonin signaling, which is itself upregulated by light, is a part of a homeostatic mechanism that regulates circadian light sensitivity. A recent study using human subjects also suggested that serotonin levels in the brain reflect the duration of prior light exposure. This change in serotonin levels with light may be relevant to the etiology and treatment of seasonal affective disorder (SAD), a mood disorder related to the reduced hours of sunlight in winter, particularly at northern latitudes. SAD patients respond to antidepression drug treatments, as well as to light therapy, both of which may produce an increase in serotonin. The interplay of serotonin, light, and the circadian system suggests a close relationship between circadian regulation and mental fitness (Yuan, 2005).

Serotonin modulates the entrainment of the circadian system. In contrast, the current results, and studies done in mammalian systems also, suggest circadian effects on serotonin signaling. (1) Based upon the differences seen in LD versus DD in the fly brain, the level of serotonin is affected by the environmental light cycle. (2) Receptor levels are modulated by circadian components, since 5-HT1B levels are altered in fly circadian mutants. In addition, serotonin release and receptor activity are regulated in a circadian fashion in mammals. Mutual regulation of the circadian and serotonin systems may be necessary to maintain the normal physiological functions of both systems (Yuan, 2005).

The Drosophila circadian network is a seasonal timer

Work in Drosophila has defined two populations of circadian brain neurons, morning cells (M-cells) and evening cells (E-cells), both of which keep circadian time and regulate morning and evening activity, respectively. It has long been speculated that a multiple oscillator circadian network in animals underlies the behavioral and physiological pattern variability caused by seasonal fluctuations of photoperiod. This study manipulated separately the circadian photoentrainment pathway within E- and M-cells and shows that E-cells process light information and function as master clocks in the presence of light. M-cells in contrast need darkness to cycle autonomously and dominate the network. The results indicate that the network switches control between these two centers as a function of photoperiod. Together with the different entraining properties of the two clock centers, the results suggest that the functional organization of the network underlies the behavioral adjustment to variations in daylength and season (Stoleru, 2007).

Two populations of circadian brain neurons, morning cells (M-cells) and evening cells (E-cells), have been connected to morning and evening locomotor activity, respectively (Grima, 2004; Stoleru, 2004). Interactions between the two oscillator populations were studied by selectively overexpressing sgg to speed up the clock in only one cell population or the other (Stoleru, 2005). This study has found that sgg overexpression gives rise to LL rhythmicity, which led to a search for the cellular substrates of entrainment. The rhythmicity is predominantly due to sgg overexpression in E-cells, which suggested that this subset of the clock network is particularly important in the light and that Sgg affects the biochemical pathway through which light impacts clock molecules and adjusts phase to the correct time of day. Indeed, strong evidence is presented that Sgg modulates Cry function, which affects in turn the core clock proteins Per and Tim. The separate manipulation of the Sgg/Cry pathway within E- and M-cells also reveals that the E-clocks drive the behavioral rhythm in light, with prominent Per oscillations of nuclear localization. This light dependence of E-cells contrasts with M-cells, which need darkness to cycle autonomously and dominate the activity output pathway. This distinction suggests a simple dual-oscillator model for how the clock adjusts to photoperiod changes, and support for this seasonal model was obtained by examining E- and M-cell cooperation under different photoperiods (Stoleru, 2007).

The free-running pacemaker and entrainment are two important and increasingly understood aspects of circadian rhythms. In contrast, little information exists about seasonal adjustment, namely, how a constant ~24-hr timekeeper accommodates dramatically different photoperiods. This study shows that the previously defined dual oscillator system in Drosophila, M-cells and E-cells, creates different rhythmic patterns by alternating master clock roles. This understanding emerged from restricting Sgg overexpression to E-cells, which allowed the E-oscillator to function and render flies rhythmic in LL. Sgg probably modulates Cry activity and, when overexpressed, provides sufficient Per and Tim to allow E-oscillator function under constant illumination conditions. The E-clocks therefore manifest free-running properties and function as the master pacemakers in LL, analogous to a previous finding that the M-oscillator is the master in DD (Stoleru, 2005). Nonetheless, these constant conditions, and even the perfect standard LD cycles commonly used in the laboratory, are poor approximations of the changing LD environments found in nature. Circadian oscillators and their entrainment mechanisms have adapted to the dramatic seasonal changes in photoperiod. The previous strategy of using oscillators with different speeds, combined with different photoperiods, has led to a model of alternating control between the M-oscillator and E-oscillator (Stoleru, 2007).

Sgg appears to attenuate, rather than inactivate, Cry activity in E-cells. This is because the LL period of timSgg/PdfGAL80 (~23.5 hr) is longer than the intrinsic period of Sgg-expressing E-clocks in DD (~21 hr) (Stoleru, 2005). A longer period in light is compatible with attenuated light perception under high light intensity conditions (1600 lx, which renders wild-type flies completely arrhythmic) and the application of Aschoff's rule to insects [Aschoff, 1979; One of the earliest observations in the study of circadian rhythms was that continuous light (LL) lengthens circadian period in most nocturnal animal species. 'Aschoff's Rule' posits that there is a positive log-linear relationship between the LL intensity and period]. As there is also a prominent effect on Cry stability, Sgg may be the regulator previously predicted to bind to the Cry C terminus (Busza, 2004; Dissel, 2004). Although Cry is favored as the major circadian substrate of Sgg, there may be others, e.g., the serotonin receptor. Biochemical support for GSK3 involvement in mammalian rhythms has recently been obtained (Yin, 2006). Since GSK3 is a proposed therapeutic target of lithium, the relationship between Sgg and Cry reported in this study recalls the intriguing relationship between mood disorders, light sensitivity, and circadian rhythms (Stoleru, 2007).

The cry^b genotype markedly affects DD period in some of the rhythmic genotypes described in this study. Although Cry is probably unnecessary for M-cell rhythmicity, this could reflect some redundancy or assay insensitivity. Moreover, the DD period of cry^b is slightly shorter than that of wild-type (23.7 versus 24.4), suggesting that 'dark Cry' makes some contribution to pacemaker function in M-cells as well as E-cells. For these reasons, it is suggested that Drosophila Cry is closer to the central pacemaker than previously believed, and therefore closer to the level of importance of its mammalian paralogs in influencing free-running pacemaker activity. Unlike mammalian Cry, however, Drosophila Cry still appears to function predominantly at a posttranslational level. Indeed, the effects of cry^b on Sgg overexpression in DD suggest that the proposed effect of Sgg on Tim stability is really an effect of Sgg on Cry followed by an altered Cry-Tim interaction. It is noted that there is a recent proposal (Collins, 2006) that Drosophila Cry, like mammalian Cry, also functions as a transcription factor in peripheral clocks (Stoleru, 2007).

The importance of E-cells in LL rhythmicity is underscored by the staining results of timSgg/PdfGAL80 brains. Only some E-cells and DN2s manifest robust cycling. It has been suspected that E-cells are important in light because they can rescue the output of arrhythmic M-cells in LD, but not in DD (Stoleru, 2004). Indeed, all of these observations make it attractive to view E-cells as autonomous pacemakers. There is, however, evidence that M-cells may not be completely dispensable. Moreover, a synchronizing or stabilization function is compatible with previous observations under different conditions (Stoleru, 2007).

In the timSgg/PdfGAL80 genotype, only Per nuclear localization changes were detectable near the end of LL cycle. The nature of the assay makes it hard to conclude that there were no differences in total Per staining intensity, i.e., no oscillations in Per levels, so the unique nature of the Per nuclear localization cycling is a tentative conclusion. The same caveat applies to the absence of Tim oscillations and nuclear staining, i.e., negative results cannot exclude low-amplitude oscillations; it is noted, however, that Tim cytoplasmic sequestration has been previously observed in cry^b flies after several days in LL. Furthermore, the circadian nuclear accumulation of Tim has been shown to respond differently than that of Per to changes in photoperiod. Nonetheless, Tim could be shuttling with a predominant steady-state cytoplasmic localization, nuclear Tim could be rapidly degraded to create a low nuclear pool, or both (Stoleru, 2007).

The importance of E-cells in entrainment is strongly supported by the potent effect of restricted Cry rescue of cry^b: E-cell rescue is much more impressive than M-cell rescue. Moreover, the differences between the two rescued phase response curves (PRCs) are striking; E-cell rescue is virtually complete, whereas the M-cell rescue is notably deficient in the delay zone. In addition, flies with Sgg overexpression in E-cells show altered PRCs, whereas flies with Sgg overexpression in M-cells respond normally to light. The results are strikingly different in darkness, as M-cell-restricted expression causes the typical short period determined by Sgg overexpression, whereas E-cell overexpression has no systemic effect (Stoleru, 2007).

The PRC delay zone is the region impacted most strongly by E-cell Sgg overexpression, indicating that the lights-off early night region is most important to E-cell function and light entrainment. Exposure to light in this interval should mimic long days (summer), which, it is speculated, will delay phase by many hours so that “evening” output of the following day will coincide with the objective evening of the environment. Even the short nights of summer are probably enough time for E-clocks to accumulate sufficient Tim and Per, shuttle them into the nucleus, and reconstitute the rhythmic substrate observed in the Sgg-overexpressing brains in LL. In contrast, M-cells need darkness to cycle robustly. They will become the master clocks and drive the system whenever lights fail to turn on more than 12 hr past lights-off, i.e., during the long nights of winter that mimic the beginning of a DD cycle. Since the intrinsic pacemaker program of M-cells in darkness relies on the changing nature of clock proteins during the night, it is hypothesized that the activity phases under long nights (winter) are locked to lights-off. This suggestion is supported by preliminary data and previous observations showing that per transcription remains locked to lights-off under different entrainment regimes. M-cells are also capable of fully entraining the system in the PRC interval that determines a phase advance (late night). This is consistent with their predicted role in generating an advanced evening output, coincident with the early evenings typical of winter. Otherwise put, long summer days should underlie light primacy as well as long and prominent evening delay zones; both suggest E-cell dominance. Night primacy and M-cells should dominate under winter conditions. This concept endows E- and M-cells with the properties originally envisioned by the Pittendrigh and Daan (1976) dual-oscillator model of entrainment (Stoleru, 2007).

A novel photoreaction mechanism for the circadian blue light photoreceptor Drosophila cryptochrome

Cryptochromes are flavoproteins that are evolutionary related to the DNA photolyases but lack DNA repair activity. Drosophila cryptochrome (dCRY) is a blue light photoreceptor that is involved in the synchronization of the circadian clock with the environmental light-dark cycle. Until now, spectroscopic and structural studies on this and other animal cryptochromes have largely been hampered by difficulties in their recombinant expression. An expression and purification scheme was establised that enables purification of mg amounts of monomeric dCRY from Sf21 insect cell cultures. Using UV-visible spectroscopy, mass spectrometry, and reversed phase high pressure liquid chromatography, this study shows that insect cell-purified dCRY contains flavin adenine dinucleotide in its oxidized state (FAD_ox) and residual amounts of methenyltetrahydrofolate. Upon blue light irradiation, dCRY undergoes a reversible absorption change, which is assigned to the conversion of FAD_ox to the red anionic FAD^- radical. These findings lead to the proposal of a novel photoreaction mechanism for dCRY, in which FAD_ox corresponds to the ground state, whereas the FAD^- radical represents the light-activated state that mediates resetting of the Drosophila circadian clock (Berndt, 2007; full text of article).

A molecular basis for natural selection at the timeless locus in Drosophila melanogaster

Diapause is a protective response to unfavorable environments that results in a suspension of insect development and is most often associated with the onset of winter. The ls-tim mutation in the Drosophila clock gene timeless has spread in Europe over the past 10,000 years, possibly because it enhances diapause. The mutant allele attenuates the photosensitivity of the circadian clock and causes decreased dimerization of the mutant Timeless protein isoform to Cryptochrome, the circadian photoreceptor. This interaction results in a more stable Timeless product. These findings reveal a molecular link between diapause and circadian photoreception (Sandrelli, 2007).

Wild European populations of Drosophila melanogaster have two major alleles of the timeless (tim) gene, ls-tim and s-tim. These alleles differ in their use of two alternative translational starts to generate longer (L-TIM₁₄₂₁) and/or shorter (S-TIM₁₃₉₈) isoforms. The ls-tim allele is derived from the s-tim allele, and directional selection is thought to have created a latitudinal gradient of ls-tim frequency within the past 10,000 years, perhaps due to an enhanced fitness of ls-tim individuals in temperate environments. TIM is a cardinal component of the circadian clock, and its light sensitivity via its physical interaction with the circadian photoreceptor cryptochrome (CRY) mediates the fly's circadian responses to light. This photoresponse can be quantified at the behavioral level by studying the fly's locomotor response to brief light pulses delivered at zeitgeber time 15 (ZT15), three hours into the night phase of a light/dark [12 hours of light alternating with 12 hours of darkness (LD12:12)] cycle that generates a phase delay of a few hours; the same light stimulus administered late at night (ZT21) generates a phase advance (Sandrelli, 2007).

The enhanced stability of L-TIM might be expected to contribute to the higher levels of TIM observed in natural ls-tim flies and to reduced circadian photoresponsiveness. Circadian light responses in Drosophila are mediated both by the canonical visual pathway, which uses rhodopsins, and by CRY. After stimulation by light, CRY can physically interact with TIM and/or PERIOD in yeast, in Drosophila S2 cells, and in vivo. These PER/TIM/CRY interactions lead to TIM degradation and subsequent PER instability, which releases the negative autoregulation of PER on the per and tim genes. Therefore the physical interaction of the L-TIM and S-TIM isoforms with CRY was examined in the yeast two-hybrid system. No interactions between TIM and CRY occurred in the dark, and the level of interaction between CRY and L-TIM in light was weaker than that between CRY and S-TIM in both plate and liquid assays. As a control, the interaction was examined of L-TIM and S-TIM with the large fragment of PER (residues 233 to 685) that is stable in yeast, but these PER/TIM interactions were not significantly different. These results indicate that the differences in interaction between the two TIM isoforms and CRY are a specific effect due to the additional N-terminal 23 residues in L-TIM, which interfere with the light-dependent dimerization of CRY (Sandrelli, 2007).

A reduced L-TIM/CRY interaction may explain the differences in the fly's circadian photoresponsiveness and the enhanced L-TIM stability. The observation that ls-tim females are more prone to ovarian diapause at any day length (Tauber, 2007) is also consistent with the results presented in this study. As in the corresponding diapause profiles (Tauber, 2007), the transformants conclusively reveal that the circadian photoresponsive phenotypes of natural tim variants are not due to linkage disequilibrium between tim and a nearby locus, but they are attributable to tim itself. Furthermore, the similarity in behavior of natural s-tim variants and P[S-TIM] transformants suggests that the residual putative truncated N-terminal 19-residue TIM product from the s-tim allele does not play any major role in the studied phenotypes (Sandrelli, 2007).

It has been argued that the light sensitivity of the circadian clock needs to be abated in temperate zones because of the dramatic increase in summer day lengths in northern latitudes. One mechanism for this process involves a reduced sensitivity to light-induced disturbance by having a higher pacemaker amplitude. However, the amplitude of TIM cycling in DD was not significantly different between the two variants, nor were there any significant differences in amplitude or phase of the tim mRNA cycle between the s-tim and ls-tim genotypes. Another way to attenuate circadian photoresponsiveness in temperate zones may be by filtering light input into the clock. The molecular changes to the L-TIM protein may buffer the circadian response to light in ls-tim individuals, even in the presence of S-TIM, and may contribute to the positive Darwinian selection observed for ls-tim in the European seasonal environment (Sandrelli, 2007).

Light-dependent interactions between the Drosophila circadian clock factors Cryptochrome, Jetlag, and Timeless

Circadian clocks regulate daily fluctuations of many physiological and behavioral aspects in life. They are synchronized with the environment via light or temperature cycles. Natural fluctuations of the day length (photoperiod) and temperature necessitate a daily reset of the circadian clock on the molecular level. In Drosophila, the blue-light photoreceptor Cryptochrome (Cry) mediates a rapid light-dependent degradation of the clock protein Timeless (Tim) via the F box protein Jetlag (Jet) and the proteasome, which initiates the resetting of the molecular clock. Cry is also degraded in the light but whereas the degradation of Tim is well characterized, the mechanism for light-dependent degradation of Cry is mostly unknown. Until now it was believed that these two degradation pathways are distinct. This study revealed that Jetlag also interacts with Cry in a light-dependent manner. After illumination, Jetlag induces massive degradation of Cry, which can be prevented in vitro and in vivo by adding Tim as an antagonist. The affinity of Tim for Cry and Jetlag determines the sequential order of Tim and Cry degradation and thus reveal an intimate connection between the light-dependent degradation of these two proteins by the same proteasomal pathway (Peschel, 2009).

Jetlag is involved in the resetting mechanism of the circadian clock. Jet is a putative component of a Skp1/Cullin/F-Box (SCF) E3 ubiquitin ligase complex that associates with Tim in a light-dependent fashion in an embryonic Drosophila cell line (S2) in the presence of Cry. This interaction promotes the ubiquitination and degradation of Tim in cultured cells. In nature, two Drosophila allelic variants of timeless can be found: one allele produces a 23 amino acid N-terminally shortened and more light-sensitive form of Tim (s-tim), the other allele encodes both forms (ls-tim). At the molecular level, S-Tim's enhanced light sensitivity is correlated with (and likely due to) enhanced binding to the circadian blue-light photoreceptor Cry (Peschel, 2009).

The hypomorphic jet^c mutation carries a single amino acid change in the leucine-rich repeat (LRR) region of Jet, which causes flies to be rhythmic in constant light (LL), but only if they express the less light-sensitive L-Tim protein (encoded by the ls-tim allele) as is the case in Veela flies (Peschel, 2006). The LL-rhythmic Veela phenotype resembles that of cry mutants. Also similar to cry^b mutants, homozygous mutant Veela flies accumulate abnormally high levels of Tim protein in the light. Strikingly, both phenotypes are also observed in transheterozygous Veela/+; cry^b/+ flies (Peschel, 2006). This strong genetic interaction between tim, jet, and cry prompted an investigation of a potential physical association between Jetlag and Cry proteins in the yeast two-hybrid system (Y2H). In addition, the two different Timeless isoforms were also tested for interaction with Jetlag or Cryptochrome. In agreement with an earlier study, light-dependent interaction between both Tim proteins and Cry was observed, whereby S-Tim interacted more strongly with Cry as compared to L-Tim. Surprisingly, a striking light-dependent interaction between Cry and Jet was observed, but not between Tim and Jet. Given that Tim and Jet do interact in S2 cells cotransfected with cry and the finding that Jet interacts with Cry in yeast, an explanation for the lack of Tim:Jet binding could be that Cry is essential for this interaction (Peschel, 2009).

The interaction between Jet^c and Cry is significantly weaker compared with the wild-type protein. Keeping in mind that the LRR is the binding region for the F box proteins' substrate, this weaker association was expected. Additionally, Jet and Jet^c were challenged with different Cry mutations. In Cry^Δ the last 20 residues from the C terminus are missing, resulting in strong, light-independent interactions of Cry^Δ with Tim. A strong light-independent interaction was observed between Jet or Jet^c and Cry^Δ. The Cry^b protein does not interact with Jet or Jet^c, correlating with its inability to bind to Tim in yeast (Peschel, 2009).

The strong biochemical and genetic interaction between cry and jet suggests that the Jet:Cry interaction is important in vivo and perhaps required for efficient light-induced Tim turnover. Given that a direct interaction between Jet and Tim was not detected in yeast, this implies that binding of Cry to Tim could modify Tim in a way that it now can bind Jet to induce degradation. Alternatively, the Jet:Cry complex binds to Tim (via Cry acting as a bridge), thereby inducing Tim degradation (Peschel, 2009).

To distinguish between these two possibilities, CoIP experiments were performed in an embryonic Drosophila cell line (S2). A full-length Jetlag protein fused to a HIS-tag (Jet-H) and untagged versions of Cry and Tim proteins were overexpressed in S2 cells and immunoprecipitated with HIS antibody. Cells were grown in darkness and exposed to light for 15 min before performing the assay. As expected from the Y2H results, Cry also interacts with Jet-H in S2 cells. Contrary to the Y2H results, Tim also interacts with Jet-H, without the addition of Cry. When Tim and Cry were simultaneously expressed in the presence of Jet-H, only minimal amounts of Tim protein could be detected in the input or CoIP fractions. It is speculated that the low Tim levels were caused because a fully functional light-sensitive clock-resetting protein complex was reconstituted. To test this, the CoIP experiments were also conducted in the presence of the proteasomal inhibitor MG-132, which led to an overall stabilization of the proteins and a clear demonstration of Tim:Jet interactions in S2 cells. The interaction of Tim with Jet is increased in the presence of Cry, supporting the idea that Cry:Tim or Jet:Cry complexes promote binding of Tim to Jet (Peschel, 2009).

Why could Tim:Jet interactions be detected in S2 cells but not in yeast? The reason for this could be that a crucial phosphorylation step necessary for the detection of Tim by Jet is not performed in yeast, but does occur in Drosophila cells. Alternatively, the low endogenous Cry levels in these cells could promote the Tim:Jet interaction, perhaps contributing to the required posttranslational modification of Tim. Therefore attempts were made to reduce the low endogenous CRY levels even further by cry-dsRNA-mediated interference before conducting the CoIP experiments. dsRNA treatment efficiently reduces transfected Cry levels, indicating that endogenous Cry levels should also be reduced by this treatment. CoIP experiments in the presence of MG-132 show that Jet:Tim interactions are dramatically reduced (but still detectable) after dsRNA treatment. This demonstrates that endogenous Cry levels are supporting Jet:Tim interaction observed in S2 cells (Peschel, 2009).

In cells transfected only with tim, a Jet:Tim interaction was detected, but not a Jet:Cry interaction. Even though the input levels of endogenous Cry and transfected Tim are very low, one would expect to precipitate equal amounts of both proteins bound to Jet, if Cry would indeed form a bridge between Tim and Jet. This was not observed, and only Tim was repeatedly precipitated, indicating the existence of Tim:Jet complexes that are free of Cry. The results therefore support a model in which Cry modifies Tim, allowing Tim to interact with Jet after dissociation of the Cry:Tim complex (Peschel, 2009).

If the Jet:Cry interaction is biologically relevant, an effect on Cry degradation should be detectable in flies with reduced jet function. Indeed, jet^c flies exhibited mildly increased Cry levels after 2 and 11 hr in light. Interestingly, in the light phase, s-tim animals accumulate higher levels of Cry compared to ls-tim flies, both in jet⁺ and jet^c genetic backgrounds. Cry associates stronger with S-Tim compared to L-Tim, and in flies this probably leads to a more efficient light-dependent degradation of S-Tim. This suggests that the affinity of the Cry:Tim interaction dictates the temporal order of Tim and Cry degradation by Jet -- in other words, S-Tim would be preferentially degraded, followed by the turnover of Cry, whereas L-Tim enhances the degradation of Cry because of its lower affinity to this photoreceptor (Peschel, 2009).

Because the differences in Cry degradation caused by jet^c were subtle, it was desirable to confirm this effect by creating a more severe reduction of Jet function. For this, the stronger jet^r allele was combined with jet^c or a deficiency of the jet locus. Both combinations lead to substantially increased Cry levels compared to controls and homozygous jet^c mutants. This unequivocally demonstrates that jet influences Cry stability in flies. It was also noticed that the absence or presence of eye pigments influences the amount of Cry degradation after light exposure, perhaps because the pigments 'protect' Cry from the light (Peschel, 2009).

Although cry is expressed in S2 cells, the endogenous Cry protein is unstable in S2 cells. jet (but not tim) is also expressed in these cells. Endogenous jet expression may explain the previous observation of Tim ubiquitination in S2 cells without cotransfection of Jet. This suggests that the amount of Jet (and) or Cry is limiting for triggering Tim degradation. To test this, S2 cells were first transformed with cry, jet, and s-tim or l-tim. Cells transfected with cry and tim showed little degradation of Tim, regardless of the Tim form present. In contrast, cotransfection of jet led to massive Tim degradation, suggesting that the endogenous Jet amount is limiting. A slight reduction of Tim degradation was also observed after cotransfection of jet^c and the long isoform of Tim, confirming previous results obtained in adult flies (Peschel, 2009).

After establishing conditions that recapitulate light-induced degradation of Tim in cell culture, it became possible to study Cry levels after illumination. Transformation of increasing amounts of jet plasmid DNA is correlated with increased degradation of Cry. This effect is indeed caused by extra jet, because transformation with equal amounts of unrelated plasmid DNA did not result in reduced Cry levels. When the jet^c mutation was used, Cry degradation in the light was reduced but still visible, confirming the results obtained with adult flies. Both effects are possibly caused by the poorer ability of Jet^c to physically interact with Cry (Peschel, 2009).

So far, these results suggest that Tim is preferentially degraded, when both Tim and Cry are present. If true, addition of Tim should stabilize Cry in S2 cells. Indeed, after 10 or 120 min of light exposure, a dramatic 'protection' of Cry by Tim was observed. Cotransfection of Jet restored the light-induced degradation of CRY, at least after 2 hr of light exposure. It is concluded that Tim indeed protects Cry from light-induced degradation, most likely because it is the preferred target of Jet (Peschel, 2009).

To further prove that both proteins are a target of Jet and subsequent proteasomal degradation, the proteasome inhibitor MG-132 was added to cells transfected with Cry and Jet. As previously shown for Tim, light- and Jet-dependent degradation of Cry was largely prevented after adding the drug, suggesting that Tim and Cry are degraded via the same pathway. Similar as in flies, a minor Jet-dependent reduction of Cry levels, which seems independent of light and the proteasome, was observed, indicating that Jet also promotes Cry degradation via a different, light-independent pathway (Peschel, 2009).

An assay was developed that allowed examination of the light-induced degradation of Cry with a higher temporal resolution and in a more quantifiable manner. A constitutively expressed firefly-luciferase cDNA was fused to full-length Cry (Luc-dCry) or to a C-terminal truncated version of Cry (Luc-dCry528). In S2 cells, the fusion protein Luc-dCry is degraded in a similar way as Cry alone, whereas the truncated Luc-dCry528 is expressed at a very low level. After transient transfection of the luc-dCry gene, luminescence was measured in an automated luminescence counter. After illumination, the Luc-dCry protein is swiftly degraded and in darkness Luc-dCry levels recover, demonstrating that the system nicely reflects the light-dependent degradation of Cry. When Jet is added to the cells the fusion protein is degraded even faster -- an effect not observed when Jet^c is added. Lower Luc-dCry levels are observed in the dark portion of the day when Jet is present. Cotransfecting luc-dCry with timeless results in a striking stabilization of Cry in S2 cells, confirming the western blot results. The magnitude of this effect depends both on the isoform and on the total amount of Tim. S-Tim inhibits Luc-dCry degradation more strongly as compared to L-Tim, indicating again that the high-affinity S-Tim:Cry interaction stabilizes Cry more efficiently. Adding Jetlag and Tim at the same time leads to decreased Cry turnover, compared to transfection with Jet alone, but Cry is less protected if Tim is added alone. Overall, these luciferase results nicely confirm the S2-cell and whole-fly western blot results and demonstrate that Jet promotes Cry degradation, which is counteracted by Tim, and especially S-Tim (Peschel, 2009).

Next the Luc-dCry protein was expressed in UAS-luc-dCry transgenic flies under the control of a tim-Gal4 driver. Robust Luc-dCry oscillations, which are due to light-dependent degradation in transgenic flies, was observed, because a sharp decrease of luciferase signals coincides exactly with 'lights-on' in every cycle, and the oscillations immediately stop after transfer to DD. This result is in agreement with light- but not clock-regulated oscillation of the Cry protein in flies. Overexpression of Tim with a UAS-tim transgene led to significantly elevated levels of Luc-dCry during the light phase, which is quite remarkable given that these flies contain the endogenous wild-type allele of jet. Because both transgenic genotypes contained the identical and single copy of the UAS-luc-dCry transgene, this difference in the level of Luc-dCry must be due to the overexpression of Tim. Therefore, the increased daytime Cry levels in the transgenic flies are most likely caused by a stabilization of Cry by Tim, similar to that observed in S2 cells. Also similar as in S2 cells, although the Luc-dCry protein is stabilized by Tim, it can still be degraded by light as long as Jet is present. Interestingly, closer inspection of luc-dCry expression in flies reveals that Cry levels in the UAS-tim flies already recover during the light phase, indicating that Tim mainly protects Cry when light is present. A western blot from flies with the same UAS-tim transgene under the control of a tim-Gal4 driver also reveals a dramatic increase in the levels of Cry and confirms the luciferase results. Both the western blot and real-time luminescence data show that Jet supports the light-dependent degradation of Cry in vitro and in flies and that Tim interferes with this process (Peschel, 2009).

The fact that Tim stabilizes Cry can most easily be explained if Tim is the preferred target for Jet. If true, one would predict that in flies light-induced degradation of both Tim and Cry occurs in sequential order; Tim being degraded ahead of Cry. Therefore Tim and Cry levels were simultaneously measured in head extracts of wild-type flies (y w; s-tim) during the first 10 hr of light in a LD cycle. Although levels of both proteins start to decrease after the lights are turned on, and trough levels are reached at the same time (ZT4), Tim degradation appears to occur more rapid in the early day. This result is in agreement with the idea that Tim is preferentially degraded after initial light exposure. Interestingly, a similar result was reported for Cry and Tim degradation kinetics in adult clock neurons (Peschel, 2009).

Recently, a genome-wide cell-culture-based RNAi screen has been performed in order to identify genes involved in the light-dependent degradation of Cry. Interestingly, Jet was not among the identified candidates. Instead, two other ubiquitin ligases encoded by Bruce and CG17735 were reported to affect light-dependent degradation in flies. The effects reported in this study were caused by eye-color differences between mutants and controls. Therefore, Bruce and CG17735 likely do not contribute to light-dependent Cry degradation in flies, which is also the case for two other ubiquitin ligases that were shown to affect Cry degradation in vitro (Peschel, 2009).

In conclusion, in Drosophila, the clock factor Timeless is degraded after illumination, resulting in a daily reset and adaptation of the circadian clock to its environment. This study has demonstrated that the blue-light photoreceptor Cryptochrome directly interacts with the F box protein Jetlag in a light-dependent manner. This interaction leads to the degradation of Cry by the proteasome and it was unequivocally shown that Jet regulates Cry turnover in vitro and in flies. This is an important and surprising observation, given that so far it was assumed that Cry and Tim are degraded via different pathways. In agreement with previous studies, it was also found that Tim also associates with Jet, but the results suggest that a posttranslational modification of Tim, induced by its binding to Cry, is a prerequisite for the Jet:Tim association. Cry is dramatically stabilized in the presence of Tim, which can be explained by an increased binding affinity of Jet toward light-activated Tim compared to Cry. Based on the results, a more complex model for light resetting is proposed: light induces a conformational change in Cry, allowing it to bind to Tim. S-Tim binds to Cry with higher affinity compared to L-Tim, which leads to more efficient S-Tim degradation by Jet and stabilization of Cry. L-Tim interacts weaker with Cry, presumably resulting in a weaker Jet-L-Tim interaction (or fewer Jet-L-Tim complexes) and less efficient L-Tim degradation. As a result, Cry is less stable in L-Tim flies, because it becomes a better substrate for Jet. Consequently, even less Cry is available to bind to L-Tim, which could further contribute to the reduced light-resetting responses observed in ls-tim flies compared to s-tim flies (Peschel, 2009).

What could be the advantage of such an interdependent binding and degradation of light-regulated clock proteins? The results suggest that Tim and Cry may be degraded in a sequential order. As long as Jet triggers the degradation of Tim, Cry would be spared, presumably because Jet's affinity to light-activated Tim is much higher than to Cry. After Tim levels have decreased to a critical amount, Cry is no longer needed and is now the prime target of Jet. Possibly the degradation of Cry then allows a reaccumulation of Tim in the next circadian cycle, which would also explain why Tim levels start to increase already during the late day (Peschel, 2009).